This paper offers insight into the 16 Myers-Briggs Type Indicator (MBTI) personality types and how they may affect the diction used by online users on social media platforms such as Twitter and YouTube. The Myers-Briggs Type Indicator categorizes individuals who take the indicator test into one of 16 different personality types, and each of these types have distinct characteristics, from the simple Introverted versus Extraverted to Intuitive or Sensing, Feeling or Thinking, and Judging or Perceiving. These 4 sets of binary characteristics produce 16 different personalities that are often used to create general pictures or summaries about the individual who was assigned a certain personality type. The characteristics can, on occasion, even predict the potential actions of the individual based on their assigned personality type. This is what allows for the objective of this paper to be achieved - to use data analysis and machine learning to identify the number of times certain words were used by those of different personalities on online platforms, find patterns, and observe if the mechanic prediction of MBTI type based on words used in online posts is possible. The three machine-learning algorithms used to predict the personality types were the Naive Bayes, Gradient, and Random Forest algorithms, with a randomly-selected 80% of the data being used to train the algorithms and the remaining 20% being used to test the machine-learning for accuracy and specificity. This paper will analyze 433,750 total individual posts made online, along with the programming-processed data and the final results of the predictions, identifying which algorithm was most effective in predicting MBTI type and what future steps could be taken to increase accuracy and capacity.
展开▼
