SVM Models for MBTI Dimension Classification

4 model SVM terpisah untuk klasifikasi dimensi kepribadian MBTI (I/E, N/S, F/T, J/P) berdasarkan vektor distribusi 6 emosi dasar Ekman (angry, disgust, fear, happy, sad, surprise), yang diekstrak menggunakan model akrvnka/distilbert-emotion-ekman-mbti.

Hasil Evaluasi (Stratified 5-Fold CV)

Dimensi MBTI Accuracy (mean ± std) Precision (mean ± std) Recall (mean ± std) Macro F1-Score (mean ± std)
IE 0.5618 ± 0.0173 0.5280 ± 0.0094 0.5302 ± 0.0115 0.5233 ± 0.0060
NS 0.5118 ± 0.0153 0.5414 ± 0.0095 0.5578 ± 0.0131 0.4862 ± 0.0065
FT 0.5782 ± 0.0131 0.5693 ± 0.0111 0.5705 ± 0.0111 0.5691 ± 0.0117
JP 0.5695 ± 0.0098 0.5648 ± 0.0103 0.5652 ± 0.0105 0.5648 ± 0.0103

Kernel & Hyperparameter Terpilih

Dimensi Kernel Terpilih C Terpilih
IE linear 100
NS poly 0.1
FT rbf 100
JP rbf 1

Cara Pakai

import joblib
model_ie = joblib.load("svm_IE_final.pkl")  # download dulu dari repo ini
pred = model_ie.predict([[0.1, 0.05, 0.15, 0.4, 0.2, 0.1]])  # urutan: angry, disgust, fear, happy, sad, surprise

Interpretasi SHAP

Lihat file feature_importance_summary.csv dan folder figures/ untuk visualisasi kontribusi tiap emosi terhadap prediksi masing-masing dimensi MBTI.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support