Semantic Approach for Big Five PersonalityPredictiononTwitter

Semantic Approach for Big Five PersonalityPredictiononTwitter

Big Five Personality, BERT, SVM,LIWC

Personality provides a deepinsight of someone and hasanimportant part in someone’s job performance. Predicting personality throughsocial media has been studied onseveral research.The problem is how to improve the performance of personality prediction system. The purpose of this research is to predict personalityon Twitter usersand increase the performance of thepersonality predictionsystem. An online survey using Big Five Inventory (BFI) questionnaire has been distributed and gathered 295Twitteruserswith 511,617tweets data. In this research, we experimenton twodifferent methods using SupportVector Machine (SVM),and the combination of SVM and BERTas the semantic approach.This research also implementsLinguistic Inquiry Word Count (LIWC) asthe linguistic feature for personality prediction system.Theresults showed thatcombination of these two methods achieve 79.35% accuracyscore and with the implementation of LIWC can improvethe accuracy score up to 80.07%.Overall, these results showed thatthecombination of SVM and BERTas the semantic approach with the implementation of LIWCis recommended to gain a better performance for the personality prediction system

GhinaDwi Salsabila1, Erwin Budi Setiawan

https://jurnal.iaii.or.id/index.php/RESTI/issue/view/24

Telkom University

20 agustus 2021

Fajar bagus W

PDF

Indonesia

Text