Floods are rare and dangerous disaster in minimum duration, which have the most destructive impact within urban and rural areas. This research in flood prediction models contributed to risk reduction, to prevent the loss of human life, and reduce the property of damage in floods. This study implements the automated machine learning models, using the Support Vector Machine (SVM) and Artificial Neural Network (ANN). The rainfall data and various meteorological parameter which include temperature data are used in this study. Concurrent daily records of inflow and discharge are taken into consideration to calculate the water level to quantify the importance of the lake flow. It aims to discovering accurate and efficient for the flood forecasting model. This paper attempts to forecast flood by modelling water level, temperature and rainfall data in the region of Korattur lake, Chennai, India. In this study, ultrasonic sensor used to capture the measurement of water level to predict from ultrasonic waves and input of same implemented in BPNN and Support Vector Machine (SVM) were used for flood forecasting. The water level flow is deducted in this research using ultrasonic sensor, proves the best efficient models applied for flood forecasting. This study can be used as a predicting the flood by choosing the proper Machine Learning (ML) algorithm such as Support Vector Machine (SVM) and Artificial Neural Network (ANN) algorithm for showing higher accuracy. To get more accurate result of the models, three standard statistical performance evaluation parameters, Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and coefficient of determination ( ) were used to analyse the performance of the model developed. As a result, the proposed model proves the most efficiency and accuracy for predicting the flood forecasting
Forecast of water level and ice jam thickness using the back propagation neural network and support vector machine methods
