IMPLEMENTASI CHATBOT WHATSAPP UNTUK LAYANAN INFORMASI DESA MENGGUNAKAN TF-IDF DAN COSINE SIMILARITY BERBASIS NATURAL LANGUAGE PROCESSING

Andika Ramadhana (UNIVERSITAS HASYIM ASY'ARI)
Hery Kristianto (UNIVERSITAS HASYIM ASY'ARI)
Edwin Hari Agus Prastyo (UNIVERSITAS HASYIM ASY'ARI)

JUTIM (Jurnal Teknik Informatika Musirawas)
Vol. 11 No. 2 (2026): JUTIM (Jurnal Teknik Informatika Musirawas) Juni

Village information services are still conducted manually, limiting public access to timely information. This study aims to develop a WhatsApp-based chatbot to automate village information services using Natural Language Processing (NLP), TF-IDF, and Cosine Similarity. The system employs a text preprocessing pipeline (case folding, tokenization, stopword removal, stemming), TF-IDF weighting, and similarity matching against a knowledge base of 15 FAQ pairs. Evaluation on 15 test queries yielded an average Cosine Similarity of 0.91, with 14 out of 15 queries exceeding the 0.80 threshold (93.33% success rate). This system contributes to reducing the administrative burden on village staff and improving public service response time.