Flood Prediction In Nigeria Using Artificial Neural Network

Apply

Flood Prediction In Nigeria Using Artificial Neural Network

Journal Title: American journal of Engineering Research - Year 2018, Vol 7, Issue 9

Abstract

Flooding has been a major problem both in urban and rural areas in Nigeria and globally. Hundreds of lives and millions dollars of properties are destroyed yearly due to flooding. This research work is based on flood prediction using ANN. These factors of which temperature and rainfall were the most significant were used to develop an ANN model for the prediction of flooding in Nigeria using deep feed-forward neural network. The network has three hidden layers sandwiched between an input layer. It accepts two input features (i.e. temperature and rainfall) and outputs the predicted Standard Precipitation Index (SPI). Two-third (67%) of the dataset was used to train the network using the back propagation algorithm. Adam’s algorithm was used as an optimizer while the loss function used was categorical cross entropy. One-third (33%) of the total dataset was used to test and validate the network during training. From the confusion matrix, the average accuracy of the model on the test set was 76% which although may not be seen as high but will be sufficient for our prediction.

Authors and Affiliations

EsiefarienrheBukohwo Michael1 ,, OfikwuEne Patience

Keywords

EP ID EP398481
DOI -
Views 64
Downloads 0