Effect of missing data on multitask prediction methods
Abstract There has been a growing interest in multitask prediction in chemoinformatics, helped by the increasing use of deep neural networks in this field. This technique is applied to multitarget data sets, where compounds have been tested against different targets, with the aim of developing model...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2018-05-01
|
| Series: | Journal of Cheminformatics |
| Subjects: | |
| Online Access: | http://link.springer.com/article/10.1186/s13321-018-0281-z |
| _version_ | 1818941570067464192 |
|---|---|
| author | Antonio de la Vega de León Beining Chen Valerie J. Gillet |
| author_facet | Antonio de la Vega de León Beining Chen Valerie J. Gillet |
| author_sort | Antonio de la Vega de León |
| collection | DOAJ |
| description | Abstract There has been a growing interest in multitask prediction in chemoinformatics, helped by the increasing use of deep neural networks in this field. This technique is applied to multitarget data sets, where compounds have been tested against different targets, with the aim of developing models to predict a profile of biological activities for a given compound. However, multitarget data sets tend to be sparse; i.e., not all compound-target combinations have experimental values. There has been little research on the effect of missing data on the performance of multitask methods. We have used two complete data sets to simulate sparseness by removing data from the training set. Different models to remove the data were compared. These sparse sets were used to train two different multitask methods, deep neural networks and Macau, which is a Bayesian probabilistic matrix factorization technique. Results from both methods were remarkably similar and showed that the performance decrease because of missing data is at first small before accelerating after large amounts of data are removed. This work provides a first approximation to assess how much data is required to produce good performance in multitask prediction exercises. |
| first_indexed | 2024-12-20T06:57:38Z |
| format | Article |
| id | doaj.art-fed18e8d4d6b4b2fa56f2d13e718a76a |
| institution | Directory Open Access Journal |
| issn | 1758-2946 |
| language | English |
| last_indexed | 2024-12-20T06:57:38Z |
| publishDate | 2018-05-01 |
| publisher | BMC |
| record_format | Article |
| series | Journal of Cheminformatics |
| spelling | doaj.art-fed18e8d4d6b4b2fa56f2d13e718a76a2022-12-21T19:49:18ZengBMCJournal of Cheminformatics1758-29462018-05-0110111210.1186/s13321-018-0281-zEffect of missing data on multitask prediction methodsAntonio de la Vega de León0Beining Chen1Valerie J. Gillet2Information School, University of SheffieldDepartment of Chemistry, University of SheffieldInformation School, University of SheffieldAbstract There has been a growing interest in multitask prediction in chemoinformatics, helped by the increasing use of deep neural networks in this field. This technique is applied to multitarget data sets, where compounds have been tested against different targets, with the aim of developing models to predict a profile of biological activities for a given compound. However, multitarget data sets tend to be sparse; i.e., not all compound-target combinations have experimental values. There has been little research on the effect of missing data on the performance of multitask methods. We have used two complete data sets to simulate sparseness by removing data from the training set. Different models to remove the data were compared. These sparse sets were used to train two different multitask methods, deep neural networks and Macau, which is a Bayesian probabilistic matrix factorization technique. Results from both methods were remarkably similar and showed that the performance decrease because of missing data is at first small before accelerating after large amounts of data are removed. This work provides a first approximation to assess how much data is required to produce good performance in multitask prediction exercises.http://link.springer.com/article/10.1186/s13321-018-0281-zMultitask predictionSparse data setsMissing dataDeep neural networksMacau |
| spellingShingle | Antonio de la Vega de León Beining Chen Valerie J. Gillet Effect of missing data on multitask prediction methods Journal of Cheminformatics Multitask prediction Sparse data sets Missing data Deep neural networks Macau |
| title | Effect of missing data on multitask prediction methods |
| title_full | Effect of missing data on multitask prediction methods |
| title_fullStr | Effect of missing data on multitask prediction methods |
| title_full_unstemmed | Effect of missing data on multitask prediction methods |
| title_short | Effect of missing data on multitask prediction methods |
| title_sort | effect of missing data on multitask prediction methods |
| topic | Multitask prediction Sparse data sets Missing data Deep neural networks Macau |
| url | http://link.springer.com/article/10.1186/s13321-018-0281-z |
| work_keys_str_mv | AT antoniodelavegadeleon effectofmissingdataonmultitaskpredictionmethods AT beiningchen effectofmissingdataonmultitaskpredictionmethods AT valeriejgillet effectofmissingdataonmultitaskpredictionmethods |