External Commercial Borrowings (ECBs), a channel to facilitate access to foreign capital by corporates, has evolved as a preferred medium of borrowing in the post-liberalization period in India. Since 2000, the total external borrowing of corporates (including short-term credit) on an average has accounted for more than four-fifths of India’s total external debt and about one-fourth of all capital flows, indicating the growing importance of ECB in India’s macroeconomy. The extant literature on India’s ECB is scarce and can be broadly categorized into macro and micro-level studies. The macro literature has analyzed India’s ECB in the context of overall capital flows, or, in comparison to other debt and equity flows. Most analyses have overwhelmingly focused on the domestic macroeconomic factors driving ECB flows to India, largely ignoring the global factors, critical to any analysis in an open economy context. The micro literature has explored several aspects of firm dynamics like -why firms borrow, which firms borrow, and which firms benefitted from ECBs. Studies have also analyzed the financial vulnerabilities and the various risks like exchange rate risk, interest rate risk, and risks from unhedged foreign borrowings emanating from rising financial dollarization of these firms. However, the majority of these studies have adopted cross-sectional analysis and limited the approach to sensitivity analysis of distress debt to macroeconomic shocks or analyzed the impact on real outcomes like investments and profitability. Thus, such studies provide little guidance on: (a) the actual nature and evolution of corporate vulnerability; (b) the impact of rising foreign currency leverage on corporate vulnerability; (c) the time-varying impact of macroeconomic shocks and the subsequent policy interventions on the financial vulnerability of the foreign-currency leveraged firms; and (d) whether there exist significant differences in leverage and vulnerability among firms across different sectors of the economy. The three essays of the thesis aim to address the critical gaps in the extant macro and micro literature, as identified above. The first essay, departing from the approach in the extant literature: (a) studies ECB on a standalone basis; and (b) examines the role of domestic and global factors in driving ECB flows to India. Using monthly data for the period 2004M1 to 2015M12, and employing the Johansen cointegration and error correction techniques, the analysis suggests that both domestic (pull) and global (push) factors like exchange rate, economic growth differential between India and the OECD, the interest rate differential between the domestic lending rate and LIBOR, and international liquidity, have significantly influenced India’s ECB inflows. In particular, the study establishes the effect that the country’s capital account openness and sovereign credit risk ratings have had on ECB flows, variables hitherto not considered in the existing literature. The second essay focuses on the study of Indian firms with ECB in their portfolio, and analyzes the impact of rising leverage, as well as foreign currency leverage, on the financial vulnerability of these firms. The study considers a sample of 825 non-financial firms, which borrowed consistently from international capital markets between 2004 and 2017 and employs the Arellano Bond dynamic GMM technique for analysis. The results of the study suggest that both leverage, and foreign currency leverage significantly explain the rise in the financial vulnerability of these firms and have been particularly pronounced during the boom periods of ECB inflows to India. Further, the study finds that bigger sized firms were more vulnerable; profitable firms were less vulnerable, and export earnings were rendered inadequate hedges to reduce the vulnerability of these firms. Finally, the spillover analysis indicates that corporate default (of distressed firms), could erode the banks’ asset quality, increasing the stress in the Indian banking system. Finally, the third essay studies the impact of exchange rate shocks, and changes in monetary policy stance adopted in the advanced economies vis-à-vis that in the domestic economy in influencing financial vulnerabilities of the Indian corporates accessing international debt markets. Employing the Dynamic Generalized Method of Moments (GMM) of Arellano Bond, the empirical investigation in the study indicates that conditional on a depreciating currency regime in India, rising foreign currency leverage has intensified the financial vulnerability of these firms, particularly in the post-GFC period. Further, the study finds that the unconventional monetary policy stance adopted by the advanced economies during and after the global financial crisis of 2007-09, further aggravated the financial vulnerability of the corporates. The sectoral analysis revealed that while a depreciating currency adversely affected the firms in the manufacturing sector, the service sector firms, being predominantly export-oriented, benefitted from a depreciating currency. The thesis makes several contributions to the existing literature on ECB in the Indian context. The first essay contributes to the literature on pull and push factors and furthers the understanding of the macroeconomic determinants of cross-border commercial debt flows to India. The second and third essays, both of which focus on firm-level analysis, contribute to the existing literature on corporate vulnerability associated with rising foreign currency leverage. Additionally, while the second essay brings new insights on the potential cascading effects of rising foreign-currency leverage of distressed non-financial firms on the banking sector stability in India, the third essay complements the literature on contagion and the potential spillover effects of monetary policy in an emerging economy context.

The Big Data revolution is upon us. The revolution had been in the making for six years, in the shape of Hadoop, till December 27, 2011 – when release 1.0.0 got announced to the world. In 2012 IIM Ranchi launched the FPM programme and announced intake into FPM(Analytics) supported by a proposed IBM Center for Analytics. Big Data was a rather fresh concept, and the technologies were barely known. Definitions itself of Big Data were still emerging and over the years, there could not have been even a consideration that the time for convergence of those definitions had arrived. There were debates continuing about the V’s of Big Data, and new V’s were emerging. Also, the use cases of the technologies were not available and multi-node cluster, distributed computing and related terminology such as IoT were buzz words – barely understood beyond mention. Concepts such as meta data and artificial intelligence – were being re-examined considering the Big Data phenomenon. Besides the daunting economics, the hardware requirements and the technical know-how required to set up multi-node distributed computing environments were esoteric knowledge- hardly known to most and continue to be a challenge, even at this time. As the research related to this thesis dug deeper, it became clear that there were considerable technological concepts and components involved inevitably in Big Data Analytics – that needed to be assimilated. Also, by 2014, doubts started emerging that Big Data could be of value– there were criticisms of the concept and Big Data was even called a “hype” in the interim. It was increasingly being pointed that the ability to derive actionable intelligence from Big Data held value for management decision making. It was in this light, that there were discussions in technical channels and academia, about the role of Context and the role of Metadata, IoT and Security – and how these impact the ability to gather intelligence, from Big Data. Also, the diversity of these so many concepts implied there was a need to first develop a visualization of the “entire Elephant in the room”, to be able to have a chance of creating some persistent knowledge – the intention of this thesis. This thesis therefore centers on the following themes: (1) Meta-data, context and Big Data (2) The Definition V’s of Big Data – conflicts within and the criticality of Veracity (3) Enabling context-aware travel for flyers from Social Big Data (4) IoT’s, Big Data, Education & Transportation and (5) Internet Security and Big Data. These concepts and their interactions with Big Data are discussed in corresponding chapters and though considerable attempts have been made to delineate the related knowledge within each chapter, so that these are in-depth and well-contained discussions as could be – yet these must continue to be aligned with the commonality brought by the thesis topic. In the process, this work delivers additions to theory and demonstrates concepts that may be applied in various situations for enabling better business performance. It is through Chapters 1 and 2 that a foundation is developed for these milestones along the