Scholarlink

The deregulation of the electricity market led to an increase in competition, which has compelled many strategies makers to opt for forecasting electricity prices because the more accuracy of prediction better the strategy they can make. Various models are present for electricity price forecasting but in this study, our main concern is how we can minimize their forecasting error and maximize their forecasting accuracy. One way of doing that is before applying any model to the dataset, we can make that dataset is understandable enough to the model that can give us possibly better results, this can be achieved by employing feature selection techniques on a dataset, and after this, those all fetched optimal features dimension were reduced using the Principle Component Analysis (PCA), then this final dataset with optimal features and reduced dimensions are fed to the Deep Neural Network (DNN) model for forecasting purposes. This paper has considered Least Absolute Shrinkage and Selection Operator (LASSO) feature selection technique and its results have been compared with other benchmark models on the scale of Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R2 score metrics to evaluate the effectiveness of the proposed model.

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

Short-Term Electricity Price Forecasting of Ontario Market by LASSO-LSTM Model of Deep Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Ashish Prajesh,
  • Prerna Jain,
  • Deepak Ranjan Nayak,
  • Navneet Sharma,
  • Amit Kumar,
  • Abhishek Harit

摘要

The deregulation of the electricity market led to an increase in competition, which has compelled many strategies makers to opt for forecasting electricity prices because the more accuracy of prediction better the strategy they can make. Various models are present for electricity price forecasting but in this study, our main concern is how we can minimize their forecasting error and maximize their forecasting accuracy. One way of doing that is before applying any model to the dataset, we can make that dataset is understandable enough to the model that can give us possibly better results, this can be achieved by employing feature selection techniques on a dataset, and after this, those all fetched optimal features dimension were reduced using the Principle Component Analysis (PCA), then this final dataset with optimal features and reduced dimensions are fed to the Deep Neural Network (DNN) model for forecasting purposes. This paper has considered Least Absolute Shrinkage and Selection Operator (LASSO) feature selection technique and its results have been compared with other benchmark models on the scale of Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R2 score metrics to evaluate the effectiveness of the proposed model.