Browsing by Author "De Wet, Dominique"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Machine learning models for mass appraisals: advancing valuations in the digital era
    (Stellenbosch : Stellenbosch University, 2023-12) De Wet, Dominique; Du Preez, Johan ; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.
    ENGLISH ABSTRACT: A property appraisal is a professional, unbiased valuation that determines the market value of a property. Traditionally, property appraisals were conducted exclusively by professional appraisers. However, with increased data availability and enhanced computational resources, automated valuation models (AVMs) have gained widespread recognition as efficient tools to assist property appraisers in conducting mass appraisals. This report investigates the suitability of various machine learning (ML) methods as AVMs. The techniques include multiple polynomial regression, random forest regression, support vector regression, and a neural network. In addition to these four models, this report also introduces and assesses the advantages of a new innovative fusion model as an AVM. The fusion model is an ensemble approach that employs a neural network to combine the predictions from the four previous ML models, aiming to achieve improved accuracy and precision. This report uses property sales data from three specific neighbourhoods located within the Western Cape province of South Africa: Edgemead, Pinehurst, and Brackenfell. The results from this study indicate that all the individual ML techniques produce highly accurate property price predictions. However, they also yielded relatively large errors for some predictions. In contrast, the fusion model achieved greater accuracies and minimised most of its errors compared to the standalone models, establishing it as an effective AVM technique. This report provides a comprehensive framework for improving mass appraisal models.