journey through Context and Big Data. For being able to arrive at the objective, i.e. enablement of enhanced context-aware air travel experience, it was important to explore and confirm the validity of context – as an information entity, and its relevance to Big Data. Therefore, concepts of context, context-awareness are introduced, and a picture of the work done so far in the area is developed. Various definitions of context are encountered, and varying concepts and applications of context-awareness are discussed. Several context-related concepts and frameworks are briefly discussed, while laying emphasis on the importance of context-awareness to concepts like Artificial Intelligence, pervasive computing and ubiquitous computing. In the process of evaluation of context-aware applications, it is observed that in the air travel sector, there is considerable potential to enable a context-aware enhanced travel experience. Many researchers, and for long – have pointed out the need for context-aware travel. Although there is rather scarce literature that has begun to examine applications of context-aware travel, it is also noticed that a focus on the cognitive aspects is yet to emerge. For example, how do flyers interact with, and feel about – the various components of their travel experience? Thus, the thesis delves into literature, to develop an understanding of the extant work in context-aware air-travel for a flyer. It is subsequently observed that there is a possibility of many contexts in air-travel, i.e. richness - because flyers often have varying needs, often have varying circumstances of travel and it is possible that all those may combine to even give rise to several travel contexts. Yet, it also becomes clear that data about air-travel is hardly available in formal airline monitoring databases and travel databases of various Governments, to the level of detail that the data may be considered more representative of flyers concerns. It also is learned that air travel data in which flyers contribute knowledge of their air-travel experience, formally – can be price muted. Thus, it becomes clear – much formally available data may not be as dense and thick in contexts of air travel experiences of flyers. At the same time, further parsing of extant literature informs that flyers have turned social media savvy, that airlines and organizations are aiming to serve up information to flyers through mobile apps and that flyers are now contributing enormous amounts of data via their social posts and by making public their air-travel experiences. This led to the identification of Social Big Data, i.e. social media flyer reviews as a potential source of context-rich information, with the advantage that flyer reviews data was publicly available for free downloading. Having developed a foundation, this thesis then starts building knowledge about the other key aspect of this thesis, i.e. Big Data. Since Big Data is a relatively new phenomenon that started accumulating literature as this research evolved, the first step was to develop an understanding of Big Data itself, its definition, the various issues that were emerging related to Big Data – i.e. issues about lack of context and therefore loss of meanings. This thread led to the need for understanding what context is, to questions like: do computers understand context as easily as humans do, if not - how is context developed or made available to computing – for example through meta data, what extant work describes context & its utility and why is context important. This leads to a methodology for understanding development of context through meta-data, and its role in Big Data through consideration of information sources that could provide valuable relevant information, since academic literature, it was found, does not provide a foundational perspective. Therefore, this work drew upon feedback from industry, also upon surveys conducted by a diverse set of industry leaders, on experiences of industry experts and on practices adopted by those working with metadata and Big Data analytics in various application areas. Though, it must be noted, this work does not shy from utilizing academic contributions, where available. Thus, with the emphasis developed, three cases are presented to create a strong foundational research paper on meta data and context. It becomes clear that meta-data is important for development of context and that, context is important for Big Data. This work forms Chapter 3 of the thesis and has been published as a research paper in the International Journal of Allied Practice, Research and Review. While exploring the phenomenon of Big Data, this research also came across several definitions of Big Data and it is noticed, that in extant literature there are conflicts and gaps in the understanding of Big Data- its definition with respect to the famous V’s of Big Data, i.e. what V’s should be included in the definition of Big Data, and the linear order of evolution and prioritization of the V’s of Big Data. It is also noticed that “Veracity” is the most neglected V of Big Data research that has somehow failed to gather sufficient interest and understanding in extant research. In Chapter 4, as further exploration uncovers various aspects of veracity, new observations emerge. These observations are further formed into research questions: [1] What does the term veracity mean, in context of Big Data? [2] Can veracity impact decision-making? Is veracity even important for business? [3] Is extant literature sufficient to provide a comprehensive understanding of veracity? How difficult or easy is it, to manage veracity? [4] Does extant literature inform about the emergence of veracity in the world of Big Data? [5] What importance have researchers accorded to veracity, i.e. is it the third or fourth or fifth or some other V- of Big Data? [6] Should veracity be included in the definition of Big Data? To develop a comprehensive understanding of veracity through those research questions, this research then utilizes the concept of “Context” as an information entity. A broad yet deep context for veracity is developed, using the 5W’s. Additional findings from extant literature that emerge, are also utilized to develop answers for the research questions. Thus, this research delves into an in-depth study of the above issues and establishes a strong understanding of Big Data, a new way of defining Big Data and a new way of looking at the V’s that presents a conflict free definition of Big Data as well. In the process, it is found that veracity is contextual – and therefore a cross-sector operational definition of veracity can be useful for research. Having examined in detail various definitions of veracity, this work then is able to offer a cross-sectoral operational definition of veracity, in the process resulting in an unprecedented yet comprehensive understanding of veracity. This work forms the basis of a research paper titled “A Comprehensive Understanding of Veracity for Big Data and Implications” under consideration(R2) for publishing in The Journal of Big Data. Having developed a deeper understanding of context, Big Data, meta-data, some of the conflicts within and having developed a bigger picture about various issues involved including veracity – this research finally focuses on developing a context-aware framework for air travel. Chapter 5 reports findings of this experimental study. Organizations have developed Big Data capabilities to process enormous amounts of data – Petabytes and such. It is the lower end of the Volume-spectrum of Big Data that has not found enough easy to use, low infrastructure solutions that may require little learning yet may offer real time interaction possibilities and human friendly visualization. Studies have found that SMEs (small and medium enterprises) need to make use of Big Data for innovation but are limited by a lack of Big Data capabilities. However, it is also reported in literature that SMEs are highly concerned about data privacy and security – which makes access to data near impossible. Since SMEs operate with IT systems and data at a scale that is much lower than typical large enterprises, it can be drawn that SMEs may also qualify for the lower end of the Big Data spectrum problem. This implies, if there are other classes of users that have a typical lower end of Big Data problem- there may be some transferable benefits from a relevant Big Data approach. It is while searching for a similar class of users, it is noticed that typical media consumers such as airlines flyers have lagged in consumption of information, because Big Data business so far has failed to even acknowledge that the social media data flyers can often have to deal with, has similar Big Data challenges, albeit at the lower end of the Big Data spectrum. The additional advantage with using Social media data often utilized by flyers is that such data is freely available without privacy and security concerns. In this part, experiment-based research establishes