Coastal Erosion Prediction at Oesapa Beach, Kupang City, Using Backpropagation and Fuzzy Time Series

Coastal erosion is a significant environmental problem that threatens shoreline stability, coastal infrastructure, and community livelihoods. This study aims to predict coastal erosion at Oesapa Beach, Kupang City, by comparing the Backpropagation algorithm and Fuzzy Time Series (FTS) using three environmental variables: sea wave height, tidal fluctuations, and wind speed. Historical environmental data obtained from the Meteorology, Climatology, and Geophysics Agency (BMKG) of East Nusa Tenggara for the 2010–2016 period were processed and analyzed to develop the prediction models. The research involved data preparation, preprocessing, model training, testing, prediction, and performance evaluation using Mean Squared Error (MSE). The results demonstrate that Backpropagation achieved lower testing errors across all three environmental variables, with MSE values of 0.089607 for sea wave height, 0.075699 for tidal fluctuations, and 0.054443 for wind speed. In contrast, FTS produced MSE values of 69.3302, 2.426, and 4647.1802, respectively. The predicted coastal erosion value for 2020 was 0.52 m using Backpropagation and 2.02 m using FTS. These findings indicate that Backpropagation provided better predictive performance than FTS within the scope of the evaluated dataset. The comparison highlights the importance of selecting an appropriate forecasting method when modeling environmental variables associated with coastal erosion. Backpropagation is therefore identified as the more suitable method among the two approaches evaluated in this study for predicting coastal erosion at Oesapa Beach, Kupang City.