California House Pricing Prediction (OLS Linear Regression)
This repository contains a Linear Regression model trained using the Ordinary Least Squares (OLS) method on the California Housing dataset to predict median house values.
📋 Project Overview
In this study, a machine learning model was developed to predict home prices based on various neighborhood features (such as median income, house age, average rooms, and location) in California.
📊 Dataset Specifications
The dataset was loaded using scikit-learn's fetch_california_housing utility. It contains 20,640 samples with 8 independent features:
- MedInc: Median income in block (The most critical feature with a strong positive correlation of ~0.69)
- HouseAge: Median house age in block
- AveRooms: Average number of rooms per household
- AveBedrms: Average number of bedrooms per household
- Population: Block population
- AveOccup: Average house occupancy
- Latitude / Longitude: Geographical location coordinates
- MedHouseVal (Target Variable): Median house value for California districts, expressed in hundreds of thousands of dollars ($100,000s)
Feature Correlation Matrix
Below is the correlation matrix of the dataset features:
🛠️ Model Pipeline and Code Steps
1. Library Imports
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error, r2_score
2. Splitting the Dataset (Train-Test Split)
The dataset is split into an 80% training set and a 20% test set:
X = df.drop(columns=['MedHouseVal'])
y = df['MedHouseVal']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
3. Training the Model (Linear Regression)
model = LinearRegression()
model.fit(X_train, y_train)
y_pred = model.predict(X_test)
📈 Model Evaluation and Performance
The performance metrics computed on the hold-out test set are as follows:
- R² Score (Coefficient of Determination):
0.5758(Our model explains approximately 57.58% of the variance in the target variable.) - Mean Squared Error (MSE):
0.5559 - Root Mean Squared Error (RMSE):
0.7456
Actual vs. Predicted Values Plot
💾 Loading and Running the Model Locally
You can load the pre-trained model file (linear_regression_model.pkl or local counterpart) in Python as follows:
import joblib
# Load the model
loaded_model = joblib.load('linear_regression_model.pkl')
# Make predictions
predictions = loaded_model.predict(X_test)