Personality Prediction From Twitter Dataset Using Machine Learning

Abstract

Social networking sites have become the most popular Internet destination, giving social scientists a unique chance to study online behavior. A rising number of study articles on social media are being published, with just a few of them focusing on personality prediction. Personality assessments computer based on data from social media platforms has proven to be more accurate than judgments given by persons who are familiar with the topic. Text on social networking sites is used to automatically detect an individual's personality qualities. We are using the Myers-Briggs Type Indicator (MBTI) dataset. These datasets have 16 types and 8675 posts. From the input text, we categorized four personality qualities using the Myers-Briggs Type Indicator. They are, in particular IntroversionExtroversion (I-E), Intuitions-Sensing(N-S), Feeling-Thinking(F-T), and JudgingPerceiving(J-P) . This data set we collected from Kaggle. Firstly, preprocessing the dataset. This is text data so that we are using NLP for preprocessing. Then using the machine learning techniques. Tokenization, word stemming, stop words deletion and feature, as well as TF IDE, are examples of text preprocessing methods. We are using six machine learning algorithm. We compare all of algorithm, among them the Support Vector Machine (SVM) have best for highest accuracy. Support vector machines outperform the other six machine learning algorithms in terms of accuracy, according to the results of an experiment.