Design for Social Sharing: The Case of Mobile Apps

2023 | Information Systems Research | Citations: 1

Authors: Karanam, Subrahmanyam Aditya; Agarwal, Ashish; Barua, Anitesh

Abstract: With a large number of mobile apps on platforms such as iOS and Android, app dev ... Expand

Abstract: With a large number of mobile apps on platforms such as iOS and Android, app developers face a significant challenge in generating market demand. Apps can incorporate social features to share information and create awareness. We focus on the impact of these features and their interplay with intrinsic features in the head, body, and tail of the demand distribution. Using a panel of version release notes from the iOS platform, we develop a novel hierarchical deep learning model to extract intrinsic and social features. Our results suggest that social features help increase the demand for tail apps and are also useful for head apps in informing users about new intrinsic features. To explore possible mechanisms, we analyze how different types of social features (personal and platform), intrinsic features (common versus differentiating), and app quality influence demand. We find that social features that allow sharing on platforms with large audiences are more effective at increasing demand than those on personal messaging systems. Furthermore, within platforms, social features perform better on those with stronger ties, such as Facebook as opposed to Twitter. Our analysis of different types of intrinsic features reveals that social features can help increase the demand for all apps when introduced with differentiating or less common intrinsic features. Furthermore, we find that there is a negative effect on the demand of tail apps when differentiating intrinsic features are combined with platform-based social features that have weaker ties. However, this negative effect is limited to low-quality tail apps only. Our results underscore the differences in the effect of different types of social and intrinsic features in various parts of the demand distribution. Our study provides managerial guidance to app developers in enabling social sharing through design choices and generating higher demand.History: Kai Lung Hui, Senior Editor; Beibei Li, Associate Editor.Supplemental Material: The online appendix is available at https://doi.org/10.1287/isre.2022.1151. Collapse

Topics: mobile application Facebook Twitter iOS accounting

Methods: deep learning robustness check computational algorithm word embedding natural language processing

Enhancing Recommendation Interpretability with Tags: A Neural Variational Model

2022 | International Conference on Information Systems | Citations: 0

Authors: Liu, Shen; Che, Shangkun; Liu, Hongyan

Abstract: Recommender systems are widely used for assisting consumers finding interested ... Expand

Abstract: Recommender systems are widely used for assisting consumers finding interested products, and providing suitable explanations for recommendation is particularly important for enhancing consumers' trust and satisfaction with the system. Tags can be used to annotate different types of items, yet their potential for providing interpretability is not well studied previously. Therefore, it is worthy to study how to leverage tags to enhance recommendation systems in terms of both interpretability and accuracy. This paper proposes a novel model that seamlessly fuse topic model and recommendation model, where the topic model can analyze tags to infer understandable topics, and the recommendation model can conduct accurate and interpretable recommendations based on these topics. We develop variational auto-encoding method to take advantage of neural networks to infer model parameters. Experiments on real-world datasets illustrate that the proposed method can not only achieve great recommendation performance, but also provide interpretability for the recommendation results. Collapse

Topics: mobile application recommender system digital distribution online community mobile application market

Methods: artificial neural network topic model experiment sensitivity analysis literature study

ML-Based Product Design: The Case of Mobile Apps

2021 | International Conference on Information Systems | Citations: 0

Authors: Karanam, Aditya; Agarwal, Ashish; Barua, Anitesh

Abstract: Firms strive to continuously improve their products to compete effectively in t ... Expand

