A user's personality prediction approach by mining network interaction behaviors on Facebook. Issue 7 (14th November 2016)

• Record Type: Journal Article
• Authors: Chen, Tsung-Yi; Tsai, Meng-Che; Chen, Yuh-Min
• Is Part Of: Online information review. Volume 40:Issue 7(2016)
• Page Start: 913
• Page End: 937
• Publication Date: 2016-11-14
• Subjects: Facebook -- Social media -- DISC theory -- Interaction feature -- Personality predicting; 025.04
• DOI: 10.1108/OIR-08-2015-0267
• Languages: English
• ISSNs: 1468-4527

Abstract

Abstract : Purpose: For an enterprise, it is essential to win as many customers as possible. The key to successfully winning customers is often determined by understanding the personality characteristics of the object of communication in order to employ an effective communication strategy. An enterprise needs to obtain the personality information of target or potential customers. However, the traditional method for personality evaluation is extremely costly in terms of time and labor, and it cannot acquire customer personality information without their awareness. Therefore, the manner in which to effectively conduct automated personality predictions for a large number of objects is an important issue. The paper aims to discuss these issues. Design/methodology/approach: The diverse social media that have emerged in recent years represent a digital platform on which users can publicly deliver speeches and interact with others. Thus, social media may be able to serve the needs of automated personality predictions. Based on user data of Facebook, the main social media platform around the world, this research developed a method for predicting personality types based on interaction logs. Findings: Experimental results show that the Naïve Bayes classification algorithm combined with a feature selection algorithm produces the best performance for predicting personality types, with 70-80 percent accuracy. Research limitations/implications: In this research, the dominance, inducement, submission, and compliance (DISC) theory was used to determine personality types. Some specific limitations were encountered. As Facebook was used as the main data source, it was necessary to obtain related data via Facebook's API (FB API). However, the data types accessible via FB API are very limited. Practical implications: This research serves to build a universal model for social media interaction, and can be used to propose an efficient method for designing interaction features. Originality/value: This research has developed an approach for automatically predicting the personality types of network users based on their Facebook interactions.