Feature configuration and fusion framework.

<div><p>The automated valuation model (AVM) has been widely used by real estate stakeholders to provide accurate property value estimations automatically. Traditional valuation models are subjective and inaccurate, and previous studies have shown that machine learning (ML) approaches per...

Full description

Saved in:
Bibliographic Details
Main Author: Lin Deng (127858) (author)
Published: 2025
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1852020268838420480
author Lin Deng (127858)
author_facet Lin Deng (127858)
author_role author
dc.creator.none.fl_str_mv Lin Deng (127858)
dc.date.none.fl_str_mv 2025-05-19T17:36:05Z
dc.identifier.none.fl_str_mv 10.1371/journal.pone.0321951.g006
dc.relation.none.fl_str_mv https://figshare.com/articles/figure/Feature_configuration_and_fusion_framework_/29101794
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Medicine
Neuroscience
Space Science
Biological Sciences not elsewhere classified
Mathematical Sciences not elsewhere classified
Information Systems not elsewhere classified
structured tabular data
spatial distribution patterns
source unstructured data
shapley additive explanations
real estate stakeholders
multiple linear regression
experimental results show
real estate valuation
automated valuation model
lightgbm ), k
approaches perform better
formulate ml pipelines
eight ml regressors
source image fusion
source image features
image feature values
remote sensing images
nonlinear associations exist
div >< p
consider integrating multi
traditional valuation models
image fusion
image features
nonlinear associations
valuation models
mlp ),
feature configuration
better understanding
enhanced ml
urban planners
street view
significantly different
reliable predictions
random forest
public authorities
previous studies
nearest neighbors
machine learning
incorporating multi
housing prices
findings contribute
extra tree
city centre
case study
dc.title.none.fl_str_mv Feature configuration and fusion framework.
dc.type.none.fl_str_mv Image
Figure
info:eu-repo/semantics/publishedVersion
image
description <div><p>The automated valuation model (AVM) has been widely used by real estate stakeholders to provide accurate property value estimations automatically. Traditional valuation models are subjective and inaccurate, and previous studies have shown that machine learning (ML) approaches perform better in real estate valuation. These valuation models are based on structured tabular data, and few consider integrating multi-source unstructured data such as images. Most previous studies use fixed feature space for model training without considering the model performance variation brought by various feature configuration parameters. To fill these gaps, this study uses Hong Kong as a case study and proposes an enhanced ML-based real estate valuation framework with feature configuration and multi-source image data fusion, including exterior housing photos, street view and remote sensing images. <sup>‌‌</sup>Eight ML regressors, namely, Random Forest, Extra Tree, XGBoost, Light Gradient Boosting Machine (LightGBM), K-Nearest Neighbors (KNN), Support Vector Regression (SVR), Multilayer Perceptron (MLP), and Multiple Linear Regression (MLR) are used to formulate ML pipelines for training. The SHapley Additive exPlanations (SHAP) method is used to examine the effects of images on housing prices. The experimental results show that the model performances using different feature configuration parameters are significantly different, indicating the necessity of feature configuration to obtain more accurate and reliable predictions. Extra Tree performs significantly better than other models. Half of the top 10 significant features are image features, and incorporating multi-source image features can improve property valuation accuracy. Nonlinear associations exist between image features and housing prices, and the spatial distribution patterns of image feature values and corresponding SHAP main effects vary significantly from the city centre to the suburbs. These findings contribute to a better understanding of AVM development with image fusion and the nonlinear associations between image features and housing prices for public authorities, urban planners, and real estate developers.</p></div>
eu_rights_str_mv openAccess
id Manara_2eb7d0ffac733dbd6370da45d73c7c39
identifier_str_mv 10.1371/journal.pone.0321951.g006
network_acronym_str Manara
network_name_str ManaraRepo
oai_identifier_str oai:figshare.com:article/29101794