Abstract: Firms strive to continuously improve their products to compete effectively in the market. Typically, firms update their products by adding novel (differentiating) features or by imitating their competitors. With the advent of social media, there is also the possibility of obtaining customer input on their desired features. Customers post reviews, which include suggestions for improvement. In the case of mobile apps, user feedback may include suggestions of novel features or features that are already present in competing apps. Leveraging the information on reviews and version release notes of iOS apps, we develop a deep learning algorithm to identify four types of app features -developer initiated innovative, developer initiated imitative, user suggested innovative and user suggested imitative, and evaluate their impact on the demand. Our results suggest that developer initiated innovative and user suggested imitative features help increase the demand, however, user suggested innovative features have a negative effect. Further, this negative effect is limited to features that are contextually "distant" from user suggestions. Contextually "close" innovative features suggested by users do have a positive effect on the demand. To the best of our knowledge, this is the first study to consider reviews as a source of user ideas and evaluate the performance impact of the ideas mentioned in the reviews. Our study also provides a novel deep learning model to extract suggested features from user generated content. Collapse

Topics: mobile application innovation management entrepreneurship crowdsourcing user requirement

Methods: deep learning BERT word embedding computational algorithm artificial neural network

LINDA-BN: An interpretable probabilistic approach for demystifying black-box predictive models

2021 | Decision Support Systems | Citations: 0

Authors: Moreira, Catarina; Chou, Yu-Liang; Velmurugan, Mythreyi; Ouyang, Chun; Sindhgatta, Renuka; Bruza, Peter

Abstract: The use of sophisticated machine learning models for critical decision-making fa ... Expand

Abstract: The use of sophisticated machine learning models for critical decision-making faces the challenge that these models are often applied as a ‘black-box’. This has led to an increased interest in interpretable machine learning, where post-hoc model-agnostic algorithms present a useful mechanism for generating interpretations of complex learning models. This paper proposes a novel approach based on Bayesian Networks to generate local post-hoc model-agnostic interpretations of a black-box predictive model. Consequently, the proposed approach presents features that are conditionally dependent between each other and that are directly influencing the class variable. This enables the decision-maker to better understand how features are related and why a certain prediction was made. Compared to the existing post-hoc interpretation methods, the contribution of our approach is three-fold: (1) as a probabilistic graphical model, the extracted Bayesian network can provide interpretations through conditional dependencies in a graphical structure regarding what input features and how/why they contributed to a prediction; (2) for complex decision problems with many features, a Markov blanket can be generated from the extracted Bayesian network to provide interpretations with a focused view on those input features that directly contributed to a prediction; (3) the extracted Bayesian network enables the identification of four different rules which can inform the decision-maker about the confidence level in a prediction, thus helping the decision-maker assess the reliability of predictions learned by a black-box model. We implemented the proposed approach, applied it in the context of two well-known public datasets and analysed the results, which are made available in an open-source repository: https://github.com/catarina-moreira/LINDA_DSS. Collapse

Topics: shapley additive explanation local interpretable model agnostic explanation explainable artificial intelligence

Methods: Bayesian network computational algorithm experiment decision tree classification deep learning

Theories: game theory

A probabilistic Bayesian inference model to investigate injury severity in automobile crashes

2021 | Decision Support Systems | Citations: 2

Authors: Topuz, Kazim; Delen, Dursun

Abstract: Big data analytics examines millions, if not billions of records, to unmask hidd ... Expand

Abstract: Big data analytics examines millions, if not billions of records, to unmask hidden patterns, provide actionable insights and interpretable results for various domains. One area that has great potential to leverage the value of big data and analytics is the critical analysis of traffic accidents. Investigation results help in providing an indepth understanding of the risks and provide measures to potentially prevent these risk factors hence enhancing the well-being of individuals who may experience such accidents. This study explains existing models and proposes a data science methodology in a field where probabilistic modeling makes much sense for faster, better decision-making. The main objective of this data analytics study is to identify the high-risk factors with their apparent significance to influence the probability of injury severity on automobile crashes using a geographically representative car crash dataset. To obtain reliable, accurate, and intuitive results, a multi-step probabilistic inference model based on Bayesian Belief Network— highly-acclaimed machine learning method­ ology—is proposed. The underlying inference model provides researchers with a causally accurate way to explore the domain (with the subject matter expert inputs) while disengaging issues related to statistical cor­ relations and causal effects. In this study, we also used the data to create a web-based probabilistic inference simulator, a Bayesian inference decision support tool, which will be a publicly available/accessible tool, to help decision-makers better understand and to conduct what-if analysis on variable interdependencies. Collapse

