Performance of Multi-Layer Perceptron-Neural Network versusRandom Forest Regression for Sea Level Rise Prediction

摘要

Sea Level Rise (SLR) is one of the most difcult elements to predict in the hydrologicalcycle. 12% of the area of Peninsular Malaysia, where the western low plains of muddysediment are home to 2.5 million people, is vulnerable to ?ooding. In this study, two ArtifcialIntelligence (AI) techniques were used to predict SLR, namely, the Multi-Layer PerceptronNeural Network (MLP-NN) and Random Forest Regression (RFR) techniques. This studied,two cases were presented. The frst case (Case 1) was to establish the prediction model forSLR by a monthly data set, while the second case (Case 2) was by means of a cyclical dataset. From sensitivity analysis result, it was found that the most e?ective meteorologicalinput parameters were rainfall (mm) and wind direction (degree). The performance ofthe models was evaluated according to three statistical indices in terms of the correlationcoeffcient (R), root mean square error (RMSE) and scatter index (SI). A comparison ofthe results of the MLP-NN and RFR showed that the MLP-NN performed better than thelatter as the R obtained in Case 2 of the MLP-NN was 0.733 with 65.652 and 2.735 forRMSE and SI respectively. Meanwhile, accuracy improvement percentage (%AI) was 8%.