RAINFALL DAILY PREDICTION BASED ON ARTIFICIAL NEURAL NETWORK

Abstract

High rainfall tends to result in flooding of some areas and affects regions that are accustomed to wet seasons, and hence, those that live in states with high rain, mainly along the east coast, need to be very careful. Due to this fact, very few people are conversant with the effects of unreliable rainfall. Statistical forecasting tools for rainfall are not helpful in the making of long-term predictions because climate factors are not static. The process of rainfall prediction using Artificial Neural Networks has taken great strides, and so far, ANN forecasting is among the most used methods for rain forecasting. The objective is to collect rainfall data and apply it along with the ANN to predict the daily rainfall. In addition, this also evaluates the model by taking the difference between the predicted output and the desired output. To execute the solution, in this section, the rainfall prediction with justifications is carried out using the Backpropagation Neural Network (BPNN) technique embedded in the sci-kit-learn library as an MLP, it is specific to regression tasks. It measures the forecast performance in direct reference to the quantification of Mean Square Error (MSE) associated with the actual output. BPNN and MLP are developed neural networks for modelling and prediction of rainfall. BPNN and MLP models can use predictive algorithms in a manner that gives good accuracy in giving predictions