Comparison of Classification rules for Predicting Personality

Abstract

Personality refers to an individual’s characteristic patterns of thought, emotion and behaviour, together with the psychological mechanisms. It has been a long-term goal for psychologists to understand human personality and its impact on human behaviour. Behavior involves an interaction between a person's underlying personality traits and situational variables. Social media is a place where users present themselves to the world, revealing personal details and insights into their lives. It is evident that there is a strong correlation between users’ personality and the way they behave on online social network and a large number of studies are made on this topic. Objective of this paper is to classify the personality traits into Extraversion and Introversion. Dataset for this purpose is collected from ‘http://personality-testing.info/_rawdata/’ which stores data on personality surveys for the purposes of academic research and personal use. The dataset consists of 140 instances with 17 attributes. Three types of classification rules like J48 Decision Tree classifier, Naive Bayes Classifier and ZeroR Rules available on Weka data mining tool are used to predict Personality traits. Performances of each of these rules are measured by the values of the Precision, Recall, and Accuracy. Comparative study of these methods shows that J48 Decision tree Classifier performs better than Naive Bayes classifier and ZeroR Rule for personality traits recognition on the Social Media.