Implementasi Algoritma ARIMA dan LSTM pada Dashboard Prediksi Saham Berbasis Web

Gautama Wijaya (Universitas International Batam)
Haeruddin Haeruddin (Universitas Internasional Batam)
Louis Lim (Universitas Internasional Batam)

Jutisi: Jurnal Ilmiah Teknik Informatika dan Sistem Informasi
Vol 15, No 2 (2026): April 2026

The volatility of banking stock prices poses challenges for investors in making the right analytical decisions, especially due to the lack of availability of flexible and data-based prediction tools. This research aims to design and build a web-based stock price prediction system that integrates two algorithms at once, namely Autoregressive Integrated Moving Average (ARIMA) and Long Short-Term Memory (LSTM), in one integrated dashboard. The system development is carried out using the System Development Life Cycle (SDLC) method of the Waterfall model, covering the stages of needs analysis, design, implementation, and testing. The system leverages the Yahoo Finance API for automatic market data updates, eliminating reliance on manual file uploads. The test results showed that all functional features were valid, and the accuracy evaluation on BBCA shares proved that LSTM produced a Mean Absolute Percentage Error (MAPE) value of 1.12%, superior to ARIMA which obtained a MAPE of 1.85%. The system provides investors with greater analytical flexibility.Keywords: Stock Prediction; Multi-Model; Yahoo Finance API; Web-Based; Waterfall. AbstrakVolatilitas harga saham perbankan menimbulkan tantangan bagi investor dalam mengambil keputusan analitis yang tepat, terutama akibat minimnya ketersediaan alat bantu prediksi yang fleksibel dan berbasis data terkini. Penelitian ini bertujuan merancang dan membangun sistem prediksi harga saham berbasis web yang mengintegrasikan dua algoritma sekaligus, yaitu Autoregressive Integrated Moving Average (ARIMA) dan Long Short-Term Memory (LSTM), dalam satu dashboard terpadu. Pengembangan sistem dilakukan menggunakan metode System Development Life Cycle (SDLC) model Waterfall, mencakup tahap analisis kebutuhan, desain, implementasi, dan pengujian. Sistem memanfaatkan Yahoo Finance API untuk pembaruan data pasar secara otomatis, sehingga menghilangkan ketergantungan pada unggah file manual. Hasil pengujian menunjukkan seluruh fitur fungsional berjalan valid, dan evaluasi akurasi pada saham BBCA membuktikan bahwa LSTM menghasilkan nilai Mean Absolute Percentage Error (MAPE) sebesar 1,12%, lebih unggul dibandingkan ARIMA yang memperoleh MAPE 1,85%. Sistem ini memberikan fleksibilitas analisis yang lebih luas bagi investor.