Personality is an individual's unique combination of traits that determines how they think, feel, and behave. There is a chance to automatically identify a person's personality qualities from the text they provide on social networking sites. The purpose of this study is to utilize the XGBoost classifier, a machine learning technique, to predict four personality qualities from input text based on the Myers- Myers Type Indicator (MBTI) model: introversion, extroversion, intuition, and sensing-thinking, respectively. Experiments make use of the Kaggle benchmark dataset, which is freely accessible to the public. The key problem with the previous work is the bias of the dataset, which is reduced by using the Re-sampling approach, also known as random over-sampling, leading to improved performance. Pre-processing methods including tokenization, word stemming from, stop words deletion, and feature selection utilizing TF IDF are also used to get more insight about the author's character from the text. This research lays the groundwork for designing an individual identification system that might help businesses better understand their consumers and attract the most qualified employees. All classifiers provide respectable results when applied to the full set of personality characteristics, but the XGBoost classifier's performance stands out because to its ability to consistently achieve above 99% precision and accuracy.