Machine Learning Project
Pre-processed dataset of 50,000+ households (1M+ data points) using integer encoding and min-max normalisation. Developed a linear regression model in MATLAB and programmed an artificial neural network for comparative analysis. Achieved 97% prediction accuracy through rigorous model evaluation and optimisation.
- Project type
- University Engineering Project
- Role
- Solo Developer
- Date
- March 2024 - April 2024
Overview
Developed machine learning models to predict household energy consumption using a large dataset of 50,000+ households. Implemented both linear regression and neural network approaches with comprehensive data preprocessing and model evaluation.
Technology
- MATLAB
Key features
- Processed 1M+ data points from 50,000+ households
- Integer encoding and min-max normalization preprocessing
- Linear regression model implementation
- Artificial neural network development
- 97% prediction accuracy achieved
Why it was built
- To apply machine learning techniques to real-world energy consumption data
- To compare traditional statistical methods with modern neural networks
- To develop skills in data preprocessing and model evaluation