A Machine Learning Approach for Personality Type Identification using MBTI Framework

In recent times, the prediction of personality traits with automated and programmed systems has caught human attention. Specifically, the use of multimodal data to predict personality types is the most considerable talk in artificial intelligence. There are a variety of techniques and methods available for personality type identification. The most popular and highly used personality type identifier is the Myers Briggs Type Indicator (MBTI) type indicator among all methods. In this paper, an exhaustive comparative analysis of all machine learning classical algorithms implementing the MBTI framework will be presented by giving a numerical and graphical representation of performance measures. To experience this study, a supervised machine learning approach is used to perform and analyze different classifiers using the phenomena of MBTI. The models are learned from a dataset to make predictions. The results show that the Ensemble Bagged Trees algorithm gives an overall good training accuracy of 98.4% and test accuracy of 70.75% at a moderate prediction speed of 11 K - Obs/sec by taking a training time of 14 sec. Other than that Coarse Tree algorithm in training time is 0.94009/sec and prediction speed 390 (K - Obs/sec), Fine KNN and Weighted KNN algorithm in training accuracy of 99.20% and Ensemble Boosted Trees algorithm in testing accuracy of 75.51% shows the efficient outcome respectively.