Perbandingan Tingkat Akurasi Prediksi Peningkatan Kasus Positif Covid-19 antara Metode Neural Network Backpropagation dan Long Short Term Memory (LSTM)

Agus Alwi Mashuri (STMIK HIMSYA SEMARANG)
Eko Riyanto (STMIK HIMSYA SEMARANG)

Jurnal Informatika Upgris
Vol 8, No 2: Desember 2022

The COVID-19 (Coronavirus) pandemic is likely to be one of the most serious globalproblems in the past year. Countries do not have similar experiences with the spreadof the virus and its effects from various fields. Estimating the number of previous casesof COVID-19 can help make decisions in the form of actions and plans to prevent thevirus. This study aims to provide a forecasting model that predicts confirmed COVID-19 cases in the city of Semarang. This study applies a machine learning algorithm,namely the Recurrent Neural Network (RNN) to predict COVID-19 cases in the city ofSemarang. The process of fine-tuning each model is described in this study andnumerical comparisons between the two models are concluded using differentevaluation measures; mean sequence error (MSE).