Effect of missing value methods on Bayesian network classification of hepatitis data
Missing value imputation methods are widely used in solving missing value problems during statistical analysis. For classification tasks, these imputation methods can affect the accuracy of the Bayesian network classifiers. This paper study’s the effect of missing value treatment on the prediction...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English English |
| Published: |
Sysbase Solution
2013
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/30217/1/Effect%20of%20missing%20value%20methods%20on%20Bayesian%20network%20classification%20of%20hepatitis%20data.pdf |
| _version_ | 1825947663359541248 |
|---|---|
| author | Nazziwa Aisha, Adam, Mohd. Bakri Shohaimi, Shamarina |
| author_facet | Nazziwa Aisha, Adam, Mohd. Bakri Shohaimi, Shamarina |
| author_sort | Nazziwa Aisha, |
| collection | UPM |
| description | Missing value imputation methods are widely used
in solving missing value problems during statistical analysis. For classification tasks, these imputation methods can affect the accuracy of the Bayesian network classifiers. This paper study’s the effect of missing value treatment on the prediction accuracy of four Bayesian network classifiers used to predict death in acute
chronic Hepatitis patients. Missing data was imputed using nine methods which include, replacing with most common attribute,support vector machine imputation (SVMI), K-nearest neighbor (KNNI), Fuzzy K-means Clustering (FKMI), K-means Clustering Imputation (KMI), Weighted imputation with K-Nearest Neighbor (WKNNI), regularized expectation maximization (EM), singular value decomposition (SVDI), and local least squares imputation (LLSI). The classification accuracy of the naive Bayes (NB), tree augmented naive Bayes (TAN), boosted augmented naive Bayes (BAN) and general Bayes network classifiers (GBN)were recorded. The SVMI and LLSI methods improved the classification accuracy of the classifiers. The method of ignoring missing values was better than seven of the imputation methods.
Among the classifiers, the TAN achieved the best average
classification accuracy of 86.3% followed by BAN with 85.1%. |
| first_indexed | 2024-03-06T08:16:50Z |
| format | Article |
| id | upm.eprints-30217 |
| institution | Universiti Putra Malaysia |
| language | English English |
| last_indexed | 2024-03-06T08:16:50Z |
| publishDate | 2013 |
| publisher | Sysbase Solution |
| record_format | dspace |
| spelling | upm.eprints-302172015-12-07T03:52:35Z http://psasir.upm.edu.my/id/eprint/30217/ Effect of missing value methods on Bayesian network classification of hepatitis data Nazziwa Aisha, Adam, Mohd. Bakri Shohaimi, Shamarina Missing value imputation methods are widely used in solving missing value problems during statistical analysis. For classification tasks, these imputation methods can affect the accuracy of the Bayesian network classifiers. This paper study’s the effect of missing value treatment on the prediction accuracy of four Bayesian network classifiers used to predict death in acute chronic Hepatitis patients. Missing data was imputed using nine methods which include, replacing with most common attribute,support vector machine imputation (SVMI), K-nearest neighbor (KNNI), Fuzzy K-means Clustering (FKMI), K-means Clustering Imputation (KMI), Weighted imputation with K-Nearest Neighbor (WKNNI), regularized expectation maximization (EM), singular value decomposition (SVDI), and local least squares imputation (LLSI). The classification accuracy of the naive Bayes (NB), tree augmented naive Bayes (TAN), boosted augmented naive Bayes (BAN) and general Bayes network classifiers (GBN)were recorded. The SVMI and LLSI methods improved the classification accuracy of the classifiers. The method of ignoring missing values was better than seven of the imputation methods. Among the classifiers, the TAN achieved the best average classification accuracy of 86.3% followed by BAN with 85.1%. Sysbase Solution 2013-06 Article PeerReviewed application/pdf en http://psasir.upm.edu.my/id/eprint/30217/1/Effect%20of%20missing%20value%20methods%20on%20Bayesian%20network%20classification%20of%20hepatitis%20data.pdf Nazziwa Aisha, and Adam, Mohd. Bakri and Shohaimi, Shamarina (2013) Effect of missing value methods on Bayesian network classification of hepatitis data. International Journal of Computer Science and Telecommunications, 4 (6). pp. 8-12. ISSN 2047-3338 http://www.ijcst.org/Volume4/Issue6/ English |
| spellingShingle | Nazziwa Aisha, Adam, Mohd. Bakri Shohaimi, Shamarina Effect of missing value methods on Bayesian network classification of hepatitis data |
| title | Effect of missing value methods on Bayesian network classification of hepatitis data |
| title_full | Effect of missing value methods on Bayesian network classification of hepatitis data |
| title_fullStr | Effect of missing value methods on Bayesian network classification of hepatitis data |
| title_full_unstemmed | Effect of missing value methods on Bayesian network classification of hepatitis data |
| title_short | Effect of missing value methods on Bayesian network classification of hepatitis data |
| title_sort | effect of missing value methods on bayesian network classification of hepatitis data |
| url | http://psasir.upm.edu.my/id/eprint/30217/1/Effect%20of%20missing%20value%20methods%20on%20Bayesian%20network%20classification%20of%20hepatitis%20data.pdf |
| work_keys_str_mv | AT nazziwaaisha effectofmissingvaluemethodsonbayesiannetworkclassificationofhepatitisdata AT adammohdbakri effectofmissingvaluemethodsonbayesiannetworkclassificationofhepatitisdata AT shohaimishamarina effectofmissingvaluemethodsonbayesiannetworkclassificationofhepatitisdata |