Determinants of Consumer Purchasing Factors through LDA Modeling Using YouTube Data

In modern times, people use the Internet every day, and they can connect to the Internet anytime, anywhere. They can watch the news, check the weather, find nearby restaurants, and navigate their way. They also post and share their feelings and opinions on social-network services such as Twitter, Facebook, Instagram, and YouTube. Due to this change, they use online review services to share their feeling, opinions, and experiences of products or services in the comments sections and make purchasing decisions based on this information. The most well-known method to identify factors related to consumers’ purchasing decisions is the survey. However, the survey method is time consuming, and the amount of data obtained is limited. Thus, we propose a system that uses YouTube data to identify such factors, reducing time constraints and supplementing the limited amount of data. The proposed method builds a Latent Dirichlet Allocation (LDA) based on the data collected to extract topics and people’s opinions. The biggest advantage of our system is that it reduces the time it takes to obtain data and uses more data for analysis.