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Feature configuration and fusion framework.Lin Deng (127858)MedicineNeuroscienceSpace ScienceBiological Sciences not elsewhere classifiedMathematical Sciences not elsewhere classifiedInformation Systems not elsewhere classifiedstructured tabular dataspatial distribution patternssource unstructured datashapley additive explanationsreal estate stakeholdersmultiple linear regressionexperimental results showreal estate valuationautomated valuation modellightgbm ), kapproaches perform betterformulate ml pipelineseight ml regressorssource image fusionsource image featuresimage feature valuesremote sensing imagesnonlinear associations existdiv >< pconsider integrating multitraditional valuation modelsimage fusionimage featuresnonlinear associationsvaluation modelsmlp ),feature configurationbetter understandingenhanced mlurban plannersstreet viewsignificantly differentreliable predictionsrandom forestpublic authoritiesprevious studiesnearest neighborsmachine learningincorporating multihousing pricesfindings contributeextra treecity centrecase study<div><p>The automated valuation model (AVM) has been widely used by real estate stakeholders to provide accurate property value estimations automatically. Traditional valuation models are subjective and inaccurate, and previous studies have shown that machine learning (ML) approaches perform better in real estate valuation. These valuation models are based on structured tabular data, and few consider integrating multi-source unstructured data such as images. Most previous studies use fixed feature space for model training without considering the model performance variation brought by various feature configuration parameters. To fill these gaps, this study uses Hong Kong as a case study and proposes an enhanced ML-based real estate valuation framework with feature configuration and multi-source image data fusion, including exterior housing photos, street view and remote sensing images. <sup>‌‌</sup>Eight ML regressors, namely, Random Forest, Extra Tree, XGBoost, Light Gradient Boosting Machine (LightGBM), K-Nearest Neighbors (KNN), Support Vector Regression (SVR), Multilayer Perceptron (MLP), and Multiple Linear Regression (MLR) are used to formulate ML pipelines for training. The SHapley Additive exPlanations (SHAP) method is used to examine the effects of images on housing prices. The experimental results show that the model performances using different feature configuration parameters are significantly different, indicating the necessity of feature configuration to obtain more accurate and reliable predictions. Extra Tree performs significantly better than other models. Half of the top 10 significant features are image features, and incorporating multi-source image features can improve property valuation accuracy. Nonlinear associations exist between image features and housing prices, and the spatial distribution patterns of image feature values and corresponding SHAP main effects vary significantly from the city centre to the suburbs. These findings contribute to a better understanding of AVM development with image fusion and the nonlinear associations between image features and housing prices for public authorities, urban planners, and real estate developers.</p></div>2025-05-19T17:36:05ZImageFigureinfo:eu-repo/semantics/publishedVersionimage10.1371/journal.pone.0321951.g006https://figshare.com/articles/figure/Feature_configuration_and_fusion_framework_/29101794CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/291017942025-05-19T17:36:05Z
spellingShingle Feature configuration and fusion framework.
Lin Deng (127858)
Medicine
Neuroscience
Space Science
Biological Sciences not elsewhere classified
Mathematical Sciences not elsewhere classified
Information Systems not elsewhere classified
structured tabular data
spatial distribution patterns
source unstructured data
shapley additive explanations
real estate stakeholders
multiple linear regression
experimental results show
real estate valuation
automated valuation model
lightgbm ), k
approaches perform better
formulate ml pipelines
eight ml regressors
source image fusion
source image features
image feature values
remote sensing images
nonlinear associations exist
div >< p
consider integrating multi
traditional valuation models
image fusion
image features
nonlinear associations
valuation models
mlp ),
feature configuration
better understanding
enhanced ml
urban planners
street view
significantly different
reliable predictions
random forest
public authorities
previous studies
nearest neighbors
machine learning
incorporating multi
housing prices
findings contribute
extra tree
city centre
case study
status_str publishedVersion
title Feature configuration and fusion framework.
title_full Feature configuration and fusion framework.
title_fullStr Feature configuration and fusion framework.
title_full_unstemmed Feature configuration and fusion framework.
title_short Feature configuration and fusion framework.
title_sort Feature configuration and fusion framework.
topic Medicine
Neuroscience
Space Science
Biological Sciences not elsewhere classified
Mathematical Sciences not elsewhere classified
Information Systems not elsewhere classified
structured tabular data
spatial distribution patterns
source unstructured data
shapley additive explanations
real estate stakeholders
multiple linear regression
experimental results show
real estate valuation
automated valuation model
lightgbm ), k
approaches perform better
formulate ml pipelines
eight ml regressors
source image fusion
source image features
image feature values
remote sensing images
nonlinear associations exist
div >< p
consider integrating multi
traditional valuation models
image fusion
image features
nonlinear associations
valuation models
mlp ),
feature configuration
better understanding
enhanced ml
urban planners
street view
significantly different
reliable predictions
random forest
public authorities
previous studies
nearest neighbors
machine learning
incorporating multi
housing prices
findings contribute
extra tree
city centre
case study