| Summary: | Myers Briggs Type Indicator (MBTI) is well-known instrument for personality evaluation and is frequently being employed in the areas of personal development, career counselling or team building. It categorised people into sixteen personality types as how people expressed themselves to certain traits such as introversion, extroversion, reasoning against feeling, sense and intuition. The main motivation for this work is to predict the classification of these personalities with the usage of machine learning and natural language processing (NLP) thus the goal of this work is to analyse and compare the performance of three NLP text pre-processing features to classify the MBTI personalities in order to increase the accuracy of the prediction model. Bag of Words (BoW), Words Embedding (Word2Vec) and Transfer Learning with Bidirectional Encoder Representation from Transformer (BERT) are methods being compared for their performances in this work. Results shown that transfer learning (BERT) has the highest accuracy followed by word embedding (Word2Vec), and lastly the BoW. © 2023 IEEE.
|
|---|