Housing Price Prediction Using Deep Neural Networks: A Case Study on California Data

Abstract

Accurate estimation of housing prices is essential for informed decision-making in urban planning, real estate investment, and economic forecasting. In this study, a Deep Neural Network (DNN) model is developed to predict median housing values using the California Housing dataset. The dataset, comprising multiple socioeconomic and geographical features, is preprocessed through feature scaling and partitioned into training and testing subsets. The proposed DNN architecture consists of multiple hidden layers employing ReLU activations, optimized using the Adam algorithm with Mean Squared Error (MSE) as the loss function. The model was trained over 100 epochs, achieving a final test MSE of 0.2578 and a Mean Absolute Error (MAE) of 0.3380. Moreover, an R-squared score of 0.8032 indicates strong predictive power and generalization capability. These results suggest that deep learning models can effectively capture complex, nonlinear relationships in housing data and offer reliable tools for real-world applications.