Abstract—This project presents a comparative study of machine learning techniques for electricity demand and price forecasting. A dataset from Kaggle, covering 2106 days between 1st January 2015 to 6th October 2020, is used to train and test models. The dataset contains 14 columns, including demand, RRP, temperature, solar exposure, rainfall, and holiday information. After preprocessing and feature engineering, the dataset is split into training and testing sets. Two machine learning models, LSTM and Random Forest, are implemented and compared. The results show that the Random Forest model achieves an accuracy of 94.4%, outperforming the LSTM model with an accuracy of 95.09%. This study contributes to the development of accurate and reliable models for electricity demand and price forecasting.
Keywords— Electricity demand prediction, Machine learning algorithms, Energy forecasting.
Methodology
Data Description
The dataset used in this project is from Kaggle, covering 2106 days between 1st January 2015 to 6th October 2020. The dataset contains 14 columns, including:
Data Preprocessing
The dataset is preprocessed by converting the date column into day, month, year, and day of the week. The values are then scaled using the Min-Max Scaler, handled missing values, encoded categorical values. And after preprocessing total number of columns are 18
Feature Engineering
The dataset is split into training and testing sets using the train_test_split function from Scikit-learn.
Model Implementation
Two machine learning models, LSTM and Random Forest, are implemented and compared.
Results
The results of the two models are compared in terms of accuracy.
