Electricity Usage Prediction using Random Forest

This project is a machine learning pipeline designed to predict electricity usage using a Random Forest Regressor. It includes data preprocessing, model training, evaluation, hyperparameter tuning, and visualization of predictions.

-- Table of Contents --

#dataset #features #technologies-used #installation #usage #model-training-and-evaluation #hyperparameter-tuning #visualization #results #contributing #license -- Dataset --

The dataset used is a public dataset (recs2015_public_v4.csv), which contains various features related to electricity consumption. Make sure to set your own file path before running the script.

Key columns:

ELECTRICITY_USAGE (Target variable) Categorical and numerical features related to households. -- Features --

Data Preprocessing:

Handles missing values and encodes categorical features using LabelEncoder. Scaling: Standardizes the feature set using StandardScaler. Model Training:

Implements a Random Forest Regressor for prediction. Hyperparameter Tuning:

Uses GridSearchCV for finding optimal hyperparameters. Evaluation:

Assesses model performance using metrics like MSE and R-squared. Visualization:

Plots actual vs. predicted electricity usage. -- Technologies Used --

Python Pandas NumPy (optional) Scikit-learn Matplotlib Seaborn Joblib -- Installation --

Clone this repository:

bash git clone (link unavailable) cd electricity-usage-prediction

Install the required dependencies:

bash pip install -r requirements.txt

Place your dataset in the appropriate directory and update the file_path variable in the script:

file_path = 'path_to_your_dataset/recs2015_public_v4.csv'

-- Usage --

Run the script to preprocess the data, train the model, and evaluate its performance:

bash python (link unavailable)

-- Model Training and Evaluation --

Training:

The Random Forest model is trained using an 80-20 train-test split. Evaluation Metrics:

Mean Squared Error (MSE) R-squared (R²) The model is saved using joblib for future predictions.

-- Hyperparameter Tuning --

A grid search is performed to find the best parameters for the Random Forest model. The following hyperparameters are tuned:

n_estimators: Number of trees in the forest. max_depth: Maximum depth of the tree. min_samples_split: Minimum number of samples required to split a node. Run the hyperparameter tuning process using:

grid_search = GridSearchCV(estimator=model, param_grid=param_grid, cv=5, n_jobs=-1, verbose=2)

-- Visualization --

A scatter plot is generated to compare actual and predicted electricity usage. The red dashed line represents the ideal prediction line (y = x).

plt.figure(figsize=(10, 6)) sns.scatterplot(x=y_test, y=y_pred, color='blue') plt.plot([min(y_test), max(y_test)], [min(y_test), max(y_test)], color='red', linestyle='--') plt.xlabel('Actual Electricity Usage') plt.ylabel('Predicted Electricity Usage') plt.title('Actual vs Predicted Electricity Usage') plt.show()

-- Results --

Best Model Performance:

Mean Squared Error (MSE): mse_best R-squared (R²): r2_best The model's performance improves after hyperparameter tuning.

-- Contributing -- Contributions are welcome! Please follow these steps:

Fork the repository. Create a feature branch: git checkout -b feature-name. Commit your changes: git commit -m 'Add feature'. Push to the branch: git push origin feature-name. Submit a pull request.

-- Members --

Aryan Kumar aka MartinLegend24, Abhinav Nirwan aka UnknownCodrr, Aakash Sharma aka Aakash5268, Aman Kumar aka satoru-coder