Information & Management最新文献

Wearable sensor-based human activity recognition (HAR) utilizes artificial intelligence models to analyze real-time data like accelerometer data to recognize daily human activities. While it greatly benefits the life of senior citizens and postoperative patients, it conventionally requires the collected data to be uploaded to a central server to train AI models, raising critical security and privacy concerns. Though Federated learning (FL) emerges as a viable way to cope with these problems, it is confronted by the data heterogeneity problem, where the varying activity patterns of different individuals result in non-identically distributed local data. Some FL models have been proposed to solve the data heterogeneity problem by leveraging the similarity between individuals to create a personalized global model for each individual. However, they are still limited by increased computation or unreliable relationships in the similarity computation. This study proposes a novel profile similarity-based personalized federated learning for wearable sensor-based HAR where the similarity between individuals can be reflected in their profile, such as age, gender, height, and weight. When personalizing a model for an individual, we compute the weighted sum of all clients’ local models, where the weight is determined by the similarity value computed from the profile. In this way, the local models from individuals who have higher similarity values will contribute more towards personalizing a model for a targeted individual than those who are less similar. Experiment results demonstrate that the proposed model outperformed the baseline FL and centralized learning on both RealWorld and SisFall datasets. We also discuss the tradeoff between privacy and personalization and FL's advantages over centralized learning.

Web seals have recently been subject to increasing skepticism from consumers. This study develops and tests a theoretical model of web seal skepticism by integrating multiple theoretical perspectives. We explore how web seal skepticism emerges and harms consumers’ perceptions. The results of an online experiment with 757 participants demonstrate that consumer skepticism increases consumers’ mistrust in information system providers and makes them feel less protected. Our findings also indicate that situational factors (web seal unreliability and seal authority incredibility) are significant antecedents of web seal skepticism and that consumers’ inferences of providers’ manipulative intent mediate the relationships.

This study primarily examines how the centralization of online content regulation increases political hate speech in a country. It also explores the roles of the government's social media surveillance and disinformation in this relationship. Calhoun's public choice theory is used as a theoretical foundation to examine relationships. Data from 179 nations are analyzed using a mixed-method approach (i.e., path analysis and fuzzy set Qualitative Comparative Analysis), and the findings reveal how centralization of online content regulation results in higher levels of political hate speech by increasing social media surveillance of political content and disinformation through social media by the government.

Disinformation researchers have much to learn about the psychological factors that lead to users’ biases. The gap in users’ new media literacy, along with personality traits, are factors that may contribute to bias. We conducted an experiment that varied levels of articles' linguistic veracity to determine how these factors influence users' engagement behavior. We found that users with higher literacy in new media were less likely to engage with disinformation featuring low veracity language. However, users with higher literacy in traditional media were more likely to engage with disinformation, regardless of the article's linguistic veracity.

We propose the design of a novel method for multidimensional data visualization, called data avatars. We rely on the design science approach and the dual-processing theory (System 1-System 2) to guide our design and test the data avatars method against two well-known data visualization methods. The test involves a lab experiment with 429 participants from the general public and uses industry data. Our results show that data avatars provide a viable method for visualizing multidimensional data in terms of user cognitive performance. Further, the results indicate the important role of perceived cognitive load and enjoyment in users’ preference for data visualization design.

Augmented reality (AR) merges virtual elements with our physical context. Although there is evidence in marketing that AR may be superior to alternative formats, there is a lack of work explaining from the ground up why this is the case. Consequently, we applied means-end chain theory to identify specific AR-features (e.g., contextualization, interactivity, portability) that drive benefits (e.g., inspiration, better decision-making, time savings, risk reduction). These benefits contribute to consumers’ goal achievement (e.g., self-confidence, self-expression, reduced purchase regret). A subsequent study organized these factors into a practical framework (SEAD and SALES). This study contributes to a better understanding of AR.

Recently, with the popularity of social investing platforms, participating in an investment club has become a good choice for investors. Following financial experts in the investment club likely generates more profit as they have higher expertise in planning an investment portfolio. In this study, we propose a portfolio selection mechanism that combines collective intelligence extracted from investors’ opinions and LSTM stock price predictions to infer a club's investment preference and predict the profitability of the extracted investment targets. Based on a club's risk tolerance and investment preference, the proposed mechanism can create an appropriate stock portfolio for the investors in the club. Utilizing StockTwits and stock historical data, the experimental results verify that the proposed portfolio selection mechanism performs better than market indices and other benchmark approaches in the market.

Although mobile payment systems offer countless advantages, they do present certain drawbacks, mainly associated with security and privacy concerns. The inclusion of biometric authentication technologies seeks to minimise such drawbacks. The aim of this article is to examine the effect of key antecedents of the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) and perceived risk on the intention to use a mobile payment system featuring biometric identification. A new hybrid analytical approach is taken. A sample of more than 2500 smartphone users was obtained through an online panel-based survey. Two techniques were used: first, structural equation modelling (PLS-SEM) was conducted to determine which variables had a significant influence on the adoption of the mobile payment system, and second, an artificial neural network (ANN) model was used, taking a deep learning approach, to rank the relative influence of significant predictors of use intention obtained via PLS-SEM. The study found that the most significant variables affecting use intention were performance expectancy, effort expectancy, facilitating conditions, hedonic motivation and risk. In contrast, subjective norms, price value and habit were found to be weak predictors of use intention. The results of the ANN analysis confirmed almost all SEM findings but yielded a slightly different order of influence among the least significant predictors. A review of the extant scientific literature revealed a paucity of published studies dealing with the adoption and use of mobile payment systems featuring biometric identification. The conclusions and managerial implications point to new business opportunities that can be exploited by firms through the use of this technology.

Although the favorable impact of sharing electronic medical records (EMRs) with other hospitals on improving clinical decision-making efficiency is widely acknowledged, the actual implementation of EMR sharing has been limited to some extent because of patient privacy protections. This study proposes a three-stage framework to retrieve medical treatment plans from multiple hospitals based on federated learning and case-based reasoning (CBR). We demonstrate that the proposed framework compensates for the privacy protection weaknesses of CBR and solves the problem of data islands among hospitals.

Gamification is used to increase the engagement of doctors on online health platforms. We track 2,106 doctors from a leading online health community (Good Doctor) over one year to investigate how gamification features (i.e., how actively doctors embellish their profiles and how they perform in popularity indicators via points and virtual gifts received) affect doctors’ future health service provision. The results showed that gamification can lead to high long-term doctor engagement. Specifically, virtual gifts and points were positively associated with doctors’ instrumental and socio-emotional participation, whereas customization was positively associated with doctors’ instrumental participation.