Scholarlink

<p>This paper presents an implementation of Support Vector Machine, Artificial Neural Network and Gated Recurrent Unit for predicting alert levels to be issued from dam based on environmental parameters to help in flood forecasting. This will help the government authorities to take better decision in controlling floods and post disaster activities. The dataset utilized includes features such as temperature, dew point, humidity, wind speed, pressure, water level, storage, rainfall, discharge of water from the dam, and inflow of water to the dam. The three models used for the research are Support Vector Machine(SVM), Artificial Neural Network(ANN) and Gated Recurrent Unit(GRU).The paper discusses the SVM model’s activities like data loading, pre-processing, model training, evaluation, and visualization of results in detail. The SVM model is trained using different kernels, and their respective performances are evaluated and compared. The study conducted shows that SVM with linear kernel performed with the highest accuracy of 98.01% in predicting the alerts of No alert, Blue alert, orange alert and Red alert compared to ANN and GRU models.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Comparative Analysis of Artificial Neural Networks, Support Vector Machines, and Gated Recurrent Units for Dam Alert Prediction

  • 期刊论文

  • 发表日期：

    2026年3月17日

查看全文
  • Nisha C.M,
  • N. Thangarasu

摘要

This paper presents an implementation of Support Vector Machine, Artificial Neural Network and Gated Recurrent Unit for predicting alert levels to be issued from dam based on environmental parameters to help in flood forecasting. This will help the government authorities to take better decision in controlling floods and post disaster activities. The dataset utilized includes features such as temperature, dew point, humidity, wind speed, pressure, water level, storage, rainfall, discharge of water from the dam, and inflow of water to the dam. The three models used for the research are Support Vector Machine(SVM), Artificial Neural Network(ANN) and Gated Recurrent Unit(GRU).The paper discusses the SVM model’s activities like data loading, pre-processing, model training, evaluation, and visualization of results in detail. The SVM model is trained using different kernels, and their respective performances are evaluated and compared. The study conducted shows that SVM with linear kernel performed with the highest accuracy of 98.01% in predicting the alerts of No alert, Blue alert, orange alert and Red alert compared to ANN and GRU models.