yazidsupriadi/bot-detector-lstm

🧠 Bot Detector LSTM

Model deteksi akun bot berbasis teks dan fitur numerik menggunakan LSTM.

---

📈 History Training

Epoch	Loss	Accuracy	Precision	Recall	F1-Score
1	0.0349	0.8472	0.8482	0.8474	0.8472
2	0.0292	0.8373	0.8402	0.8374	0.8370
3	0.0236	0.8390	0.8431	0.8392	0.8386
4	0.0210	0.8395	0.8400	0.8396	0.8395
5	0.0222	0.8470	0.8473	0.8471	0.8470
6	0.0111	0.8458	0.8492	0.8459	0.8454
7	0.0246	0.8387	0.8432	0.8390	0.8383
8	0.0147	0.8575	0.8575	0.8575	0.8575
9	0.0122	0.8433	0.8440	0.8433	0.8432
10	0.0176	0.8273	0.8314	0.8275	0.8268

---

🧪 Evaluasi pada Data Uji

Confusion Matrix

Training

Classification Report ROC AUC Score
ROC AUC: 0.9370 Test Metrics Summary Test Loss : 0.7760 Accuracy : 0.8273 Precision : 0.8314 Recall : 0.8275 F1-Score : 0.8268