Prediction of Flood by Rainfall using MLP Classifier of Neural Network Model
Floods are one of the most cataclysmic events in the face of Earth, which are exceptionally unpredictable to demonstrate and difficult to predict at an earlier point of time. The examination of improvement of flood-forecast designs has added to diminish hazard, policy suggestion, minimization of the loss of lives and damage caused to properties pertaining to floods. To reenact the compound numerical interpretations of physical actions of floods, during the last two decades, neural system blueprints have assisted in improving and advancing flood prediction structure providing better execution and cost-effective solutions. This project helps predict the occurrence of a flood by rainfall dataset with neural network-based techniques to prevent this problem of floods. The analysis of the dataset by Multi-Layer Perceptron Classifier (MLP) is carried out to acquire certain information like; variable identification, treatment of missing values, data validation and data cleansing and preparation will be done on the complete provided dataset. The performance of the algorithms utilized for flood forecasting is seen by the accuracy computation with an evaluation classification report, confusion matrix identification and result. It shows the effectiveness of the graphical user interface-based application by predefined attributes and gives an early alarm for an impending disaster.
Keywords: SK-learn, rainfall, disaster, natural language processing, forecast.