Topological Data Analysis Via Unsupervised Machine Learning For Recognizing Atmospheric Rivers Conditions On Flood Detection
Flooding is a natural disaster that annually destroys buildings, farmland, properties, and life in many regions of the world. Less than two decades ago, Topological data analysis (TDA) and machine learning (ML) were used in predictions, which have advantages over the common method. Thus, the present...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/60102/ http://eprints.usm.my/60102/1/24%20Pages%20from%20OHANUBA%20FELIX%20OBI.pdf |
| _version_ | 1848884351638437888 |
|---|---|
| author | Obi, Ohanuba Felix |
| author_facet | Obi, Ohanuba Felix |
| author_sort | Obi, Ohanuba Felix |
| building | USM Institutional Repository |
| collection | Online Access |
| description | Flooding is a natural disaster that annually destroys buildings, farmland, properties, and life in many regions of the world. Less than two decades ago, Topological data analysis (TDA) and machine learning (ML) were used in predictions, which have advantages over the common method. Thus, the present work introduces a hybrid method of TDA and unsupervised ML (TDA-uML) for flood management. The TDA-uML blends topological algebra with computer science to become a new study area in statistics, handling shapes in big data. Three properties make TDA distinct from common methods; they are coordinate invariance, deformation invariance, and compressed representation. The method involves training, testing, computation, obtaining of optimal values and validation of optimal value. Some common flood management models such as Hydrologic, hydraulic, and statistical models that researchers had used are inaccurate in the prediction, costly, lack the implementation of hybrid models, and are not validated compared to the TDA-uML method. The technique is aimed at developing a hybrid method of TDA-uML for flood prediction; evaluating the accuracy of the hybrid method (TDA-uML) in predicting flood, choosing the best validity tests for the study, and determining whether there is a relationship in the feature patterns. |
| first_indexed | 2025-11-15T19:05:20Z |
| format | Thesis |
| id | usm-60102 |
| institution | Universiti Sains Malaysia |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-15T19:05:20Z |
| publishDate | 2022 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | usm-601022024-03-11T03:58:15Z http://eprints.usm.my/60102/ Topological Data Analysis Via Unsupervised Machine Learning For Recognizing Atmospheric Rivers Conditions On Flood Detection Obi, Ohanuba Felix QA1 Mathematics (General) Flooding is a natural disaster that annually destroys buildings, farmland, properties, and life in many regions of the world. Less than two decades ago, Topological data analysis (TDA) and machine learning (ML) were used in predictions, which have advantages over the common method. Thus, the present work introduces a hybrid method of TDA and unsupervised ML (TDA-uML) for flood management. The TDA-uML blends topological algebra with computer science to become a new study area in statistics, handling shapes in big data. Three properties make TDA distinct from common methods; they are coordinate invariance, deformation invariance, and compressed representation. The method involves training, testing, computation, obtaining of optimal values and validation of optimal value. Some common flood management models such as Hydrologic, hydraulic, and statistical models that researchers had used are inaccurate in the prediction, costly, lack the implementation of hybrid models, and are not validated compared to the TDA-uML method. The technique is aimed at developing a hybrid method of TDA-uML for flood prediction; evaluating the accuracy of the hybrid method (TDA-uML) in predicting flood, choosing the best validity tests for the study, and determining whether there is a relationship in the feature patterns. 2022-08 Thesis NonPeerReviewed application/pdf en http://eprints.usm.my/60102/1/24%20Pages%20from%20OHANUBA%20FELIX%20OBI.pdf Obi, Ohanuba Felix (2022) Topological Data Analysis Via Unsupervised Machine Learning For Recognizing Atmospheric Rivers Conditions On Flood Detection. PhD thesis, Perpustakaan Hamzah Sendut. |
| spellingShingle | QA1 Mathematics (General) Obi, Ohanuba Felix Topological Data Analysis Via Unsupervised Machine Learning For Recognizing Atmospheric Rivers Conditions On Flood Detection |
| title | Topological Data Analysis Via Unsupervised Machine Learning For Recognizing Atmospheric Rivers Conditions On Flood Detection |
| title_full | Topological Data Analysis Via Unsupervised Machine Learning For Recognizing Atmospheric Rivers Conditions On Flood Detection |
| title_fullStr | Topological Data Analysis Via Unsupervised Machine Learning For Recognizing Atmospheric Rivers Conditions On Flood Detection |
| title_full_unstemmed | Topological Data Analysis Via Unsupervised Machine Learning For Recognizing Atmospheric Rivers Conditions On Flood Detection |
| title_short | Topological Data Analysis Via Unsupervised Machine Learning For Recognizing Atmospheric Rivers Conditions On Flood Detection |
| title_sort | topological data analysis via unsupervised machine learning for recognizing atmospheric rivers conditions on flood detection |
| topic | QA1 Mathematics (General) |
| url | http://eprints.usm.my/60102/ http://eprints.usm.my/60102/1/24%20Pages%20from%20OHANUBA%20FELIX%20OBI.pdf |