Topics: missing data Python decision support system decision making database system

Methods: Bayesian analysis machine learning probabilistic graphical model simulation Bayesian network

Theories: information theory

Measuring Brand Favorability Using Large-Scale Social Media Data

2021 | Information Systems Research | Citations: 0

Authors: Zhang, Kunpeng; Moe, Wendy

Abstract: For decades, brand managers have monitored brand health with the use of consumer ... Expand

Abstract: For decades, brand managers have monitored brand health with the use of consumer surveys, which have been refined to address issues related to sampling bias, response bias, leading questions, etc. However, with the advance of Web 2.0 and the internet, consumers have turned to social media to express their opinions on a variety of topics and, subsequently, have generated an extremely large amount of interaction data with brands. Analyzing these publicly available data to measure brand health has attracted great research attention. In this study, we focus on developing a method to measure brand favorability while accounting for the measure biases exhibited by social media posters. Specifically, we propose a probabilistic graphical model–based collective inference framework and implement a block-based Markov chain Monte Carlo sampling technique to obtain an adjusted brand favorability measure that is correlated with traditional survey-based measures used by brands. For analysis, we collect and examine Facebook data for more than 3,300 brands and about 205 million unique users that interact with those brands via their Facebook brand pages. Our data set is large and contains 6.68 billion likes and full text for 1.01 billion user comments, creating challenges for any modeling efforts. We evaluate the effectiveness of our model via out-of-sample prediction, external ground truth testing, and simulation. All demonstrate that our model performs very well, providing brand managers with a new method to more accurately measure consumer opinions toward the brand using social media data. Collapse

Topics: social media Facebook financial performance marketing management Google+

Methods: computational algorithm simulation survey machine learning sentiment analysis

Social media analytics: Extracting and visualizing Hilton hotel ratings and reviews from TripAdvisor

2019 | International Journal of Information Management | Citations: 8

Authors: Chang, Yung-Chun; Ku, Chih-Hao; Chen, Chun-Hung

Abstract: Analyzing and extracting insights from user-generated data has become a topic of ... Expand

Abstract: Analyzing and extracting insights from user-generated data has become a topic of interest among businesses and research groups because such data contains valuable information, e.g., consumers’ opinions, ratings, and recommendations of products and services. However, the true value of social media data is rarely discovered due to overloaded information. Existing literature in analyzing online hotel reviews mainly focuses on a single data resource, lexicon, and analysis method and rarely provides marketing insights and decision-making information to improve business’ service and quality of products. We propose an integrated framework which includes a data crawler, data preprocessing, sentiment-sensitive tree construction, convolution tree kernel classification, aspect extraction and category detection, and visual analytics to gain insights into hotel ratings and reviews. The empirical findings show that our proposed approach outperforms baseline algorithms as well as well-known sentiment classification methods, and achieves high precision (0.95) and recall (0.96). The visual analytics results reveal that Business travelers tend to give lower ratings, while Couples tend to give higher ratings. In general, users tend to rate lowest in July and highest in December. The Business travelers more frequently use negative keywords, such as “rude,” “terrible,” “horrible,” “broken,” and “dirty,” to express their dissatisfied emotions toward their hotel stays in July. Collapse

Topics: online review visual analytics Google+ cloud computing Tableau software

Methods: product review exploratory data analysis support vector machine sentiment analysis machine learning

Predicting Security Events in Cloud Computing

2018 | Americas Conference on Information Systems | Citations: 0

Authors: Singh, Vivek; Dutta, Kaushik

Abstract: Organizations migrate to the cloud computing environment to achieve cost reducti ... Expand

