In-Training Multicalibrated Survival Analysis For Healthcare via Constrained Optimization

2022 | International Conference on Information Systems | Citations: 0

Authors: Suttaket, Thiti; Kok, Stanley

Abstract: Survival analysis is an important problem in healthcare because it models the r ... Expand

Abstract: Survival analysis is an important problem in healthcare because it models the relationship between an individual's covariates and the onset time of an event of interest (e.g., death). It is important for survival models to be well-calibrated (i.e., for their predicted probabilities to be close to ground-truth probabilities) because badly calibrated systems can result in erroneous clinical decisions. Existing survival models are typically calibrated at the population level only, and thus run the risk of being poorly calibrated for one or more minority subpopulations. We propose a model called GRADUATE that achieves multicalibration by ensuring that all subpopulations are well-calibrated too. GRADUATE frames multicalibration as a constrained optimization problem, and optimizes both calibration and discrimination in-training to achieve a good balance between them. Empirical comparisons against state-of-the-art baselines on real-world clinical datasets demonstrate GRADUATE's efficacy. In a detailed analysis, we elucidate the shortcomings of the baselines vis-à-vis GRADUATE's strengths. Collapse

Topics: online learning system

Methods: survival function nonparametric test logrank test experiment survival analysis

The Effect of Platform Intervention Policies on Fake News Dissemination and Survival: An Empirical Examination

2021 | Journal of Management Information Systems | Citations: 19

Authors: Ng, Ka Chung; Tang, Jie; Lee, Dongwon

Abstract: Fake news on social media has become a serious problem, and social media platfor ... Expand

Abstract: Fake news on social media has become a serious problem, and social media platforms have started to actively implement various interventions to mitigate its impact. This paper focuses on the effectiveness of two platform interventions, namely a content-level intervention (i.e., a fake news flag that applies to a single post) and an account-level intervention (i.e., a forwarding restriction policy that applies to the entire account). Collecting data from China's largest social media platform, we study the impact of a fake news flag on three fake news dissemination patterns using a propensity score matching method with a difference-in-differences approach. We find that implementing a policy of using fake news flag influences the dissemination of fake news in a more centralized manner via direct forwards and in a less dispersed manner via indirect forwards, and that fake news posts are forwarded more often by influential users. In addition, compared with truthful news, fake news is disseminated in a less centralized and more dispersed manner and survives for a shorter period after a forwarding restriction policy is implemented. This study provides causal empirical evidence of the effect of a fake news flag on fake news dissemination. We also expand the literature on platform interventions to combat fake news by investigating a less studied account-level intervention. We discuss the practical implications of our results for social media platform owners and policymakers. Collapse

Topics: fake news social media Sina Weibo filter bubble social network

Methods: regression analysis method hazard function propensity score matching term frequency–inverse document frequency proportional hazards model

Being Counted and Remaining Accountable: Maintenance of Quarterly Earnings Guidance by U.S. Public Companies

2021 | Organization Science | Citations: 0

Authors: Park, Sun Hyun; Patterson, Kelly

Abstract: Organizations are often pressured to adopt and maintain institutionally supporte ... Expand

Abstract: Organizations are often pressured to adopt and maintain institutionally supported practices. Why do some companies remain committed to these practices, despite high operational cost and widespread frustration with them? Although prior theorists have emphasized the importance of institutional pressure at the broader population level, less research attention has been paid to the abandonment of a practice as a result of resource dependence between a firm and market intermediaries. In this paper, we theorize intermediary coverage breadth and depth as two important structural indicators of resource dependence. Firms lacking in coverage breadth (as indicated by the degree of reporting by market intermediaries) and firms with deeper coverage (as evidenced by a prolonged relationship with market intermediaries) are less likely to abandon a practice due to an increase in power imbalance and mutual dependence within the firm-intermediary relationships. We also theorize how a firm’s resource dependence, as determined by coverage structure, moderates the firm’s sensitivity to (1) observed peer support for the practice, (2) the intermediary’s expectation regarding the continued use of the practice, and (3) performance deviations that fail to meet intermediary expectations. Our empirical study of the abandonment of quarterly earnings guidance by U.S. public companies during 2001–2010 provides overall support for our theoretical arguments. Collapse

Topics: organizational productivity research and development database system strategic management information exchange

Methods: theory development descriptive statistic robustness check longitudinal research data transformation

Theories: resource dependency theory institutional theory

Deterrent Effects of Warnings on User’s Behavior in Preventing Malicious Software Use

