Deep Learning based personality recognition model for the hindi language

Abstract

Personality is a concern with individual differences in characteristics and patterns of human thinking, feeling, and behavior. Computational recognition of user personality is likely to be useful in many computational applications and technologies. The usage of words by individuals provides cues to their personalities. Therefore, in the existing studies user-generated content is employed to understand their emotions and personality traits. However, research on personality detection has primarily focused on foreign languages, with no research work reported on the Hindi language. Hindi is the fourth most widely spoken language in the world and many people in India express their thoughts and emotions online or on other platforms in the Hindi language. The recent increase in Hindi language content on the web highlights the need to identify personality traits through Hindi text analysis. Our objective is to gain insight and evaluate human perceptions expressed by individuals and various communities by addressing the challenge of recognizing personality in Hindi language content. newline In this work, we focus on the development of the Hindi Personality Dataset and the Hindi Psycholinguistic Dictionary to achieve our research objective to recognize users personalities from Hindi text. This work describes the Data Preprocessing task, Feature Extraction process, and Classification process. The Psycholinguistic features, Sentiment features, TF-IDF, and Deep Learning based Word Embedding features have been extracted for utilizing psycholinguistic, emotional, and word embedding features in the classification task. Furthermore, the feature selection process has been carried out to improve the performance of the classification models. The Machine Learning algorithms such as Multinomial Naïve Bayes (MNB) and Support Vector Machine(SVM) are utilized for the classification of personality traits. The five binary MNB and SVM classifiers are trained for each personality trait.