and demonstrates a context-aware approach that can help flyers self-learn from social media data such as flyer reviews. It show-cases ContextSys- a framework and, InContext, an in-browser Social Media data processor that has been tested with nearly 42,000 customer reviews expressed as nearly 0.5 million lines of text equivalent to nearly 17,500 A4 sized pages of text. InContext uses light-weight technologies for processing enormous amounts of Social Media data and carries out considerable Big Data tasks, real-time in an interactive manner without requiring the use of any (1) external dictionaries, (2) SQL or NoSQL like querying, and (3) libraries. InContext enables: (1) exploration of text data (2) keyword and phrase matching (3) data partitioning (4) data reduction (5) extraction of entities and dictionaries (6) extraction of relationships between entities (7) extraction of specific information in-context, say regarding food and drinks, in-flight entertainment or seating comfort and (8) comparison of flyer reviews with that specificity. It must be understood that the lack of capabilities for flyers to process social big data is not without implications. While organizations continue to flood social media channels with potent information, which may even be customized and useful - for flyers that is addition to the already overwhelming information overload. Unable to dig through mounds of data in a timely and efficient manner - instead of feeling helped, flyers can feel thwarted from making timely informed decisions that could ensure an enhanced context-aware air travel experience. Therefore, it is not surprising that the social data ecosystem is facing abandonment from at least some users unable to cope up with the information overload. If the pace at which flyers, or users in general, are being bombarded with information becomes too high - it may not be too long before more flyers start questioning participation in the information democracy unleashed by social media. That has the potential to spoil the gains of the information revolution. This work has potential for significant impact on business. Especially the high usability and low resource requirements have implications for design of comparative apps by business. Also, by enabling flyers to consume more information – this work proposes to alleviate information load – thus increasing the availability of flyers to even more information from business. These principles of Big Data information management could also be extended to the Big Data problem of SMEs. Thus, it becomes possible for various classes of users- flyers, SMEs included - to partake the fruit of Big Data at that end of the Big Data spectrum, that otherwise may not be possible. The ContextSys framework could be implemented using any of the latest technologies. This work forms the basis of a research article titled “A Long Desired Self-Learning System To Aid Big Data Business Social Media Consumption” accepted for publication in the Journal of Academy of Business and Economics. Chapter 6 presents the research and findings on the application of IoT’s and Big Data as explored in the education sector through transportation- in the form of a case study, and on the application of Context to Internet Security. Work in these emerging areas demonstrates how IoTs, Context, Big Data, Education, Transportation and Internet Security can co-relate, and how these can together help answer questions that can be critical for some sectors. Chapter 7 presents the conclusions of this work while also mentioning the direction future work may be able to take.

There is a renewed interest in Recommender Systems (RS) research after the Netflix prize competition. In parallel, there is also a growth of social media which has driven research in link prediction, with the aim of predicting labels about friends, groups, etc. Both these problems involve the prediction of labels (star ratings or friendship) between a pair of entities (user movie or user-user). This type of problem is known as a dyadic prediction. Dyadic prediction in context of ‘personalization’ on e-service platforms is a broad research area of this dissertation. In particular, this dissertation seeks to contribute three folds on the implementation of personalization tools on e-commerce. The first contribution of this work is to develop novel optimization algorithms in dyadic prediction that improves the precision of RS. Since the foremost utility of RS is to recommend good items to users, it is worth concentrating on improving the precision of the RS model. Keeping in line with this thought, this work pursues to develop an entirely novel latent factor model which optimizes its parameters on the likes and dislikes of users for items instead of numerical ratings. The novelty of the proposed algorithm is to optimize latent factors of user and items that are derived from the likeability of users to items. The proposed algorithm has also been extended to the ranking task which considers matching the order of labels rather than the actual label itself. Secondly, the work extends regularized singular value decomposition (RSVD), state-of the- art algorithm in rating prediction, to provide optimal matching for long-tail items to idiosyncratic users based on prior likeness to long-tail items. The second model is also extended to an additive latent factor model that promotes long-tail items to users based on their prior interests. This work discusses business goals that a firm has and pushes according to its requirement so that the trade-off between novelty and diversity can be linked with the accuracy of the RS. Lastly, this work also develops a new model for context-aware recommender systems (CARSs) which is an upcoming and challenging research area. The priority of this work would be to model contexts with lesser complexity. Comprehensive experiments are conducted on publicly available benchmark datasets such as Amazon and Movie Lens to verify proposed research work with existing state-of-the-art algorithms. The experimentation evaluation frameworks are in line with existing frameworks proposed in related literatures.

Early marketing thoughts were highly descriptive of commodities and inherited manufacturing-based models. Recent studies have advocated for the development of service focus and contribute to the reorientation of marketing by aligning customer processes. The corresponding shift in service strategies has focused on the convergence of the new perspectives to form a new logic for marketing which is based upon co-creation of value and relationships. Furthermore, service providers operate in a rapidly changing environment and need a customer-centric philosophy for value creation. Therefore, a constant reinvention by the service provider in terms of flexibility has become a fundamental premise. Flexibility provides the potential to the firm to adjust in multiple fronts and deliver customized services rapidly. Despite the relevancy of flexibility in modern service business environment, the extant literature doesn’t explicate the dynamics of flexibility and its link with experiential value. . This thesis attempts to understand the dynamics of service flexibility and its link with customer experience and satisfaction in the context of public healthcare system in India. A multi-level case study was conducted in a public healthcare organization to explicate the multi-dimensionality of service flexibility. The qualitative findings proposed a framework of flexibility that assimilates several dimensions and provides a unique classification of service flexibility in terms internal flexibility and market-focused flexibility. The network of flexibility dimensions also explains the constituents of internal flexibility and provides several dimension of market-focused flexibility. The relationships proposed in the framework indicate that the dimensions of market-focused flexibility have a line of visibility with customers regarding adaptations to their changing demand and the consumption of customized services. Likewise, internal flexibility represents the adjustments in service capacity and process mechanisms of providers which influences the genesis of market-focused flexibility. Thus, the framework explains how the phenomena of flexibility across the delivery system amplifies the consumption experience. Further, a survey of healthcare professionals was conducted in public healthcare setting in Indian in order to investigate the relationships depicted in the framework. The results of the structural equation modelling support that internal flexibility as a higher order factor positively influences the dimensions of market-focused flexibility which further impacts on customer satisfaction. The effects of market-focused flexibility on customer satisfaction was also gauged by giving voice to the customers and including customer data into the study. Thus, the findings were also validated through a dyadic sample. The study provides several implications for theory and practice.

