Exploiting time-series analysis to predict customers’ behavioural dynamics in social networks

In the recent years, companies have focused on ways to take advantage of the rich and endless data that flow from the social media. The effective exploitation of such data has a crucial role in businesses’ growth and development. Businesses in order to survive in a highly competitive market need to analyze and understand customers’ needs and behaviours. This research work addresses the problem of predicting the behavioural dynamics of the customers regarding a brand name, using time-series analytics in social media. While in the existing literature many researchers have focused on studying the individual analytics, in this work we propose a more coarse-grained approach that analyses the overall behavioural dynamics in social networks. In particular, a data mining model is being presented, which exploits users’ linguistic and emotional characteristics in social networks, in order to predict the future collective behavioural trends. The purpose of this work is to provide an efficient and automated tool, that will enable businesses to predict the relationships with their customers. Thus, businesses will be able to strengthen and reestablish their bonds with their customers, through appropriate social media marketing campaigns in a timely manner.