Abstract:Kepribadian Myers-Briggs Type Indicator ( MBTI ) telah menjadi topik populer dalam memahami karakteristik individu dan dampaknya pada interaksi sosial, karir, dan pengambilan keputusan. Model Machine Learning dengan algoritma Naive Bayes Classifier sering digunakan untuk memprediksi kepribadian MBTI berdasarkan data Twitter. Namun, seringkali terjadi ketidakseimbangan kelas, dengan beberapa jenis kepribadian yang memiliki sampel lebih sedikit. Untuk mengatasi hal ini, penelitian ini menggunakan teknik Synthetic Minority Over-sampling Technique (SMOTE) untuk meningkatkan jumlah sampel pada kelas minoritas. Selain itu, metode Term Frequency-Inverse Document Frequency (TF-IDF) digunakan untuk mengekstraksi fitur penting dari teks. Penelitian ini bertujuan menerapkan teknik SMOTE untuk mengatasi ketidakseimbangan kelas dalam klasifikasi kepribadian MBTI menggunakan beberapa algoritma Naive Bayes Classifier, termasuk Gaussian, Multinomial, Bernoulli, Complement, dan Logistic Regression berdasarkan model Keirsey: Artisan, Guardian, Rational, dan Idealist. Evaluasi menggunakan metode Hold-Out-Validation dengan membagi data menjadi 90% data latih dan 10% data uji. Hasil evaluasi menunjukkan performa rendah algoritma Naive Bayes Classifier untuk kelas Artisan dan Guardian, tetapi baik untuk kelas Rational dan Idealist. Algoritma Logistic Regression memiliki akurasi tertinggi 80% dan performa yang lebih baik secara keseluruhan, meskipun masih rendah untuk kelas Artisan dan Guardian. Dengan demikian, penelitian ini memberikan pemahaman tentang penggunaan algoritma Naive Bayes Classifier dan teknik SMOTE dalam prediksi kepribadian MBTI, dengan potensi peningkatan kinerja melalui penggunaan algoritma Logistic Regression. Abstract Myers-Briggs Type Indicator (MBTI) personality is becoming a popular topic in understanding individual characteristics and their impact on social interaction, career, and decision-making. Machine Learning models with Naive Bayes Classifier algorithms are often used to predict MBTI personalities from Twitter data. However, there is often a class imbalance, with some personality types having a smaller sample. To overcome this, this study used the Synthetic Minority Over-sampling Technique (SMOTE) technique to increase the number of samples in minority classes. Additionally, the Term Frequency-Inverse Document Frequency (TF-IDF) method is used to extract important features from text. This study aims to apply SMOTE techniques to address class imbalances in MBTI personality classification using several Naïve Bayes Classifier algorithms, including Gaussian, Multinomial, Bernoulli, Complement, and Logistic Regression based on Keirsey's model: Artisan, Guardian, Rational, and Idealist. Evaluation using the Hold-Out-Validation method by dividing the data into 90% training data and 10% test data. The evaluation results showed low performance of the Naive Bayes Classifier algorithm for the Artisan and Guardian classes, but both for the Rational and Idealist classes. The Logistic Regression algorithm has the highest accuracy of 79% and better performance overall, although it is still low for the Artisan and Guardian classes. Thus, this study provides insight into the use of Naive Bayes Classifier algorithm and SMOTE technique in MBTI personality prediction, with potential performance improvement through the use of Logistic Regression algorithm.

PersoNet: A Novel Framework for Personality Classification-Based Apt Customer Service Agent Selection

A Machine Learning Approach for Personality Type Identification using MBTI Framework

Personality Trait Detection via Transfer Learning

A Personality-Based Instant Messaging Software User Classification Model

Can Large Language Models Understand You Better? An MBTI Personality Detection Dataset Aligned with Population Traits

A sentiment-aware deep learning approach for personality detection from text

MBTI Personality Prediction Using Machine Learning and SMOTE for Balancing Data Based on Statement Sentences

TranSentGAT: A Sentiment-Based Lexical Psycholinguistic Graph Attention Network for Personality Prediction

Extending the Abstraction of Personality Types based on MBTI with Machine Learning and Natural Language Processing

Enhancing MBTI Personality Trait Prediction from Imbalanced Social Media Data Using Hybrid Query Expansion Ranking and Glo Ve- BiLSTM

Personality Type Based on Myers-Briggs Type Indicator with Text Posting Style by using Traditional and Deep Learning

Myers-Briggs personality classification from social media text using pre-trained language models

Two-Faced Humans on Twitter and Facebook: Harvesting Social Multimedia for Human Personality Profiling

Penerapan SMOTE untuk Mengatasi Imbalance Class dalam Klasifikasi Kepribadian MBTI Menggunakan Naive Bayes Classifier

INTELLIGENT SYSTEM FOR CLASSIFICATION OF STUDENT PERSONALITY WITH NAIVE BAYES ALGORITHM

Combining machine learning algorithms for personality trait prediction

Personality Recognition Based On User Generated Content

PersonaGym: Evaluating Persona Agents and LLMs

Automatic Text-based Personality Recognition on Monologues and Multiparty Dialogues Using Attentive Networks and Contextual Embeddings

Do we behave differently on Twitter and Facebook: Multi-view social network user personality profiling for content recommendation

Designing LLM-Agents with Personalities: A Psychometric Approach