Principal Component Analysis of Process Datasets with Missing Values
Datasets with missing values arising from causes such as sensor failure, inconsistent sampling rates, and merging data from different systems are common in the process industry. Methods for handling missing data typically operate during data pre-processing, but can also occur during model building....
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020
|
| Online Access: | https://hdl.handle.net/1721.1/125630 |
| _version_ | 1826194838162243584 |
|---|---|
| author | Severson, Kristen Molaro, Mark Braatz, Richard D |
| author2 | Massachusetts Institute of Technology. Department of Chemical Engineering |
| author_facet | Massachusetts Institute of Technology. Department of Chemical Engineering Severson, Kristen Molaro, Mark Braatz, Richard D |
| author_sort | Severson, Kristen |
| collection | MIT |
| description | Datasets with missing values arising from causes such as sensor failure, inconsistent sampling rates, and merging data from different systems are common in the process industry. Methods for handling missing data typically operate during data pre-processing, but can also occur during model building. This article considers missing data within the context of principal component analysis (PCA), which is a method originally developed for complete data that has widespread industrial application in multivariate statistical process control. Due to the prevalence of missing data and the success of PCA for handling complete data, several PCA algorithms that can act on incomplete data have been proposed. Here, algorithms for applying PCA to datasets with missing values are reviewed. A case study is presented to demonstrate the performance of the algorithms and suggestions are made with respect to choosing which algorithm is most appropriate for particular settings. An alternating algorithm based on the singular value decomposition achieved the best results in the majority of test cases involving process datasets. Keywords: principal component analysis; missing data; process data analytics; chemometrics; machine learning; multivariable statistical process control; process monitoring; Tennessee Eastman problem |
| first_indexed | 2024-09-23T10:02:56Z |
| format | Article |
| id | mit-1721.1/125630 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T10:02:56Z |
| publishDate | 2020 |
| publisher | MDPI AG |
| record_format | dspace |
| spelling | mit-1721.1/1256302022-09-26T15:24:10Z Principal Component Analysis of Process Datasets with Missing Values Severson, Kristen Molaro, Mark Braatz, Richard D Massachusetts Institute of Technology. Department of Chemical Engineering Datasets with missing values arising from causes such as sensor failure, inconsistent sampling rates, and merging data from different systems are common in the process industry. Methods for handling missing data typically operate during data pre-processing, but can also occur during model building. This article considers missing data within the context of principal component analysis (PCA), which is a method originally developed for complete data that has widespread industrial application in multivariate statistical process control. Due to the prevalence of missing data and the success of PCA for handling complete data, several PCA algorithms that can act on incomplete data have been proposed. Here, algorithms for applying PCA to datasets with missing values are reviewed. A case study is presented to demonstrate the performance of the algorithms and suggestions are made with respect to choosing which algorithm is most appropriate for particular settings. An alternating algorithm based on the singular value decomposition achieved the best results in the majority of test cases involving process datasets. Keywords: principal component analysis; missing data; process data analytics; chemometrics; machine learning; multivariable statistical process control; process monitoring; Tennessee Eastman problem 2020-06-02T18:39:46Z 2020-06-02T18:39:46Z 2017-07 2017-05 2019-08-14T18:20:09Z Article http://purl.org/eprint/type/JournalArticle 2227-9717 https://hdl.handle.net/1721.1/125630 Severson, Kristen et al. “Principal Component Analysis of Process Datasets with Missing Values.” Processes 5, 4 (July 2017): 38. © 2017 The Authors en http://dx.doi.org/10.3390/pr5030038 Processes Creative Commons Attribution 4.0 International license https://creativecommons.org/licenses/by/4.0/ application/pdf MDPI AG MDPI |
| spellingShingle | Severson, Kristen Molaro, Mark Braatz, Richard D Principal Component Analysis of Process Datasets with Missing Values |
| title | Principal Component Analysis of Process Datasets with Missing Values |
| title_full | Principal Component Analysis of Process Datasets with Missing Values |
| title_fullStr | Principal Component Analysis of Process Datasets with Missing Values |
| title_full_unstemmed | Principal Component Analysis of Process Datasets with Missing Values |
| title_short | Principal Component Analysis of Process Datasets with Missing Values |
| title_sort | principal component analysis of process datasets with missing values |
| url | https://hdl.handle.net/1721.1/125630 |
| work_keys_str_mv | AT seversonkristen principalcomponentanalysisofprocessdatasetswithmissingvalues AT molaromark principalcomponentanalysisofprocessdatasetswithmissingvalues AT braatzrichardd principalcomponentanalysisofprocessdatasetswithmissingvalues |