2017 | HICSS | Citations: 0

Authors: Silic, Mario; Back, Andrea

Abstract: Despite the fact that a number of technical counter-measures do exist to mitigat ... Expand

Abstract: Despite the fact that a number of technical counter-measures do exist to mitigate the risks related to malicious software, in reality users are the last line of defense against security incidents. In this technology-human interaction, warning messages can represent an important tool to help users when making a decision. Understanding the effects of computer warnings on the progression and duration of the malicious software use would bridge the existing knowledge gap. Supported by the restrictive deterrence model and psychological factors, we conducted a non-controlled field experiment in which we collected data from no previously recruited participants. We found that in the presence of the warning message, the progression of the software use will be decreased and the duration of both first and repeated software uses will be reduced. Finally, we offer important findings for further theorizing and interesting practitioner insights that could help to leverage the interaction between the human and the computer technology with an objective to reduce the risk. Collapse

Topics: information system use malware computer system ergonomics decision making

Methods: experimental group theory development field experiment nonparametric test Kaplan-Meier estimator

Theories: deterrence theory

Scarcity Strategy in Crowdfunding: An Empirical Exploration

2017 | International Conference on Information Systems | Citations: 4

Authors: Yang, Lusi; Wang, Zhiyi; Hahn, Jungpil

Abstract: Scarcity-based marketing strategy has been increasingly adopted in reward-based ... Expand

Abstract: Scarcity-based marketing strategy has been increasingly adopted in reward-based crowdfunding in the form of reward limits, whereby project creators are able to restrict the quantity of contributors in each reward tier. This study uncovers the effect of reward limits in eventual and concurrent funding performance. Specifically, we performed campaign and campaign-day level analysis. At the campaign level, we find that setting reward limit at the beginning of a campaign is beneficial for final funding outcomes across four different performance measures. The number of limited reward tiers is shown to have an inverted U-shaped relationship with fundraising performance. Potential endogeneity issues are addressed with propensity score matching and the Heckman selection model. At the campaign-day level, we find incorporating new limited reward tiers is helpful for attracting new backers, but having reward tiers being “sold out” will demotivate subsequent backers to contribute thus lead to lower funding speed in subsequent days. Collapse

Topics: crowdsourcing crowdfunding marketing management accounting electronic market

Methods: propensity score method experimental group econometric modeling regression analysis method propensity score matching

Aircraft re-routing optimization and performance assessment under uncertainty

2017 | Decision Support Systems | Citations: 0

Authors: Zhang, Xiaoge; Mahadevan, Sankaran

Abstract: The need for aircraft re-routing arises when there is disruption in the system, ... Expand

Abstract: The need for aircraft re-routing arises when there is disruption in the system, such as when an airport is closed due to extreme weather. In this paper, we investigate a simulation-based approach to optimize the aircraft re-routing process, by considering multiple sources of uncertainty. The proposed approach has four main components: system simulation, uncertainty representation, aircraft re-routing algorithm, and system performance assessment. Several sources of uncertainty are accounted for in this approach, related to incoming aircraft, space availability in neighboring airports, radar performance, and communication delays. An aircraft re-routing optimization model is formulated to make periodic re-routing decisions with the objective of minimizing the overall distance travelled by all the aircraft, subject to the system resources. We analyze the performance of this aircraft re-routing system using system failure time as the metric. Since the simulation time is limited, right-censored data arises with respect to system failure time. A novel methodology is developed to compute the lower bound of system failure time in the presence of right-censored data, and to analyze the sensitivity of the system performance metric to the uncertain variables relating to the aircraft, radars, nearby airports, and communication system. Since the simulation is time-consuming, we build a Support Vector Regression (SVR) surrogate model to efficiently construct the system failure time distribution. Collapse

Topics: IS failure IT productivity communication service infrastructure system reliability IS evaluation

Methods: simulation sensitivity analysis computational algorithm nonparametric test Kaplan-Meier estimator

Examining Hacker Participation Length in Cybercriminal Internet-Relay-Chat Communities

2016 | Journal of Management Information Systems | Citations: 0

Authors: Benjamin, Victor; Zhang, Bin; Jr, Jay F. Nunamaker; Chen, Hsinchun

Abstract: To further cybersecurity, there is interest in studying online cybercriminal com ... Expand

