Utilization of Real Time Behavior and Geographical Attraction for Location Recommendation

Personalized location recommendation is an increasingly active topic in recent years, which recommends appropriate locations to users based on their temporal and geospatial visiting patterns. Current location recommendation methods usually estimate the users’ visiting preference probabilities from the historical check-ins in batch. However, in practice, when users’ behaviors are updated in real-time, it is often cost-inhibitive to re-estimate and updates users’ visiting preference using the same batch methods due to the number of check-ins. Moreover, an important nature of users’ movement patterns is that users are more attracted to an area where have dense locations with same categories for conducting specific behaviors. In this paper, we propose a location recommendation method called GeoRTGA by utilizing the real time user behaviors and geographical attractions to tackle the problems. GeoRTGA contains two sub-models: real time behavior recommendation model and attraction-based spatial model. The real time behavior recommendation model aims to recommend real-time possible behaviors which users prefer to visit, and the attraction-based spatial model is built to discover the category-based spatial and individualized spatial patterns based on the geographical information of locations and corresponding location categories and check-in numbers. Experiments are conducted on four public real-world check-in datasets, which show that the proposed GeoRTGA outperforms the five existing location recommendation methods.

Session interest model for CTR prediction based on self-attention mechanism

Click-through rate prediction, which aims to predict the probability of the user clicking on an item, is critical to online advertising. How to capture the user evolving interests from the user behavior sequence is an important issue in CTR prediction. However, most existing models ignore the factor that the sequence is composed of sessions, and user behavior can be divided into different sessions according to the occurring time. The user behaviors are highly correlated in each session and are not relevant across sessions. We propose an effective model for CTR prediction, named Session Interest Model via Self-Attention (SISA). First, we divide the user sequential behavior into session layer. A self-attention mechanism with bias coding is used to model each session. Since different session interest may be related to each other or follow a sequential pattern, next, we utilize gated recurrent unit (GRU) to capture the interaction and evolution of user different historical session interests in session interest extractor module. Then, we use the local activation and GRU to aggregate their target ad to form the final representation of the behavior sequence in session interest interacting module. Experimental results show that the SISA model performs better than other models.

Identifying Distinct Patterns of Social Media Usage During the Transition into College as Predictors of First-Year Alcohol Use and Consequences: A Latent Profile Analysis

Social media (SM) users are a combination of several behaviors across platforms. Patterns of SM use across platforms may be a better indicator of risky drinking than individual behaviors or sets of behaviors examined previously. This longitudinal study addressed this gap in the literature using latent profile analysis (LPA) to identify subpopulations of SM users during the college transition (N=319). Indicators included in the LPA were general SM (checking, time spent, and posting to Instagram/Facebook/Snapchat; Finstagram ownership) and alcohol-related posting (alcohol, partying, and marijuana content) behaviors. LPA results revealed three SM user subpopulations at baseline: low general use with low alcohol-related posting (LGU+LAP), and high general use with low alcohol-related posting (HGU+LAP) or high alcohol-related posting (HGU+HAP). Baseline drinking, injunctive norms, and alcohol beliefs were associated with greater odds of HGU+HAP membership. Prospective analyses revealed that HGU+HAP was associated with greater alcohol use and consequences relative to HGU+LAP and LGU+LAP. Results suggest that there are distinct patterns of SM use during the college transition associated with risky drinking that can inform interventions combating SM-related alcohol risks. These findings also illustrate the importance of analyzing multiple SM user behaviors across multiple platforms simultaneously in future studies.

Addiction to Social Networking Sites and User Responses

Prior studies have primarily used "variable-centered" perspectives to identify factors underlying user responses to social networking site (SNS) addiction, their predictors and outcomes. This paper extends this perspective by taking a person-centered approach to examine (1) the prototypical subpopulations (profiles) of users' extent of SNS addiction and responses to it, (2) how affiliations with these profiles can explain user behaviors toward SNS use, and (3) how personality traits can predict affiliations with these profiles. To this end, we propose a typological theory of SNS addiction and user responses to it via two empirical, personcentered studies. Study 1 draws on survey data from 188 SNS users to develop a typology of users based on the extent of their SNS addiction and their responses to it. It further examines the relations between affiliation with these profiles and users' SNS discontinuance intention, as a typical behavioral response to SNS addiction. Study 2 uses survey data from 284 SNS users to validate the user typology developed in Study 1 and investigate its relations to users' Big Five personality traits. Our findings shed light on a typology of five prototypical profiles of SNS users-cautious, regular, consonant, dissonant, and hooked-who differ in their extent of SNS addiction and their cognitive, emotional, and behavioral responses to it. Our findings also demonstrate how Big Five personality traits can predict user affiliations with these prototypical profiles.

Google Trends as a predictive tool for COVID-19 vaccinations in Italy: a retrospective infodemiological analysis.

