Open AccessClassification and identification the most important features of cervical cancer based on the expression of microRNA gene with the random forest (RF) algorithm
Author(s) -
E. Aminudin Aziz,
Adi Wibowo,
Panji Wisnu Wirawan
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1217/1/012123
Subject(s) - overfitting , random forest , cervical cancer , cancer , algorithm , identification (biology) , generalization , microrna , statistical classification , computer science , machine learning , medicine , artificial intelligence , gene , mathematics , biology , artificial neural network , genetics , mathematical analysis , botany
Cervical cancer is the leading cause of death women in the world and number one in Indonesia. An effort that can be done for this case is early detection, for example, an IVA test (visual inspection test with acetic acid). However, the IVA test is not able to indicate patients who have potential cancer before cancer’s physical characteristics are seen. Thus a new solution is needed for early detection of cervical cancer that can indicate patients who have potential cancer before cancer’s physical characteristics are seen. In recent years, various types of miRNA that play a role in cancer malignancies have been identified and can be used as non-invasive biomarkers for cancer diagnosis and monitoring. The use of classification based on miRNA gene expression is a solution for early detection, but the use of high accuracy classification algorithms is something that must be considered. Random Forest (RF) algorithm is the solution to these problems because better generalization performance and is less susceptible to overfitting. In this study also identified important features that are very influential in the classification process. The results showed that the Random Forest algorithm was able to have 100% accuracy for classification and most important features supporting the cancer were miR-549c-5p, miR-183 and miR-515-5p.