Classification of Immunosignature Using Random Forests for Cancer Diagnosis
The non-invasive cancer diagnosis can be considered as one of the most feasible challenges of ground-breaking medicine. Cancer has been characterized as a miscellaneous disease consisting of numerous disparate subtypes. Subsequently, diagnosing and combating cancer is extremely significant. The sign...
| Main Authors: | , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/48060/ http://irep.iium.edu.my/48060/ http://irep.iium.edu.my/48060/1/ID_134.pdf |
| Summary: | The non-invasive cancer diagnosis can be considered as one of the most feasible challenges of ground-breaking medicine. Cancer has been characterized as a miscellaneous disease consisting of numerous disparate subtypes. Subsequently, diagnosing and combating cancer is extremely significant. The significance of classifying cancer patients has led numerous research parties, from the bioinformatics and the biomedical domains, to rout out the enforcement of data mining methods. The fundamental target of data mining and machine learning is to achieve efficacious cancer classification mechanisms which provide considerable and trustworthy classification accuracy. To attain this essential research purpose, a minimum set of genes that can assure higher performance in classification using data mining algorithms need to be detected. The evolution of authoritative immunofingerprint mining technology is exerting a growing influence on comprehensive cancer diagnosis biology. In this work, we will develop a robust classification model that can be utilized in cancer diagnosis using immunofingerprint data. We have used the Random Subset gene selection method to avoid overfitting and improve model performance in order to make the input data suitable for the classification stage, which has been implemented using the Random Forest (RF) classifier. This novel model has been examined in diagnosing and classifying cancer over two benchmark cancer datasets. Altogether, the empirical results show that the combination of random subset reduction technique with the Random Forest classification method offers a promising tool that is masterful in the diagnosis of carcinoma. |
|---|