Factors Dominating Individual Information Disseminating Behavior on Social Networking Sites.
Juan Shi,Kin Keung Lai,Ping Hu,Gang Chen
DOI: https://doi.org/10.1007/s10799-017-0278-8
2017-01-01
Information Technology and Management
Abstract:Identifying dominating features that affect individual information retweeting behavior on social networking sites (SNSs) is crucial to understanding individual retweeting behaivor and developing effective marketing strategies on SNS. However, there is little agreement on what factors are dominating individual information disseminating behavior on SNS, and what’s worse, more and more factors are added into the prediction model, without examining the relevance of them and even why these factors are added is rarely discussed. This leads to undesirable outcomes such as increasing the cost of measuring and computing irrelevant/redundant features. Most importantly, it hinders us from understanding what discriminative features are affecting individual information disseminating behavior. Using a unique real-life Twitter data set consisting of 55,575 twitterers and 9,440,321 tweets, the authors examine what discriminative features are dominating individual information disseminating behavior. The results indicate that topic distance is the most discriminative factor, highlighting that self-presentation motives play an important role in information disseminating decisions. Besides, the amount of information, social relationship and the popularity of the tweet also contribute to individual information disseminating decisions. Experiments demonstrate that adopting only dominating factors can improve prediction performance in terms of various indicators, compared with adopting the full features set. Finally, we conclude the paper by discussing theoretical and practical implications of our findings.