Abstract: To further cybersecurity, there is interest in studying online cybercriminal communities to learn more about emerging cyber threats. Literature documents the existence of many online Internet Relay Chat (IRC) cybercriminal communities where cybercriminals congregate and share hacking tools, malware, and more. However, many cybercriminal community participants appear unskilled and have fleeting interests, making it difficult to detect potential long-term or key participants. This is a challenge for researchers and practitioners to quickly identify cybercriminals that may provide credible threat intelligence. Thus, we propose a computational approach to analyze cybercriminals IRC communities in order to identify potential long-term and key participants. We use the extended Cox model to scrutinize cybercriminal IRC participation for better understanding of behaviors exhibited by cybercriminals of importance. Results indicate that key cybercriminals may be quickly identifiable by assessing the scale of their interaction and networks with other participants. Collapse

Topics: computer crime IT security online community participatory design denial of service attack

Methods: proportional hazards model survival analysis support vector machine Kaplan-Meier estimator nonparametric test

Knowledge-acquisitions and post-acquisition innovation performance: a comparative hazards model

2015 | European Journal of Information Systems | Citations: 0

Authors: Datta, Pratim; Roumani, Yaman

Abstract: Do acquisitions lead to instrumental innovations related to the acquired knowled ... Expand

Abstract: Do acquisitions lead to instrumental innovations related to the acquired knowledge? Past arguments on vertical integration espouse how a quest for knowledge drives acquisitions culminating in innovation performance. Using Google and Yahoo as cases-in-point, we examine how facets of acquired innovation knowledge impact post-innovation performance. In particular, the apparently opposing fortunes of Google and Yahoo allow us to investigate the pace of their innovation performance as a hazards model. Results from our investigation highlight Google’s ambidexterity over Yahoo with a swifter, systematic pace of innovation performance – from hastening time to patenting new ideas to the time to releasing new applications from acquisitions. Collapse

Topics: innovation management Google+ knowledge base business process innovation idea generation

Methods: hazard function proportional hazards model regression analysis method computational algorithm descriptive statistic

Theories: core competency theory

Insider Threats in a Financial Institution: Analysis of Attack-Proneness of Information Systems Applications

2015 | Management Information Systems Quarterly | Citations: 0

Authors: Jingguo Wang; Gupta, Manish; Rao, H. Raghav

Abstract: This study investigates the risk of insider threats associated with different ap ... Expand

Abstract: This study investigates the risk of insider threats associated with different applications within a financial institution. Extending routine activity theory (RAT) from criminology literature to information systems security, hypotheses regarding how application characteristics, namely value, inertia, visibility, accessibility, and guardians, cause applications to be exposed to insider threats are developed. Routine activity theory is synthesized with survival modeling, specifically a Weibull hazard model, and users’ system access behavior is investigated using seven months of field data from the institution. The inter-arrival times of two successive unauthorized access attempts on an application are employed as the measurement of risk. For a robustness check, the daily number of unauthorized attempts experienced by an application as an alternative measurement of risk are introduced and a zero-inflated Poisson-Gamma model is developed. The Markov chain Monte Carlo (MCMC) method is used for model estimations. The results of the study support the empirical application of routine activity theory in understanding insider threats, and provide a picture of how different applications have different levels of exposure to such threats. Theoretical and practical implications for risk management regarding insider threats are discussed. This study is among the first that uses behavioral logs to investigate victimization risk and attack proneness associated with information assets. Collapse

Topics: insider threat criminality IT manager data protection risk management

Methods: nonparametric test survey hazard function survival function robustness check

Theories: activity theory

THE QUEST FOR CONTENT: THE INTEGRATION OF PRODUCT NETWORKS AND SOCIAL NETWORKS IN ONLINE CONTENT EXPLORATION

2010 | International Conference on Information Systems | Citations: 0

Authors: Goldenberg, Jacob; Oestreicher-Singer, Gal; Reichman, Shachar

Abstract: Without the "guidance" of traditional marketing campaigns, how do consumers fin ... Expand

Abstract: Without the "guidance" of traditional marketing campaigns, how do consumers find "good" usergenerated-content online?Recently, websites have begun to present content through combined product and social networks, linked by hyperlinks. We focus on the role of this dual network structure in facilitating ill-defined exploration of the content space. We first analyze the YouTube.com and show that user pages have unique structural properties and act as content brokers in the dual network. Using simulation analysis we show that random rewiring of the product network cannot replicate the brokering effect of the self-organizing social network. Finally, we introduce an experiment in which consumers browse a YouTube-based website that offers video recommendations through different networks. Using survival analysis we show that the dual network structure leads to faster access to "good" content and to overall higher satisfaction. Our work suggests that integrating a selforganizing social network with product networks significantly improves content exploration. Collapse

Topics: social network YouTube website user satisfaction internet technology

Methods: experiment hazard function simulation survival analysis computational algorithm