FEATURE LEARNING AND PRICE ESTIMATION USING DEEP LEARNING IN REAL ESTATE

Abstract

Taiwanese houses, flats, and other types of urban homes should be looked at very carefully. Real estate researchers and forecasters frequently implement surveys and statistical instruments to accumulate data. Nevertheless, when confronted with multidimensional data, these methodologies become inefficient and time-consuming. The objective of this initiative was to develop a real estate forecasting model that could adapt to the current environment and circumstances. The data was standardized to ensure that it was categorized correctly after being extracted from public government records. In order to guarantee that the appropriate clustering method was implemented, cross-statistical analysis was implemented. In order to enhance the precision of classification, an automatic encoder for deep learning was implemented. Double-bottom map ppaper swarm optimization (DBM-PSO) was implemented to identify the optimal clustering solution. We sought to determine the factors that contributed to the increase in property prices in Taiwan over the past decade by employing cluster analysis and deep learning on publicly available data. The findings indicate that the average unit price, the frequency of real estate transactions, and the building material and construction index all significantly influence the prices of property in Taiwan. These discoveries can assist policymakers and researchers in addressing specific aspects of real estate development and improving market regulation to prevent excessive growth. The framework that was proposed is a novel approach to analyzing and forecasting real estate trends.