Regression Neural Networks Advantage over Classical Regression Analysis

In this study, two analyzing methods are used to predict housing prices in California: neural network forecasting methods and methods based on regression analysis. Using the example of individual forecast indicators produced on the basis of two methods, the forecast results are compared. The purpose of this study is to show that the accuracy of prediction by neural networks is higher than that of the classical method. The assessment is carried out by creating a product in Python, which was chosen for reasons of ease of implementation of this analysis, ease of implementation of the product, as well as ease of constructing a graphical analysis of the results obtained. An open data source consisting of sixteen thousand items, which includes a number of housing criteria and prices based on these criteria, was used as resources for training the neural network. A broad review of studies comparing the predictive performance of artificial neural network-based methods and other forecasting methods is conducted. Much attention is paid to comparing artificial neural network methods and linear regression methods. Based on the results of this study, it was revealed that the accuracy of the neural network model is much higher when predicting results using linear regression methods, depending on the introduction of new forecasting criteria.