Abstract: Organizations migrate to the cloud computing environment to achieve cost reduction, business agility, and elasticity. However, cloud migration brings another set of challenges such as data security & privacy, resource management, and compliance. Recently, cloud vendors have started offering services to their customers to collect detailed logs of events and resource utilization. However, the cloud computing ecosystem lacks frameworks, models, and IT artifacts to analyze such logs to draw business insights and address the above challenges. Following Design science research paradigm, we present a Gaussian Bayesian Network based approach for learning the underlying dependencies among the events in cloud services to determine the antecedents and consequences of security related events. Moreover, using our model, we predict the security related events with average mean error of 0.13 events for one day ahead forecast. We further discuss our research implications for software development, security, and IT audit in cloud computing environment. Collapse

Topics: cloud computing IT security defense software developer systems development IT skill

Methods: Bayesian network computational algorithm descriptive statistic design science probabilistic graphical model

Predicting graft survival among kidney transplant recipients: A Bayesian decision support model

2018 | Decision Support Systems | Citations: 10

Authors: Topuz, Kazim; Zengul, Ferhat D.; Dag, Ali; Almehmi, Ammar; Yildirim, Mehmet Bayram

Abstract: Predicting the graft survival for kidney transplantation is a high stakes undert ... Expand

Abstract: Predicting the graft survival for kidney transplantation is a high stakes undertaking considering the shortage of available organs and the utilization of healthcare resources. The strength of any predictive model depends on the selection of proper predictors. However, despite improvements in acute rejection management and short-term graft survival, the accurate prediction of kidney transplant outcomes remains suboptimal. Among other approaches, machine-learning techniques have the potential to offer solutions to this prediction problem in kidney transplantation. This study offers a novel methodological solution to this prediction problem by: (a) analyzing the retrospective database including >31,000 U.S. patients; (b) introducing a comprehensive feature selection framework that accounts for medical literature, data analytics methods and elastic net (EN) regression (c) using sensitivity analyses and information fusion to evaluate and combine features from several machine learning approaches (i.e., support vector machines (SVM), artificial neural networks (ANN), and Bootstrap Forest (BF)); (d) constructing several different scenarios by merging different sets of features that are optioned through these fused data mining models and statistical models in addition to expert knowledge; and (e) using best performing sets in Bayesian belief network (BBN) algorithm to identify non-linear relationships and the interactions between explanatory factors and risk levels for kidney graft survival. The results showed that the predictor set obtained through fused data mining model and literature review outperformed the all other alternative predictors sets with the scores of 0.602, 0.684, 0.495 for F-Measure, Average Accuracy, and G-Mean, respectively. Overall, our findings provide novel insights about risk prediction that could potentially help in improving the outcome of kidney transplants. This methodology can also be applied to other similar transplant data sets. Collapse

Topics: analytical information system decision making missing data database system procurement management

Methods: machine learning support vector machine computational algorithm naïve bayes literature study

A Customized and Interpretable Deep Neural Network for High-Dimensional Business Data - Evidence from an E-Commerce Application

2017 | International Conference on Information Systems | Citations: 0

Authors: Urbanke, Patrick; Uhlig, Alexander; Kranz, Johann

Abstract: Extracting actionable information from complex data is a key challenge for busin ... Expand

Abstract: Extracting actionable information from complex data is a key challenge for business analytics researchers (Hedgebeth, 2007). This is particularly difficult for high-dimensional datasets, to which an increasing number of businesses have access (Martens et al., 2016). In this study, we develop a customized neural network for extracting interpretable features from very highdimensional datasets. These features can be interpreted both at an aggregated as well as a very fine-grained level. Interpreting non-linear interactions is almost as easy as interpreting a linear regression. We apply the algorithm to a dataset related to product returns in online retail which contains a total of 3,637,654 transactions and 13,533 dimensions. Comparing 85 different models, we demonstrate that, in addition to being interpretable, our algorithm yields higher predictive accuracy than extant methods. The approach is sufficiently abstract to be applicable to a wide variety of business analytics datasets. Collapse

Topics: analytical information system logistics management electronic commerce Python operating system

Methods: artificial neural network deep learning computational algorithm machine learning data transformation