Household-Electric-Power-Consumption-Prediction-using-Machine-Learning

⚡ Household Electric Power Consumption Prediction

A complete end-to-end machine learning project that analyzes and predicts household electricity consumption using real-world time series data. Seven models are trained, evaluated, and compared to find the best predictor.

📌 Project Overview

This project covers the full data science pipeline — from raw data cleaning to model deployment-ready predictions — using the UCI Household Power Consumption dataset recorded at one-minute intervals over nearly 4 years.

🔄 Pipeline

Raw Data → Cleaning → Feature Engineering → EDA → Modeling → Evaluation → Insights

📊 Dataset

Source: UCI Machine Learning Repository — Individual Household Electric Power Consumption

Size: ~2 million minute-level records

Target: Global_active_power (kW)

Features: Voltage, Global reactive power, Sub-metering 1/2/3, date/time

🛠️ What's Inside

1. Data Preprocessing

Parsed datetime index from Date + Time columns

Time-based interpolation for missing values

Resampled from minute-level to hourly averages

Removed duplicates and statistical outliers (IQR method)

2. Feature Engineering

Extracted: hour, day_of_week, month, year, is_weekend

Normalized features using MinMaxScaler

3. Exploratory Data Analysis

Time series trends

Correlation heatmap

Consumption by hour, weekday, and month (boxplots)

Distribution histogram with outlier visualization

4. Models Trained

Model Type

Linear Regression Baseline

Support Vector Machine (SVR) Kernel-based

Gradient Boosting Ensemble

Random Forest Ensemble

MLP Neural Network Deep Learning

XGBoost Boosting

LSTM Recurrent Neural Network

5. Evaluation Metrics

RMSE (Root Mean Squared Error)

R² Score

MAE (Mean Absolute Error)

Training Time

📈 Key Results

✅ Best models: XGBoost & Random Forest (highest R², lowest RMSE)

🕐 Most important features: Hour of day, Sub-metering 3

🌙 Peak consumption: Evening hours (6–9 PM)

📅 Weekend patterns differ significantly from weekdays

🧰 Tech Stack

Library Purpose

Pandas / NumPy Data manipulation

Scikit-learn ML models & preprocessing

XGBoost Gradient boosting

TensorFlow / Keras LSTM model

Matplotlib / Seaborn Visualization