International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 77 - Number 3 |
Year of Publication: 2013 |
Authors: Jogendra Singh Kushwah, Divakar Singh |
10.5120/13373-0977 |
Jogendra Singh Kushwah, Divakar Singh . A Comparative Result Analysis of Human Cancer Diagnosis using Ensemble Classification Methods. International Journal of Computer Applications. 77, 3 ( September 2013), 14-18. DOI=10.5120/13373-0977
Cancer research has been an interesting and challenging research area in the field of medical science. Classification techniques have been found useful in early diagnosis of cancer and better treatment. For diagnosis of cancer various classification methods are used but they suffer with one or more disadvantages. In this paper ensemble based classification methods which combine the prediction of individual classifiers to generate the final prediction are discussed. The methods discussed are Bagging, Boosting and Random Forest Algorithm. These ensemble methods have shown improvement in quality of result as compared to commonly used single classifier e. g. decision tree or neural network . The improvement in classification is however at the cost extra processing time and higher storage as decision tree or neural network are faster as compared to ensemble based techniques. The ideas for further improvement in this field are also discussed in this paper. Methods discussed in the paper are applied on human cancer data for appropriate cancer gene selection which leads to classification of cancer.