Online Consumer Behaviour in Social Media Post Types: A Data Mining Approach
(A2020-63455)

AUTHORS

Dimitrios Gkikas, University of Patras; Theodoros Theodoridis, University of Salford; Prokopis Theodoridis, University of Patras; Androniki Kavoura, Prof. University of West Attica

KEYWORDS

social; decision; Facebook

ABSTRACT

Our research focuses in justifying the performance of different types of social media posts extracted from real posts in fashion and cosmetics Facebook business pages after a live video was introduced as a new posting type. The data we used include posts of a different nature like video, photos, statuses, and links. User engagement metrics consist of comments, shares, and reactions. The dataset is analysed through a study of the averages of the different engagement metrics for different timeframes. We applied machine learning and data mining classification techniques on benchmarked dataset using the WEKA platform to highlight a variety of reactions on different status posts. Finally, we present the classified posts performances upon several status posts and users’ reactions. We hope that our research will reveal to decision makers, marketers and managers valuable information incorporating new social media strategy for leveraging their fashion businesses.