The world is transitioning from the traditional way of delivering public services to the electronic way; this is referred to e-government. There is a huge investment is being done on the e-government initiatives across the world including India as well. However, for a successful e-government implementation it is important to understand citizens’ perception towards these e-government initiatives because no matter how well-designed e-government services are, they will not succeed in realizing promised benefits if citizens are unwilling to adopt them and shift from offline to online engagement. With the current advent of e-government services, user experience and factors influencing the adoption of e-government services poses interesting questions in information systems research and practice. Also, how masses, in a country like India with the digital divide, are accessing and getting benefits out of the e-government initiatives poses an interesting question and still remains unanswered in e-government research, a phenomenon of the last two decades. The questions that this dissertation seeks to illuminate are (i) how users experience e-government services? (ii) what are the critical factors for indirect adoption of e-government services? (iii) what are the critical factors to adopt e-government services at different maturity levels? This dissertation addresses each of these three questions in three respective parts. Qualitative research has been used in each part of the study using semi-structured interviews. The first research question directing towards exploring users’ experience of e-government services, hence phenomenology approach is found to be most appropriate. Purposive sampling was used to recruit the participants for the first part of the study. The second and third research questions are grounded on the existing theoretical gaps to explain the relationship between the influencing factors and adoption intention, both at different maturity levels as well as indirectly; hence grounded theory approach was found to be more appropriate. Hence, participants for the grounded theory approach are recruited using theoretical sampling. Thematic analysis has been used to analyze the data of each part of the study. The findings of the second part have been further validated by quantitative methods. The study has been conducted in India. The findings of the first part of the study highlight that users have both positive and negative experience with e-government services. The study revealed that time savings, convenience, ease-of-use, corruption avoidance, fairness in the process, transparency, image, trustworthiness, and connectedness are the key factors that give improved experience of e-government services to citizens and change their behavior by avoiding visiting government offices. Citizens value the e-government services from multiple dimensions like time savings, convenience, enhancement of their image etc. Although e-government services have positive impact on citizens' attitude and behavior, citizens experience challenges like unreliable service, delay in response, website unavailability, inaccurate and insufficient information, and transaction failure with several e-government services that lead to their dissatisfaction and keep them dependent on conventional government mode of government service delivery. The study has also revealed that e-government services at different levels of government have mixed experience. Users found central government websites comparatively better than state and local governments’ websites in terms of performance and ease of use. In the context of m-government services, time-saving, convenient, ease-of-use, compatibility, procedural fairness in delivering government services, image, connectedness are the key experiences of the users that change their behavior to access public service from mobile rather than visiting government offices. Affordability, portability/mobility, quick access, easy to handle, less electricity requirement, easy to charge, long battery life, SMS based services, mobile apps, multi features, and less effort requires to learn mobile usage etc. are the key factors that leads citizens' to experience m-government services better than conventional e-government services. Citizens also sometimes have poor experience with m-government services due to small screen size, high memory usage due to several mobile apps, and poor mobile network that results in call drop and poor internet connectivity. The second part of this study reveals thirteen factors influencing indirect adoption of e-government services. Access to an intermediary, intermediaries’ service charge, risk-averse characteristics, and value-added services are novel factors revealed by this study. Lack of resources, lack of computer self-efficacy, perceived difficulty in use, and lack of multilingual option as factors for indirect adoption have been found to be contrary to the respective influencing factors like availability of resources, computer self-efficacy, perceived ease-of-use, multilingual option, which are significant for direct e-government adoption discussed in the extant literature. Remaining five factors perceived awareness, perceived functional benefits, perceived compatibility, trust of intermediary and social influence were also identified by extant studies as factors for direct adoption. The quantitative techniques found all thirteen factors significant to indirect e-government adoption. The third part of the study has considered four e-government services maturity levels: information, two-way communication, transaction, and political participation. The study reveals that factors influencing adoption of e-government services at different maturity levels differ. The study has identified thirteen, thirteen, twenty-six, and fifteen factors to be influencing the adoption of e-government services at information level, two-way communication level, transaction level, and political participation levels respectively. There are six novel identified factors which are previously unexplored in e-government adoption literature. These novel factors are auxiliary facilities, connectedness, corruption avoidance, transparency and fairness, customer support and forced adoption. The study advances the literature by answering the above research questions. The findings of this research may guide public administrators and policy-makers to develop and deliver e-government services in a manner that will foster quick adoption by citizens for fast benefit percolation.

Entrepreneurial Orientation (EO) has been a subject of more than three decades of research and has become a cornerstone of strategic management and corporate entrepreneurship literature. Currently, two prominent conceptualizations of EO exists in literature – the Miller (1983) unidimensional view of EO and the Lumpkin and Dess (1996) multidimensional view of EO. Between these two competing conceptualizations, the Miller (1983) EO concept has been studied extensively and currently forms the major corpus of extant research in EO. On the contrary, despite being present for almost two decades and garnering a large number of citations, empirical studies examining the Lumpkin and Dess (1996) notion of EO have been few. Since co-existence of both the conceptualizations would provide unique insights for EO theory, scholars have called for studies pertaining to a better comprehension of the Lumpkin and Dess (1996) formulation of EO. Simultaneously, the examination of EO phenomena in extant research has mostly been limited to matured western economies and among emerging economies, to China. Scholars suggested that significant insights for the EO theory could be gained by attending to the specific characteristics of other emerging economies like Brazil, India or Russia. Finally, both the dominant conceptualization of EO has been developed in western economies context. Scholars have questioned the valence of such theories when examining entrepreneurial phenomena in emerging economies and have insisted on developing EO conceptualizations based on careful attention to the conditions of such settings. To address these research gaps, this thesis builds on the Lumpkin and Dess’ (1996) ideation of EO and proposes a novel conceptualization of EO for emerging economies. Owing to the substitution for formal institutional mechanisms by informal institutional constraints fostered through networking, as well as recent developments in literature which suggests that networking is entrepreneurial action, we argue that EO for emerging economies includes networking as a salient dimension in addition to the original five dimensions. This thesis then undertakes three related studies to verify this novel theorization of EO empirically. While a suitable instrument for measuring the dimensions of EO was derived in the first study, the second and the third study respectively established networking as a dimension of EO and substantiated the effects of the individual dimensions of EO on firm performance in emerging economies. The first study of this dissertation derived a