Inflammatory Bowel Disease Classification Using Neural Network and Support Vector Machine
| Main Authors: | Asmaa Mohamed Hassan, Yasser M. K. Omar |
|---|---|
| Format: | Article eJournal |
| Terbitan: |
, 2020
|
| Online Access: |
https://zenodo.org/record/3712094 |
Daftar Isi:
- Due to the importance and difficulty of diagnosis of Inflammatory Bowel Disease (IBD) computer-aided diagnosis plays an essential role in the early detection of the IBD. In this paper we explore the feasibility of Machine Learning (ML) algorithms for classification of the activation of IBD and its subtypes Chon's Disease (CD) or Ulcerative Disease (UC). Two commonly used ML algorithms were applied: The Artificial Neural Network (ANN) and the Support Vector Machine (SVM). A detailed experiment is conducted to evaluate the classification capability of each algorithm. Moreover, using different datasets containing normal and IBD patients, the experiment result reveals that ANN and SVM with linear kernel have comparable performance in classification with accuracy of 80% and 79.9% respectively; however, SVM with Radial Basis Function (RBF) kernel outperforms ANN with accuracy of 70.9% and 67% respectively. ANN performance may become much worse by increasing the training samples and the ANN number of layers.