Pengembangan AISHA: AI Sistem Harga Berbasis Telegram Bot Menggunakan Data SP2KP

Abstract

Informasi harga komoditas pangan penting bagi pedagang dalam menentukan harga jual, namun umumnya diperoleh melalui pengalaman pribadi, distributor, atau observasi langsung ke pasar. Sistem Pemantauan Pasar dan Kebutuhan Pokok (SP2KP) menyediakan data harga nasional, tetapi akses melalui website dinilai kurang praktis. Observasi terhadap delapan pedagang menunjukkan bahwa seluruh responden belum mengetahui SP2KP, sementara 62,5% responden menilai informasi harga nasional penting sebagai acuan dalam menentukan harga jual. Penelitian ini bertujuan mengembangkan AISHA berbasis Telegram Bot untuk menyediakan informasi harga komoditas berdasarkan wilayah dan periode waktu, membandingkan karakteristik workflow Rule-Based Parsing dan AI-Based Parsing dalam memproses masukan pengguna, serta mengevaluasi persepsi pengguna terhadap AISHA setelah menggunakan kedua workflow tersebut. Penelitian dilaksanakan melalui tahapan identifikasi masalah, analisis kebutuhan sistem, perancangan sistem, pengembangan sistem, pengujian sistem, dan evaluasi pengguna. Hasil penelitian menunjukkan bahwa workflow Rule-Based Parsing memiliki karakteristik simple chatbot, sedangkan workflow AI-Based Parsing memiliki karakteristik intelligent chatbot. Workflow AI-Based Parsing mampu memahami variasi masukan pengguna dan memberikan respons yang lebih spesifik, sedangkan workflow Rule-Based Parsing menghasilkan alur interaksi yang lebih sederhana dan langsung. Hasil evaluasi menunjukkan bahwa AISHA memperoleh persepsi positif dari pengguna, di mana mayoritas responden lebih memilih workflow Rule-Based Parsing karena dinilai lebih sederhana dan mudah digunakan dalam memperoleh informasi harga komoditas.

Commodity price information plays an important role for traders in determining selling prices. However, such information is generally obtained through personal experience, distributors, or direct market observations. The Market Monitoring and Basic Needs Information System (Sistem Pemantauan Pasar dan Kebutuhan Pokok / SP2KP) provides national commodity price data, but access through its website is considered less practical. Observations conducted with eight traders showed that none of the respondents were familiar with SP2KP, while 62.5% considered national commodity price information important as a reference for determining selling prices. This study aims to develop AISHA, a Telegram Bot-based system that provides commodity price information based on region and period, compare the characteristics of the Rule-Based Parsing and AI-Based Parsing workflows in processing user input, and evaluate users' perceptions of AISHA after using both workflows. The study was conducted through the stages of problem identification, system requirements analysis, system design, system development, system testing, and user evaluation. The results show that the Rule-Based Parsing workflow exhibits the characteristics of a simple chatbot, whereas the AI-Based Parsing workflow exhibits the characteristics of an intelligent chatbot. The AI-Based Parsing workflow can understand various user input formats and provide more specific responses, whereas the Rule-Based Parsing workflow offers a simpler and more direct interaction. User evaluation indicates that AISHA received positive perceptions from users, with most respondents preferring the Rule-Based Parsing workflow because it was considered simpler and easier to use for obtaining commodity price information.