content adequate EO scale. Content adequacy is the most basic psychometric property that should be assessed of any measurement instrument. However, evidence for content adequacy was not found for any of the few, extant five-dimensional EO scales. Utilizing the methods developed by Schriesheim and colleagues (1993, 1999), a two-stage study was undertaken first to explore and then confirm the content adequacy of a five-dimensional EO scale. The final scale had factor structures in congruence with the theoretical dimensions of interest. Thus, the resultant scale showed fine agreement with theory and was utilized in the subsequent studies of this thesis. In the second study, three sets of hypotheses were drawn to validate networking as the sixth dimension of EO in emerging economies and to test for the enhanced effect of EO on new venture performance due to networking. While the first set of hypotheses aimed at verifying the reliability and validity of the networking construct, the second replicated that for the six-dimensional EO construct. The third set of hypotheses strived to establish that the six-dimensional EO model was a better predictor of new venture performance than the five-dimensional model. Analysis of data, collected from new ventures in India, was done using confirmatory factor analysis and structural equation modeling. The results supported the hypotheses charted out in this study and thus, furnished evidence to assert networking as the sixth dimension of EO in emerging economies. The final study supplemented the findings of the second one by undertaking a deconstructive approach to the examination of the individual dimensions of EO from a universalistic perspective. By analyzing the same set of data, using multiple linear regression, this study tested the Lumpkin and Dess (1996) theory of EO in emerging economies context and highlighted the dimensions of EO that are critical in securing superior firm performance in such settings. Simultaneously, this part also demonstrated that including networking as a dimension of EO enhances its explanatory prowess and thus, presented additional empirical evidence to justify that networking must be a dimension of EO. Through these three studies, this thesis makes contributions to the Lumpkin and Dess’ (1996) conceptualization of EO and the understanding of EO phenomena in emerging economies context. Most significantly, this thesis produces evidence that testifies networking to be the sixth dimension of EO in emerging economies. This novel conceptualization of EO for emerging economies is expected to advance the scholarly discourse towards building a conceptual family of EO constructs. Additionally, implications for researchers and practitioners are also drawn from this thesis. For researchers, the content adequate EO scale is expected to be instrumental in promoting further research in the sparsely studied five-dimensional EO research domain. For entrepreneurs and managers, the findings add relevance for implementing the networking concept and suggesting the key dimensions of EO for success in emerging economies. This thesis thus serves to build a more comprehensive theory of EO.

Even after the introduction of various employment legislations over the past fifty years and the growing popularity of diversity management programs, multiple forms of gender bias continue to exist in the workplace. Gender bias towards women in the workplace negatively affects women’s opportunities and earnings, which are manifested in the form of gender pay gap, the underrepresentation of women in leadership positions and the length of time required for career advancement to name a few. These disadvantages result in lower job satisfaction, lower organizational commitment and reduced performance among women leading to stress and health issues. India mirrors the same global situation of women representation in senior positions in organizations with women choosing to leave the workforce due to institutional and personal gender bias. Gender bias against women can occur during the process of recruitment and selection, pay, performance evaluations and promotion. Subtle gender bias also occurs in the way men and women are appraised in the workplace. Women are more frequently disturbed during work and their contributions are generally ignored. This continued negative bias indicates that organization and government policies have not succeeded in mitigating gender bias. This suggests adoption of different measures to tackle the issue. Formulating policies with the right perspective is important because it contributes to the social construction of individual’s values and preferences which in turn also impacts the creation and execution of policies. Thus, the attitude of an individual towards gender roles, appears to be an underlying factor perpetuating gender bias. Though the societal and organizational structure exacerbates gender discriminatory behaviour, it is the individual attitude that collectively transforms a neutral structure into an inhospitable environment for women. In the last few decades, one of the biggest determinants of perceptual change related with traditional gender roles has been the participation of women in the workforce. Another important aspect is that, although women's aspirations regarding their career have increased, they have not compromised on motherhood. In consequence, we expect to find that attitudes towards these roles have different orientations, assuming that people’s beliefs are highly correlated with their behaviours. Further, the way these attitudes correlate with the new distribution of gender roles remains unclear and will be empirically explored in this study. The aim of this study is also to see whether the specific support given by government and organizations through special policy measures have encouraged men to be more egalitarian in their attitudes towards women, or otherwise. For this, the present study explores the explicit and implicit attitudes towards gender by using different mechanisms to explore and compare the outcomes. The main contribution of the study in understanding gender attitudes consists of developing an implicit test in sync with the Indian culture and using neuroscience tools to further enhance and support the findings. The study has been conducted in two phases with the first phase being studied in Indian Public Sector organizations. The choice of the organizations was essentially based on the fact that individuals within the organization had awareness of gender issues and the organisations were large enough to enable random sampling across departments and hierarchy for selecting the required number of participants. The second phase of the study was conducted in a neuroscience lab to gain a deeper insight into mechanism of bias. In both the phases, the gender attitude of individuals was assessed using different attitude measures. In the first phase, the explicit and implicit measurement techniques were used. In the second phase, implicit measure and cognitive neurosciences tool was used to assess the gender attitude. The novel contribution of the study includes developing an implicit measure to suit the Indian context. The Implicit Association Test (IAT) used extensively to understand subtle attitudes in the west was re-modelled and used for to measure implicit attitudes in the Indian workplace. Further neuroscience techniques were used to develop a better understanding of attitude functioning. The study employed ANOVA and t-tests to understand the difference in gender attitudes among individuals at different levels of gender and work experience within the organization. The divergence in findings among explicit, implicit and neuroscience measures indicate the importance of using all the three measurement techniques for better understanding of gender attitudes The study applied the tripartite theory of attitude, attitude models and dual process theory to develop an understanding of the gender attitude in the workplace. The study attempted to address the gap in literature by instituting the need for studying implicit gender attitude in the Indian workplace for a holistic understanding of gender biased attitudes and its direct and indirect effect on individual behaviour The study of gender attitudes in workplace has critical ramifications for an individual, organization and society as a whole. The result of this study will provide managers with useful information when developing gender equitable policies as the results will indicate whether using individual’s explicit attitudes as an indicator of the true gender attitude might be an inadequate measure. The study has methodological implications as it tests the applicability of the suggested measurement technique (Implicit Association Test) within an organization, simultaneously as it explores socially desirable responding through existing explicit measures. The findings will have implications for the way policies are instituted and gender equitable attitudes practised by individuals as it questions the reliance on existing explicit measures as a predictor for behaviour.

