Accuracy achieved by the Intensified LSTM based rainfall prediction model is 87.99%. Figure 9. Comparison of accuracy for Neural Network based prediction models. The comparison of accuracies and RMSE of the predictive models, including the statistical methods, is shown in the form of a bar chart in Figure 10.

A comparison of the performance of models based on LSTM-Networks, Stacked-LSTM Networks, Bidirectional-LSTM Networks, XGBoost, and an ensemble model resulted from performing an AutoML approach was performed in the task of forecasting the amount of rainfall per hour using time-series data from five major UK cities.

Intensified LSTM based Recurrent Neural Network is constructed to predict rainfall. This approach is compared with other methods, namely, Holt–Winter, ARIMA, ELM, RNN and LSTM, in order to demonstrate the improvement of rainfall prediction in the proposed approach. For attaining good accuracy, it is necessary to consider previous datasets.

With MAAPE value was obtained as . LSTM algorithm is good in capturing rainfall time series patterns. This paper employs an efficient prediction model based on LSTM for rainfall prediction of Nigeria. A type of RNN called long short-term memory networks was developed to solve the problem of long-term dependency learning.