Kernel dependence analysis and graph structure morphing for novelty detection with high-dimensional small size data set
In this study, we propose a new approach for novelty detection that uses kernel dependence techniques for characterizing the statistical dependencies of random variables (RV) and use this characterization as a basis for making inference. Considering the statistical dependencies of the RVs in multiva...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Elsevier BV
2020
|
| Online Access: | https://hdl.handle.net/1721.1/126504 |
| _version_ | 1826205903527870464 |
|---|---|
| author | Mohammadi Ghazi Mahalleh, Reza Welsch, Roy E Buyukozturk, Oral |
| author2 | Massachusetts Institute of Technology. Department of Civil and Environmental Engineering |
| author_facet | Massachusetts Institute of Technology. Department of Civil and Environmental Engineering Mohammadi Ghazi Mahalleh, Reza Welsch, Roy E Buyukozturk, Oral |
| author_sort | Mohammadi Ghazi Mahalleh, Reza |
| collection | MIT |
| description | In this study, we propose a new approach for novelty detection that uses kernel dependence techniques for characterizing the statistical dependencies of random variables (RV) and use this characterization as a basis for making inference. Considering the statistical dependencies of the RVs in multivariate problems is an important challenge in novelty detection. Ignoring these dependencies, when they are strong, may result in inaccurate inference, usually in the form of high false positive rates. Previously studied methods, such as graphical models or conditional classifiers, mainly use density estimation techniques as their main learning element to characterize the dependencies of the relevant RVs. Therefore, they suffer from the curse of dimensionality which makes them unable to handle high-dimensional problems. The proposed method, however, avoids using density estimation methods, and rather, employs a kernel method, which is robust with respect to dimensionality, to encode the dependencies and hence, it can handle problems with arbitrarily high-dimensional data. Furthermore, the proposed method does not need any prior information about the dependence structure of the RVs; thus, it is applicable to general novelty detection problems with no simplifying assumption. To test the performance of the proposed method, we apply it to realistic application problems for analyzing sensor networks and compare the results to those obtained by peer methods. |
| first_indexed | 2024-09-23T13:20:53Z |
| format | Article |
| id | mit-1721.1/126504 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T13:20:53Z |
| publishDate | 2020 |
| publisher | Elsevier BV |
| record_format | dspace |
| spelling | mit-1721.1/1265042022-10-01T14:42:18Z Kernel dependence analysis and graph structure morphing for novelty detection with high-dimensional small size data set Mohammadi Ghazi Mahalleh, Reza Welsch, Roy E Buyukozturk, Oral Massachusetts Institute of Technology. Department of Civil and Environmental Engineering Sloan School of Management In this study, we propose a new approach for novelty detection that uses kernel dependence techniques for characterizing the statistical dependencies of random variables (RV) and use this characterization as a basis for making inference. Considering the statistical dependencies of the RVs in multivariate problems is an important challenge in novelty detection. Ignoring these dependencies, when they are strong, may result in inaccurate inference, usually in the form of high false positive rates. Previously studied methods, such as graphical models or conditional classifiers, mainly use density estimation techniques as their main learning element to characterize the dependencies of the relevant RVs. Therefore, they suffer from the curse of dimensionality which makes them unable to handle high-dimensional problems. The proposed method, however, avoids using density estimation methods, and rather, employs a kernel method, which is robust with respect to dimensionality, to encode the dependencies and hence, it can handle problems with arbitrarily high-dimensional data. Furthermore, the proposed method does not need any prior information about the dependence structure of the RVs; thus, it is applicable to general novelty detection problems with no simplifying assumption. To test the performance of the proposed method, we apply it to realistic application problems for analyzing sensor networks and compare the results to those obtained by peer methods. 2020-08-06T21:26:25Z 2020-08-06T21:26:25Z 2020-09 2019-12 2020-08-04T17:00:22Z Article http://purl.org/eprint/type/JournalArticle 0888-3270 https://hdl.handle.net/1721.1/126504 Mohammadi-Ghazi, Reza et al. "Kernel dependence analysis and graph structure morphing for novelty detection with high-dimensional small size data set." Mechanical Systems and Signal Processing 143 (September 2020): 106775 © 2020 Elsevier Ltd en http://dx.doi.org/10.1016/j.ymssp.2020.106775 Mechanical Systems and Signal Processing Creative Commons Attribution-NonCommercial-NoDerivs License http://creativecommons.org/licenses/by-nc-nd/4.0/ application/pdf Elsevier BV Prof. Buyukozturk via Elizabeth Soergel |
| spellingShingle | Mohammadi Ghazi Mahalleh, Reza Welsch, Roy E Buyukozturk, Oral Kernel dependence analysis and graph structure morphing for novelty detection with high-dimensional small size data set |
| title | Kernel dependence analysis and graph structure morphing for novelty detection with high-dimensional small size data set |
| title_full | Kernel dependence analysis and graph structure morphing for novelty detection with high-dimensional small size data set |
| title_fullStr | Kernel dependence analysis and graph structure morphing for novelty detection with high-dimensional small size data set |
| title_full_unstemmed | Kernel dependence analysis and graph structure morphing for novelty detection with high-dimensional small size data set |
| title_short | Kernel dependence analysis and graph structure morphing for novelty detection with high-dimensional small size data set |
| title_sort | kernel dependence analysis and graph structure morphing for novelty detection with high dimensional small size data set |
| url | https://hdl.handle.net/1721.1/126504 |
| work_keys_str_mv | AT mohammadighazimahallehreza kerneldependenceanalysisandgraphstructuremorphingfornoveltydetectionwithhighdimensionalsmallsizedataset AT welschroye kerneldependenceanalysisandgraphstructuremorphingfornoveltydetectionwithhighdimensionalsmallsizedataset AT buyukozturkoral kerneldependenceanalysisandgraphstructuremorphingfornoveltydetectionwithhighdimensionalsmallsizedataset |