Past three decades have witnessed a significant transformation in the global banking scenario due to restructuring of the financial sector. These reforms have resulted in deregulation of interest rates, liberalization of credit control, privatization of financial institutions and removal of entry barriers in the capital markets (Beju & Ciupac-Ulicii, 2012). These reforms have emphasized on developing robust risk frameworks to counter the new forms of risks (such as exchange-rate risk, interest-rate risk) that have crept into the banking system. The Basel Committee on Banking Supervision (BCBS) in Basel II accord describes ‘concentration risk’ as one such new form of risk. In the context of financial institutions, concentration risk entails the risk arising from a narrow distribution of credit borrowers, business associations, focus in terms of geographical boundaries and business segments, which in turn produces inefficient risk-return tradeoffs, and put at risk the solvency of the institution (Acharya, Hasan, & Saunders, 2006). This new form of risk has assumed attention of researchers, managers and regulators who recognize the key role of concentration risk in determining bank’s performance. George and Singla (2015) suggested two principal strategies for managing risk: risk hedging and diversification. Hedging entirely removes an exposure, eliminating both upside and downside potential by using derivatives and structured products (Smith, 1995). Diversification, on the other hand, can mitigate certain kind of risk, by holding a large collection of imperfectly correlated assets (Stulz, 1996). As per the theory, hedging is considered as one time activity and considered to be expensive as it has a premium attached to it; whereas diversification has been considered as inexpensive; moreover, it is considered as a structural response to financial risk management. Therefore, diversification soon became the new innovative technique for handling risk management in banks. Traditionally banks have been focused on risk management using credit analysis which primarily focuses on high-frequency low-severity risk. We argue that this approach though useful in the micro management of risk does not provide an affordable response to low-frequency high-severity risk which is a major cause of recent bank failures (Ivashina & Scharfstein, 2010). Thus it is particularly important to manage the later as banks are getting exposed to new form of opportunities and the resulting risks. Through this study, we extend the discussion on management of low-frequency high-severity risk using diversification as a structural measure. This study proposes that diversification would lead to firm performance in two significant ways. One, it will reduce the unsystematic risk by first identifying efficient bases for diversification and then by distributing this risk through an efficient distribution of assets/businesses on each of the bases. Two, for banks with suboptimal distribution of resources on a marginal return basis, an efficient diversification shall improve the return of the bank. As a result, banks will benefit either from a reduction in the overall risk or an increased returns or both which will indicate an improvement in the firm performance. In this direction, the three most common bases of diversification noted in the literature and observed in practice are income, sectoral and geographical. This study examines the role of these diversification bases in reducing risk and/or improving returns of the banks aiding overall performance. In the process the study also examines efficiency of each of the considered diversification bases in improving the firm performance. Various approaches have been applied till date to examine the relationship between a bank’s performance and diversification which have contradicting findings. However, the existent literature on this subject have investigated the impact of diversification either on risk or on a bank’s profitability; this dissertation looks to examine the overall impact of diversification taking into account both risk and return perspectives of a bank’s performance. The study also analyzes the role of bank ownership, asset size and age on the above association. Further, this dissertation comprehensively examines diversification on three different bases: sector, income and geography and assesses their independent impact on a bank’s performance. This is in contrast with previous studies where generally, the single basis of diversification is discussed. The institutional framework, macroeconomic conditions and the political environment in developing nations are distinct from developed nations. Therefore, this study sampled annual data of Indian banks, representing developing economies, whereas most of the earlier studies were based on developed economies. Thus, taking into account the importance of concentration risk in determining a bank's performance, this study examines the impact of various diversification strategies adopted by banks to mitigate concentration risk and thereby improve its performance. The study raises a few critical concerns regarding the impact of bank diversification on performance: first, whether banks realize the importance of concentration risk and in turn, they start diversifying themselves? Second, whether adoption of different bases of diversification is effective for risk management or not? Third, do bank specific characteristics like ownership, size and age affect the effectiveness of diversification? The primary objective of the study is to examine the impact of Sectoral, Income and Geographical diversification on the banks’ performance. The study further examines the role of bank ownership, asset size and age on the relationship between diversification and banks’ performance. This dissertation focuses on the impact of diversification on bank’s return and risk. It consists of four essays: First essay discusses in length the various risk management practices adopted by financial institutions and the rationale behind using diversification as a structural response mechanism to the management of concentration risk. It then focuses on the alternate bases of diversification used by the banks to improve their risk management framework. The subsequent three essays analyse the impact of income, sectoral and geographical diversification respectively on the performance of banks in India. Annual data of Indian banks are sampled; a series of regressions are operated with performance measures as a dependent variable and diversification measures as an explanatory variable, after controlling for a set of firm specific variables. For an unbalanced dataset, a distributed lag model is employed in order to estimate a bank’s performance, using Fixed Effect Models (FEM) based on firm and time fixed effects. Extending the findings in this field, our result suggests that in general, sectoral diversification positively affects the bank profitability by improving risk adjusted returns and reduces risk of Indian banks, which is in line with the studies, Bebczuk & Galindo, 2008 (US markets); Elsas et al., 2010 (global); Williams & Prather, 2010 (Australia). Results are consistent across all ownership types’, i.e., public, private or foreign, indicating existence of an efficient monitoring and control systems which seems to have resulted in a better asset quality management. Income diversification is also found to be associated with increasing bank returns and diminishing risk across all ownership categories. Although, for private and foreign banks, income diversification helps in reducing risk, for public banks, no significant impact in risk reduction is observed. This indicates a better coordination and control across different business lines both in private and foreign banks but in public banks there is a dearth of synchronization across different departments which bring new kind of risk into the system. Geographical diversification helps in increasing bank returns for the overall sample but does not have any significant impact on bank risk. For foreign and public banks, geographical diversification helps in increasing bank returns but does not have any significant impact on bank risk. Interestingly, geographical diversification results for private banks shows a decline in bank returns, which is indicative of the severity of the agency problem. The results indicate that diversification as a risk mitigating tool can be effective, but the regulators should not emphasize on the ‘one size fit all’ approach for all kind of banks. Therefore, the policy frameworks for controlling concentration risk should be developed keeping in mind bank specific factors like bank size, customer base and core competencies of managers. This study has important implications for managers and policy makers. The policy makers should work towards developing banks risk management frameworks which cater to both low-frequency high-severity and high-frequency low-severity risks. Therefore, guidelines like BASEL should be developed to cater importance to both kinds of risks. In Indian banks, use of diversification as a risk mitigating tool has increased considerably over the past 15 years, which should be closely monitored by the managers and policy makers. Our results suggest that the extent of diversification in Indian banks is less as compared to the developed nations. So Indian managers should learn from their mistakes and do not over diversify their resources because studies suggest that after a certain extent the benefits of diversification starts diminishing as diseconomies creep into the system.

