LSTM-Based NLP Chatbot for Fish E-Marketplace at BBI Cangkiran Mijen
Efficient and responsive information services are essential to support the fish sales process at the Cangkiran Mijen Fish Hatchery Center (Balai Benih Ikan), Semarang City. Interviews with hatchery staff revealed that the fish trading process is still conducted conventionally, requiring buyers to vi...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | Indonesian |
| Published: |
Islamic University of Indragiri
2025-09-01
|
| Series: | Sistemasi: Jurnal Sistem Informasi |
| Subjects: | |
| Online Access: | https://sistemasi.ftik.unisi.ac.id/index.php/stmsi/article/view/5170 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Efficient and responsive information services are essential to support the fish sales process at the Cangkiran Mijen Fish Hatchery Center (Balai Benih Ikan), Semarang City. Interviews with hatchery staff revealed that the fish trading process is still conducted conventionally, requiring buyers to visit the hatchery in person. Currently, information regarding fish sales is only available through the official Semarang City Government website and Google Maps, which provides limited and often incomplete details. To obtain more comprehensive information, the public must contact staff via WhatsApp or visit the site directly. Moreover, customer inquiries tend to be repetitive, making the information service less effective. To address these issues, this study aims to develop a web-based fish e-marketplace system integrated with a natural language processing (NLP) chatbot using the Long Short-Term Memory (LSTM) algorithm. The system is expected to provide more informative, responsive, and always-available information services without relying on staff availability. The chatbot was trained using 757 question-and-answer pairs as training data. The system was developed using the Software Development Life Cycle (SDLC) waterfall model. Testing results indicate that the system demonstrates good functionality, is compatible across multiple devices and web browsers, and received positive feedback from users regarding ease of interaction and the relevance of chatbot responses. Algorithm validation results show an accuracy of 97%, precision of 94%, and recall of 95%. |
|---|---|
| ISSN: | 2302-8149 2540-9719 |