Supervised versus Semi-Supervised Urban Functional Area Prediction: Uncertainty, Robustness and Sensitivity
To characterize a community-scale urban functional area using geo-tagged data and available land-use information, several supervised and semi-supervised models are presented and evaluated in Hong Kong for comparing their uncertainty, robustness and sensitivity. The following results are noted: (i) A...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-01-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/2/341 |
| _version_ | 1797437516488376320 |
|---|---|
| author | Rui Deng Yanning Guan Danlu Cai Tao Yang Klaus Fraedrich Chunyan Zhang Jiakui Tang Zhouwei Liao Zhishou Wei Shan Guo |
| author_facet | Rui Deng Yanning Guan Danlu Cai Tao Yang Klaus Fraedrich Chunyan Zhang Jiakui Tang Zhouwei Liao Zhishou Wei Shan Guo |
| author_sort | Rui Deng |
| collection | DOAJ |
| description | To characterize a community-scale urban functional area using geo-tagged data and available land-use information, several supervised and semi-supervised models are presented and evaluated in Hong Kong for comparing their uncertainty, robustness and sensitivity. The following results are noted: (i) As the training set size grows, models’ accuracies are improved, particularly for multi-layer perceptron (MLP) or random forest (RF). The graph convolutional network (GCN) (MLP or RF) model reveals top accuracy when the proportion of training samples is less (greater) than 10% of the total number of functional areas; (ii) With a large amount of training samples, MLP shows the highest prediction accuracy and good performances in cross-validation, but less stability on same training sets; (iii) With a small amount of training samples, GCN provides viable results, by incorporating the auxiliary information provided by the proposed semantic linkages, which is meaningful in real-world predictions; (iv) When the training samples are less than 10%, one should be cautious using MLP to test the optimal epoch for obtaining the best accuracy, due to its model overfitting problem. The above insights could support efficient and scalable urban functional area mapping, even with insufficient land-use information (e.g., covering only ~20% of Beijing in the case study). |
| first_indexed | 2024-03-09T11:20:28Z |
| format | Article |
| id | doaj.art-a8c99370f50f4b2cae5ff0c5c7dcddca |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T11:20:28Z |
| publishDate | 2023-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-a8c99370f50f4b2cae5ff0c5c7dcddca2023-12-01T00:19:11ZengMDPI AGRemote Sensing2072-42922023-01-0115234110.3390/rs15020341Supervised versus Semi-Supervised Urban Functional Area Prediction: Uncertainty, Robustness and SensitivityRui Deng0Yanning Guan1Danlu Cai2Tao Yang3Klaus Fraedrich4Chunyan Zhang5Jiakui Tang6Zhouwei Liao7Zhishou Wei8Shan Guo9Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, ChinaThe School of Architecture, Tsinghua University, Beijing 100084, ChinaMax Planck Institute for Meteorology, 20146 Hamburg, GermanyAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, ChinaCollege of Resources and Environment, University of Chinese Academy of Sciences, Beijing 100049, ChinaYangtze River Basin Operation Management Center, China Three Gorges, Co., Ltd., Yichang 443133, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, ChinaTo characterize a community-scale urban functional area using geo-tagged data and available land-use information, several supervised and semi-supervised models are presented and evaluated in Hong Kong for comparing their uncertainty, robustness and sensitivity. The following results are noted: (i) As the training set size grows, models’ accuracies are improved, particularly for multi-layer perceptron (MLP) or random forest (RF). The graph convolutional network (GCN) (MLP or RF) model reveals top accuracy when the proportion of training samples is less (greater) than 10% of the total number of functional areas; (ii) With a large amount of training samples, MLP shows the highest prediction accuracy and good performances in cross-validation, but less stability on same training sets; (iii) With a small amount of training samples, GCN provides viable results, by incorporating the auxiliary information provided by the proposed semantic linkages, which is meaningful in real-world predictions; (iv) When the training samples are less than 10%, one should be cautious using MLP to test the optimal epoch for obtaining the best accuracy, due to its model overfitting problem. The above insights could support efficient and scalable urban functional area mapping, even with insufficient land-use information (e.g., covering only ~20% of Beijing in the case study).https://www.mdpi.com/2072-4292/15/2/341geotagged dataurban functional areagraph convolutional networkmulti-layer perceptronsemi-supervised learning |
| spellingShingle | Rui Deng Yanning Guan Danlu Cai Tao Yang Klaus Fraedrich Chunyan Zhang Jiakui Tang Zhouwei Liao Zhishou Wei Shan Guo Supervised versus Semi-Supervised Urban Functional Area Prediction: Uncertainty, Robustness and Sensitivity Remote Sensing geotagged data urban functional area graph convolutional network multi-layer perceptron semi-supervised learning |
| title | Supervised versus Semi-Supervised Urban Functional Area Prediction: Uncertainty, Robustness and Sensitivity |
| title_full | Supervised versus Semi-Supervised Urban Functional Area Prediction: Uncertainty, Robustness and Sensitivity |
| title_fullStr | Supervised versus Semi-Supervised Urban Functional Area Prediction: Uncertainty, Robustness and Sensitivity |
| title_full_unstemmed | Supervised versus Semi-Supervised Urban Functional Area Prediction: Uncertainty, Robustness and Sensitivity |
| title_short | Supervised versus Semi-Supervised Urban Functional Area Prediction: Uncertainty, Robustness and Sensitivity |
| title_sort | supervised versus semi supervised urban functional area prediction uncertainty robustness and sensitivity |
| topic | geotagged data urban functional area graph convolutional network multi-layer perceptron semi-supervised learning |
| url | https://www.mdpi.com/2072-4292/15/2/341 |
| work_keys_str_mv | AT ruideng supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity AT yanningguan supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity AT danlucai supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity AT taoyang supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity AT klausfraedrich supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity AT chunyanzhang supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity AT jiakuitang supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity AT zhouweiliao supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity AT zhishouwei supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity AT shanguo supervisedversussemisupervisedurbanfunctionalareapredictionuncertaintyrobustnessandsensitivity |