Akademik Veri Yönetim Sistemi
EGE 12th INTERNATIONAL CONFERENCE ON APPLIED SCIENCES, İzmir, Türkiye, 26 - 30 Aralık 2024, ss.798-805
The valuation of real estate plays a crucial role in informing investment decisions, determiningtax obligations, and conducting financial risk assessments. However, traditional valuationapproaches are often characterized by high costs and lengthy processes. This inherently intricateprocedure is shaped by factors such as market price fluctuations, regional economic conditions,and the state of infrastructural development. Conventional methods typically rely on humanexpertise and subjective judgment, making them prone to inaccuracies and biases. Machinelearning methodologies stand out for their ability to derive meaningful insights from vastdatasets and generate more accurate predictions. Techniques such as machine learning and deeplearning enhance predictive capabilities by analyzing complex data structures and detectingpatterns that may elude human analysts. By leveraging data mining, statistical analysis, andautomated learning algorithms, these models surpass the constraints of traditional approaches.This study explores the feasibility of employing machine learning methodologies to estimatereal estate unit values. The effectiveness of algorithms, linear regression, k-nearest neighbor,decision tree, and random forest was evaluated through empirical analysis. The resultsdemonstrate that machine learning-based models achieve higher accuracy rates compared totraditional regression methods, significantly minimizing errors and biases. Incorporatingmachine learning-driven valuation systems has the potential to revolutionize the real estatesector by enabling more transparent, consistent, and reliable property valuations. The valuation of real estate plays a crucial role in informing investment decisions, determining tax obligations, and conducting financial risk assessments. However, traditional valuation approaches are often characterized by high costs and lengthy processes. This inherently intricate procedure is shaped by factors such as market price fluctuations, regional economic conditions, and the state of infrastructural development. Conventional methods typically rely on human expertise and subjective judgment, making them prone to inaccuracies and biases. Machine learning methodologies stand out for their ability to derive meaningful insights from vast datasets and generate more accurate predictions. Techniques such as machine learning and deep learning enhance predictive capabilities by analyzing complex data structures and detecting patterns that may elude human analysts. By leveraging data mining, statistical analysis, and automated learning algorithms, these models surpass the constraints of traditional approaches. This study explores the feasibility of employing machine learning methodologies to estimate real estate unit values. The effectiveness of algorithms, linear regression, k-nearest neighbor, decision tree, and random forest was evaluated through empirical analysis. The results demonstrate that machine learning-based models achieve higher accuracy rates compared to traditional regression methods, significantly minimizing errors and biases. Incorporating machine learning-driven valuation systems has the potential to revolutionize the real estate sector by enabling more transparent, consistent, and reliable property valuations.