Bayesian Random Forest for the Classification of High-dimensional mRNA Cancer Samples Oyebayo Ridwan Olaniran*, Mohd Asrul Affendi Bin Abdullah Department of Mathematics and Statistics, Faculty of Applied Sciences and Technology, Universiti Tun Hussein Onn Malaysia, Pagoh Educational Hub, 84600 Pagoh, Johor, Malaysia.
[email protected];
[email protected] iCMS 2017, Langkawi, Malaysia.
November 8, 2017
Oyebayo Ridwan Olaniran*, Mohd Asrul Affendi Bin Bayes Abdullah Random (UTHM) Classification Forest
November 8, 2017
1 / 30
Outline 1
Overview General Introduction
2
Random Forest
3
Bayesian Random Forests
4
Application to Cancer Data Data Calibration
5
Results
6
Results
7
Conclusion
8
Funding
Oyebayo Ridwan Olaniran*, Mohd Asrul Affendi Bin Bayes Abdullah Random (UTHM) Classification Forest
November 8, 2017
2 / 30
Overview
General Introduction
Introduction
Recent findings reveal that various cancer types can be diagnosed using non-clinical approach which involves monitoring of the biological samples using their genes expression profiles. However, this advancement is possible due to the enhancement of microarray technology which made it possible to observe gene expression levels of several gene chips concurrently (Olaniran et al., 2016).
Oyebayo Ridwan Olaniran*, Mohd Asrul Affendi Bin Bayes Abdullah Random (UTHM) Classification Forest
November 8, 2017
3 / 30
Overview
General Introduction
A Typical Cancerous and Normal Cells in Human Colon
Figure: Colon cancer cells and staging in human
Oyebayo Ridwan Olaniran*, Mohd Asrul Affendi Bin Bayes Abdullah Random (UTHM) Classification Forest
November 8, 2017
4 / 30
Overview
General Introduction
Introduction contd’
Several authors have discussed the health benefits of the non-clinical diagnosis breakthrough, but the major problem that still exists is how to adequately identify the few subsets of thousands genes whose information can be used to reliably classify the mRNA (messenger Ribonucleic acid) samples into their respective biological groups. In addition, it has been observed that adequacy of any method strongly depends on the health problems (Sim et al., 2012).
Oyebayo Ridwan Olaniran*, Mohd Asrul Affendi Bin Bayes Abdullah Random (UTHM) Classification Forest
November 8, 2017
5 / 30
Overview
General Introduction
Introduction contd’
Many types of machine learning algorithms have been proposed to perform the task of non-clinical diagnosis mRNA. mRNA samples are usually collected on several biological features (genes). The resulting data structure are of the form n
