Power Consumption Prediction using Random Forest

This project involves building a predictive model to estimate power consumption using a Random Forest Regressor. The goal is to analyze historical data, train a model, and evaluate its performance for predicting future power consumption.

Project Overview

The dataset contains daily power consumption data, and the objective is to predict future power consumption based on various features such as:

• Date of the measurement
• Day of the week
• Month
• Semester
• Quarter

We use Random Forest Regressor to predict power consumption, evaluate the model performance using Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), and visualize the actual vs predicted power consumption.

Data

The project uses two CSV files for training and testing:

1. df_train.csv - Training data containing historical power consumption records.
2. df_test.csv - Test data used to evaluate the model’s performance.

Columns:

Available in df_train.csv and df_test.csv:

Column	Type	Description
date	chr	Date of the measurement
power_consumption	dbl	Daily power consumption (in kilowatts)
year	int	Year of the measurement
semester	int	Semester of the measurement (1 for Jan-Jun, 2 for Jul-Dec)
quarter	int	Quarter of the measurement (1 for Q1, 2 for Q2, 3 for Q3, 4 for Q4)
day_in_week	chr	Day of the week of the measurement (e.g., Monday, Tuesday)
week_in_year	int	Week number in the year of the measurement
day_in_year	int	Day number in the year of the measurement
month	int	Month of the year of the measurement

Model

The model used for predicting power consumption is the Random Forest Regressor from sklearn.ensemble. Random Forest is an ensemble learning method that builds multiple decision trees and aggregates their predictions. It is widely used for regression tasks and can handle complex non-linear relationships in data.

Steps Involved:

1. Data Preprocessing:

   • One-hot encoding for day_in_week to convert categorical variables into numerical format.
   • Ensuring that the test data has the same features as the training data.

2. Training:

   • Splitting the data chronologically into training and validation sets (80% train, 20% validation).
   • Training the Random Forest model on the training set.

3. Evaluation:

   • Predicting power consumption on the validation set.
   • Calculating MAE and RMSE to evaluate the model's performance.
   • Plotting actual vs predicted power consumption for validation and test sets.

4. Test Prediction:

   • The model is then used to predict power consumption on the test set and evaluate the results.

Setup

To run the project on your local machine, follow these steps:

Prerequisites:

Ensure you have the following installed:

• Python 3.x
• pip (Python package installer)

Installation:

1. Clone the repository:

   git clone https://github.com/rd9437/predict_energy_consumption.git
   cd predict_energy_consumption