Discovering the Key Factors for Increasing Viewers of Live Streaming

Live streaming has become one of the most important audio-visual behaviours of users in Taiwan, and it has gradually become popular to enable people to interact with each other. According to available literatures, it was reported that the audience's satisfaction, participation and consumption motivations directly influence the audience's watching intention. However, most of related literatures still use qualitative models, questionnaires, and hypotheses to study related issues. Only few works attempt to find key factors of increasing viewers of live streaming. Therefore, this study attempts to define potential factors of influencing watching intention of live streaming. Then, text mining techniques and feature selection methods, including Decision Trees (DT) and Neural Network Pruning (NNP) have been employed to discover the important factors. Some live streaming films and text comments about games in Twitch will be collected for further analysis. The results show that “the number of like＂, and “Tokens＂ are important for users to watch live streaming programs. It can help live-streamers to attract more viewers.