With the rapid rise in popularity of e-commerce applications and the dramatic increase in the size of data present in such applications along with the related social media data generated by customers, information filtering technique like Recommender Systems are widely used to help the customer in finding the item they might find useful. This prediction also helps the companies in cross-selling, upselling and to increase the loyalty of their customers. Collaborative Filtering algorithms, the most widely used technique in Recommender Systems, are primarily based on the phenomenon of homophily which means that humans tend to associate and bond with other humans who are similar to them. This also gets reflected in their preferences for products and the feedback they give to them. The similarity between users in Collaborative Filtering approach is primarily measured through the inverse of the distance between their rating vectors. However, due to lack of sufficient amount of relevant ratings data despite having vast numbers of users and products in the system, there exists a sparsity problem which impacts the quality of recommendations. There also exist few users whose opinions do not consistently agree or disagree with any group of people. Hence such users don’t benefit much from the system. Moreover, their presence also reduces the overall efficiency of the system impacting the quality of recommendations made to other users. This is referred to as Gray Sheep problem, and their identification is an important area of research in Recommender systems. For increasing the efficiency of recommendations and cross-selling, Cross Domain Recommender Systems are used, which use the knowledge from domain to make recommendations in different domain. With the increasing capability to capture more and more contextual data about the users, Context Aware Recommender Systems are also gaining popularity, particularly to make real-time recommendations. In this thesis work, five research gaps have been identified which are present in the existing research literature related to Recommender Systems. They are primarily related to lack of any work done to study the Gray Sheep behavior in different domains and contexts. Since Gray Sheep behavior is a reflection of a user’s eccentric taste, we have also explored research related to the usage of a user’s personality data in an attempt to resolve Gray Sheep problem and also sparsity problem through Transfer Learning approach. Transfer Learning is a machine learning paradigm that makes use of knowledge learned in one task in a different but related source domain, to solve the task in other target domain. This is particularly useful when there is a scarcity of fewer high-quality training data. For Transfer Learning, auxiliary data sources used in this work comprise of users’ social media data and their ratings in other domains. Users’ characteristics are learned using the latest advancement in natural language processing which makes this approach more practical and scalable. Based on our observation about Gray Sheep users, we also propose the concept of Gray Sheep Items, in scenarios where item-based collaborative filtering is used. We have also empirically verified their existence using various publicly available dataset.

Many authors in the fields of strategic management, organizational theory and organizational behavior have suggested that long-term organization performance depends on the organization’s ability to simultaneously exploit current capabilities and explore new opportunities/capabilities. Organizational researchers have utilized the concept of ambidexterity “…the ability of humans to use both hands with equal skill, as a metaphor for organizations that are equally dexterous at exploiting and exploring” (Simsek, 2009, p. 597). Organizational ambidexterity refers to “an organization’s ability to be aligned and efficient in its management of today’s business demands while simultaneously being adaptive to changes in the environment” (Raisch & Birkinshaw, 2008, p. 375). March (1991) simulated the interplay between distinct sets of organizational activities (exploration, and exploitation) and highlighted inherent incompatibilities between the two using an organizational learning approach. Since then, significant research in the field has uncovered three distinct types of organizational ambidexterity: structural, temporal and contextual. These types focus on the bifurcation of exploration and exploitation activities across organizational structures, time and contexts respectively. Although previous works in the field have highlighted various aspects of organizational ambidexterity, recent ambidexterity studies have also pointed at the lack of integrative models spanning multiple levels of analysis (for example, see Birkinshaw and Gupta, 2013; Simsek, 2009) in the field and call for more research to address the gap. The scarcity of ambidexterity research incorporating multiple levels of analysis, in the same study, using the available conceptual lenses led us to explore nested dynamics of organizational ambidexterity within the new product introduction initiatives at two large automotive firms in India. The study utilized two in-depth multi-level case studies to highlight the ambidexterity dynamics across strategic, business-unit and functional levels using a semi-structured guide. The respondents were scattered across various geographical locations in India. Thematic analysis was used to analyze the interview responses. It was found that the existing mechanisms of understanding ambidexterity at play do not offer full explanatory power in the sample. On the other hand, the visualization of the new product introduction efforts in the sample firms as metaroutines offered satisfactory explanatory power. It is found that certain ostensive (like lean processes and focus on commonality) and performative aspects (like creative recombination) of the new product introduction efforts of sample firms, visualized as metaroutines, aid in promoting multi-level ambidexterity. Organizational routines are defined as “repetitive patterns of interdependent organizational actions.” (Parmigiani & Howard-Grenville, 2011, p. 417) involving multiple actors. The routines literature suggests three kinds of routines: metaroutines (higher order routines bringing about changes in or creating new lower level organizational routines), decision-making routines (routines aiding in strategic decisions like investment, pricing etc.) and operating routines (standardized procedures for day to day activities). Previous work on routines suggest two aspects of routines: ostensive (ideal/schematic) and performative (actual enactment). Further, extant research identifies process, structure and governance mechanisms coordinating multiple actors as micro-foundations of routines. Although some researchers have pointed at a relationship between routines (including metaroutines) and organizational ambidexterity, studies detailing the various interactions of routines (and specifically metaroutines) with multi-level ambidexterity have been largely absent in the literature. Further, the study argues that metaroutine enabled approaches to ambidexterity offer a distinct fourth form when compared with other academically established forms of structural, contextual and temporal ambidexterity. Using a data/text mining based systematic literature review, this thesis also identifies the current boundaries of research in the field of ambidexterity. Analyzing abstracts of 504 works on ambidexterity (in the last two decades), eight themes are inductively derived to indicate the current boundaries of the field. The analysis in the study suggests that although the role of processes in organizational ambidexterity has been studied, focus on routines in ambidexterity literature is very low. The field of organizational ambidexterity is rooted in the organizational learning literature. To further understand the impact of the identified embedded theme of commonality, this study implements a stochastic probability simulation extending the base model used by March (1991). It is found that in the case of low socialization rates of individuals within the organization and high organizational learning rate from the individuals, commonality has largely no benefits for organizational knowledge. However, in other combinations of socialization rates and organizational learning rates benefits of commonality are evident. In summary, the study attempts to add to the organizational ambidexterity literature by delving into the dynamics of how metaroutines impact multi-level ambidexterity. Metaroutine enabled ambidexterity is offered as a distinct and novel type of ambidexterity. Centralized, firm level metaroutines like the new product introduction metaroutines (found in the sample firms) are uniquely positioned to handle multi-level ambidexterity dynamics. Metaroutines, due to certain embedded aspects (discussed in the thesis), can affect the nature of ambidexterity questions across levels while coordinating the various levels towards common organizational goals. One such aspect, i.e. commonality has been explored in further detail. Although the thesis encapsulates how metaroutines enable multi-level ambidexterity in a new product development efforts of automotive firms, the thesis cannot generalize the qualitative findings beyond the context and industry studied.

