A machine learning project that applies regression techniques to predict house prices using real-world tabular data. This end-to-end workflow demonstrates core skills in data preprocessing, exploratory data analysis, feature selection, model building, evaluation, and comparison using Python’s data science ecosystem.
This project highlights fundamental machine learning competencies by walking through a complete regression pipeline using Python. It includes data exploration, cleaning, feature engineering, model implementation, and performance evaluation.
- Load data from an Excel file.
- Use
.info()and.describe()to understand data structure and summary statistics. - Identify missing values using
.isnull().sum().
- Drop irrelevant columns using
.drop()based on initial inspection.
- Generate a correlation heatmap to identify relationships between variables.
- Select features based on a defined correlation threshold with the target variable.
Implemented and trained two regression models using scikit-learn:
- ✅ Decision Tree Regressor
- ✅ Polynomial Linear Regression (via pipeline)
- Evaluate model performance using Mean Absolute Percentage Error (MAPE).
- Visualize predicted vs. actual values using line plots for model interpretability.
- Compare the performance of different models using bar plots of evaluation metrics.
- Correlation Heatmap
- Actual vs Predicted Value Plots
- Model Comparison Bar Chart
- Language: Python
- Libraries:
pandas,numpymatplotlib,seabornscikit-learn