Background: Google Trends is an infoveillance tool widely used by the scientific community to investigate different user behaviors related to COVID-19. However, several limitations regarding its adoption are reported in the literature. Objective: This brief paper aims to provide an effective and efficient approach to investigating vaccine adherence against COVID-19 via Google Trends. Methods: Through the cross-correlational analysis of well-targeted hypotheses, we investigate the predictive capacity of web searches related to COVID-19 towards vaccinations in Italy from November 2020 to November 2021. The keyword "vaccine reservation" (VRQ) was chosen as it reflects a real intention of being vaccinated (V). Furthermore, the impact of the second-largest Italian national newspaper on vaccines-related web searches was investigated to evaluate the role of the mass media as a confounding factor. Results: Simple and generic keywords are more likely to identify the actual web interest in COVID-19 vaccines than specific and elaborated keywords. Cross-correlations between VRQ and V were very strong and significant (min r^2 = .460, P<.001, lag = 0 weeks; max r^2 = .903, P < .001, lag = 6 weeks). Cross-correlations between VRQ and news about COVID-19 vaccines have been markedly lower and characterized by greater lags (min r^2 = .190, P=.001, lag = 0 weeks; max r^2 = .493, P < .001, lag = -10 weeks). No correlation between news and vaccinations was sought since the lag would have been too high. Conclusions: This research provides strong evidence in favor of using Google Trends as a surveillance and prediction tool for vaccine adherence against COVID-19 in Italy. These findings prove that the search for suitable keywords is a fundamental step to reduce confounding factors. Additionally, targeting hypotheses helps diminish the likelihood of spurious correlations. It is recommended that Google Trends be leveraged as a complementary infoveillance tool by government agencies to monitor and predict vaccine adherence in this and future crises by following the methods proposed in this manuscript.

Google Trends as a Predictive Tool for COVID-19 Vaccinations in Italy: a Retrospective Infodemiological Analysis (Preprint)

BACKGROUND Google Trends is an infoveillance tool widely used by the scientific community to investigate different user behaviors related to COVID-19. However, several limitations regarding its adoption are reported in the literature. OBJECTIVE This brief paper aims to provide an effective and efficient approach to investigating vaccine adherence against COVID-19 via Google Trends. METHODS Through the cross-correlational analysis of well-targeted hypotheses, we investigate the predictive capacity of web searches related to COVID-19 towards vaccinations in Italy from November 2020 to November 2021. The keyword "vaccine reservation" (VRQ) was chosen as it reflects a real intention of being vaccinated (V). Furthermore, the impact of the second most read Italian newspaper on vaccines-related web searches was investigated to evaluate the role of the mass media as a confounding factor. RESULTS Simple and generic keywords are more likely to identify the actual web interest in COVID-19 vaccines than specific and elaborated keywords. Cross-correlations between VRQ and V were very strong and significant (min r² = .460, P<.001, lag = 0 weeks; max r² = .903, P < .001, lag = 6 weeks). Cross-correlations between VRQ and news about COVID-19 vaccines have been markedly lower and characterized by greater lags (min r² = .190, P=.001, lag = 0 weeks; max r² = .493, P < .001, lag = -10 weeks). No correlation between news and vaccinations was sought since the lag would have been too high. CONCLUSIONS This research provides strong evidence in favor of using Google Trends as a surveillance and prediction tool for vaccine adherence against COVID-19 in Italy. These findings prove that the search for suitable keywords is a fundamental step to reduce confounding factors. Additionally, targeting hypotheses helps diminish the likelihood of spurious correlations. It is recommended that Google Trends be leveraged as a complementary infoveillance tool by government agencies to monitor and predict vaccine adherence in this and future crises by following the methods proposed in this manuscript.

Analysis on Influence of Business Intelligence Information Quality over User Information Adoption Based on Multiple Mediating Effects

From the perspective of information, this paper constructs a theoretical model based on information system (IS) success model and information adoption model (IAM), aiming to further disclose the action mechanism of business intelligence (BI) on user information adoption (UIA). Firstly, the BI factors affecting UIA were modeled in reference to the existing literature. Next, a questionnaire survey was carried out to collect valid samples from 423 Chinese enterprises. After that, an empirical analysis on structural equations was conducted on Amos 24.0 and SPSS 26.0. The results show that BI information content quality (ICQ), expected performance (EP), expected ease-of-use (EEOU), and perceived risk (PR) have a direct and significant influence on UIA, and ICQ further significantly influences UIA via mediators such as EP, EEOU, and PR; BI information access quality (IAQ) has a direct yet insignificant influence on UIA, but exerts a significant positive effect on UIA of BI via EP and EEOU. The research provides a new perspective into BI user behaviors in the context of big data, and a good reference for the successfully implementation and use of BI in practice.

Joint Modeling of User Behaviors Based on Variable-Order Additive Markov Chain for POI Recommendation

The POI recommendation system has become an important means to help people discover attractive and interesting places. Based on our data analysis, we observe that users pay equal attention to conservatism and curiosity. In particular, adopting analysis corresponding to different time intervals, we find that users lean towards old POIs in the short term and look for new POIs with the increase of the time interval. However, existing approaches usually neglect users’ conservatism and curiosity preferences. Therefore, they are confronted with a bottleneck of depicting accurate user needs, making it difficult to improve the recommendation performance further. Besides, we further find that the number of user daily check-ins has uneven distribution, which is not conducive to capture the accurate transition patterns of user behaviors. In light of the above, we design a single POI sequential method. On this basis, we propose a recommendation method of the variable additive Markov chain. We consider the user sequential preferences, especially liking old and pursuing new features. In addition, our model exploits the geographical tendency of user behaviors. Finally, we conduct abundant experiments on four cities in the two real datasets, i.e., Foursquare and Jiepang. The experimental results show its superiority over other competitors.