Corporate finance deals with capital structure decisions, capital budgeting decisions, dividends theory or firm valuation decisions and working capital management decisions. Working capital management (WCM) is one of the crucial decisions to be dealt by managers as it impacts the liquidity and profitability of the firm (Appuhami, 2008). Working capital (WC) as the excess of current assets over the current liabilities and is the lifeblood of a business enterprise and if not managed properly, it may result in the failure of the business (Padachi et al., 2012). This thesis has three research objectives. Existing literature assumes an indirect relationship between working capital efficiency and profitability of the firms. Hence, if one needs to consider only that part of profit which is generated by efficient working capital management (WCM); then there is a need for a new measure which can describe an explicit and direct relationship between working capital efficiency and financial performance of the firms. The first research objective of this thesis introduces Working Capital Efficiency Multiplier (WCEM) as a direct financial performance (outcome) measure of working capital efficiency which can be widely used by academicians and practitioners. The second research objective explores the previous studies on the relationship between working capital management (WCM) and firm profitability and indicates a linear relationship. However, the studies have ignored the higher risk of loss of sales and interruptions in the production process that is related to the low levels of working capital. There might be a working capital level at which reduction in working capital negatively affects a firm’s profitability. The relation between working capital and a firm’s profitability may be concave rather than linear and might be better captured by a quadratic relationship. Hence, the second research study attempts to find if the listed Indian non-financial firms have optimality in their investments in working capital with respect to their profitability. Finally, the third research objective opines from the extant literature that the Managerial Efficiency - Working Capital Management (WCM) is the pre-requisites for the survival of a firm and significant element of an organization’s success. Level of holding cash is a managerial discretion and holding excess cash raises concerns and the conflicting interest between mangers and shareholders could be minimized by establishing an effective corporate governance mechanism. Few studies have examined the impact of corporate governance on the working capital management (working capital efficiency aspect) of firms in developed countries. Given this backdrop, the third research study of this thesis studies the impact of corporate governance on the liquidity dimension of the working capital management of the firms operating in emerging economies like India.

Natural language processing (NLP) deals with analyzing, understanding and generating the languages which humans use to communicate naturally. This technique has come into prominence from early 2000s. NLP has become one of the major techniques for analysis of User Generated Content (UGC) as it provides a distinct advantage in text analysis when compared to conventional data mining techniques. The present proposal uses NLP based techniques for social media analytics, the findings from which could be extended to other domains, including IT incident management. Analyzing user generated content in case of microblogs is a challenging problem. This is because of various reasons, such as, informal language, smileys, word limit (140 characters for twitter) and deliberate as well as genuine spelling mistakes. This kind of unstructured data needs to be cleaned and accumulated in larger amount before pattern identification and knowledge creation. One of the interesting problems in microblog analysis is identifying user demography through tweets. It has multiple social as well as business implications. From a social perspective, demographic knowledge could provide information, such as, how different genders behave online, views of a certain age group, opinions of people belonging to a certain region etc. From a business perspective, opinions about products/services based on gender, age and region could be crucial markers from a firm’s perspective. Surprisingly, all this information could be acquired from UGC (tweets and microblogs in this case). The first problem, this proposal deals with is predicting demographic features using NLP, specifically, gender of a Twitter user. A supervised learning method is developed which predicts the gender of a tweeter by capturing unique linguistic features with an accuracy of above 70%, which is considerably better than commercially available gender detection software. Further, the distinguishing text features of both the genders are identified, based on the probability of them belonging to one of the classes. The demographic information which can be obtained using the above technique gives one a good starting point for knowledge generation using UGC, however, this might not be enough in a highly competitive business environment where customer behavior can be highly unpredictable. A seemingly innocuous sentence which we have come across in our dataset is, “I love this book’s cover #neverjudgeabookbyitscover”, looks positive, but on a closer look it becomes clear that it actually uses irony to represent the user’s unhappiness with the book. One of the major examples of this could be, “faking news communities” which have a humorous take on day to day events and clearly propagate false information for humor’s sake. The pace with which the online community is evolving, one could safely assume that the online communication by users is going to be more complex than ever in future. Hence, the second problem this proposal tries to deal with is this complexity in online communication, by identifying irony or sarcasm in microblog sentences. It is to be done by extracting authorial style based linguistic features in tweets. Finally, the knowledge generated through the above two works is to be applied in an IT infrastructure support scenario, specifically, in an attempt to analyze the incident management process. Incident management process is a crucial part of IT business support where micro-text plays a crucial role.

Internet today has surrounded the users with its symbolic problem of ‘information overload’ where a consumer has far more than required choices. In the realm of e-commerce industry, recommender systems (RS) utilize the information filtering technology to proactively handle ‘information overload’, the ubiquitous problem in cyberspace. Out of all known recommendation techniques, collaborative filtering (CF) remains the most successful and widely accepted method. The principle behind CF is the automated form of word-of-mouth communication between like-minded people. These like-minded users have similar tastes in terms of rating preferences for a common set of items in the past. CF framework broadly has two components, viz., neighbour identification and rating prediction. The neighbourhood formation process or the process of neighbour (recommendation partner) identification is central in a CF mechanism. Traditionally, only those neighbours are selected as recommendation partners who are most similar to the given user and have rated the item under consideration. In this thesis, the shortcoming in the above described neighbourhood formation process is highlighted, where the selected neighbours are weak as the most similar users in terms of similarity strength have not necessarily rated the item under consideration. In order to select neighbours by the existing practices, their similarity deteriorates as many closer neighbours might not have rated the item under consideration. Consequently, the selected neighbours are weak in strength which leads to poor neighbourhood formation and, in turn, affects the overall recommendation quality. Thus, a new approach for neighbourhood formation is proposed here, where the neighbours are selected as recommendation partners only on the basis of highest similarity magnitude, irrespective of rating the item under consideration. The above shortcoming in the neighbourhood formation process has never been discussed in the extant literature, hence, making this work first-of-its-kind. By virtue of the proposed approach for the process of neighbour identification, there will be missing ratings for the active item by many strong neighbours. To solve this problem arising primarily due to data sparsity, regression based techniques are used to predict these missing ratings. The proposed approach is based on the motivation that in the presence of meaningful recommendation partners, i.e., more similar neighbours, reliance should not be made on weak neighbours. The objective of this thesis is to improve the recommendation quality by identifying meaningful recommendation partners. Further, this thesis aims to provide improved recommendations to address the cold-start user problem, i.e., users who are relatively new or inactive in the system. Experiments performed using different publicly available datasets establish the efficacy of the proposed approach in improving the overall recommendation quality.