Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal KOMPUTA : Jurnal Ilmiah Komputer dan Informatika
Fiani, Mia Al
Unknown Affiliation
Computer Science & IT

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Prediksi Temperatur Maksimum di Kota Tanjungpinang Menggunakan Model CNN-LSTM Nurfalinda; Fiani, Mia Al; Rathomi, Muhamad Radzi
Komputa : Jurnal Ilmiah Komputer dan Informatika Vol 14 No 1 (2025): Komputa : Jurnal Ilmiah Komputer dan Informatika
Publisher : Program Studi Teknik Informatika - Universitas Komputer Indonesia (UNIKOM)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.34010/komputa.v14i1.15377

Abstract

The prediction of maximum temperature is important for supporting decision process related to public activities and reducing the consequences of climate change. The goal of this study is to analyze the performance of the CNN-LSTM hybrid method in forecasting maximum temperature in Tanjungpinang City by utilizing average humidity and rainfall as input variables. Historical weather data was obtained through the BMKG website, covering the period from January 1, 2022, to November 30, 2024, and was used as the research dataset. The CNN-LSTM model was developed by optimizing the advantages of CNN in recognizing spatial patterns and the capability of LSTM in capturing temporal patterns. The model was trained using an optimal configuration consisting of 128 CNN filters, a kernel size of 7, 200 LSTM units, a batch size of 16, and 120 epochs. Performance evaluation was conducted using two key metrics: Root Mean Squared Error (RMSE) of 1.65 and Mean Absolute Percentage Error (MAPE) of 4.19%. The findings indicate that the model can be used to predict maximum temperature based on available historical weather data. Additionally, the model has been implemented in a web-based platform that allows users to input historical data and select prediction periods ranging from 1, 3, 7, to 10 days ahead. The prediction results are presented in tables and graphical visualizations to facilitate users in understanding and evaluating the generated information.
Co-Authors Muhamad Radzi Rathomi Nurfalinda, Nurfalinda