Date |
Nov 21, 2017 |
Speaker |
Dr. Limin Li, Associate professor, Xi'an Jiaotong University
|
Title |
Multi-view learning models and the application in cancer subtype discovery
|
Abstract |
Multi-view learning aims to integrate multiple data information from
different views to improve the learning performance.
In this talk, I will first present our recent work on multi-view
models by uniform projection,which can be used to do clustering with
different data sources. Two multi-view models including uniform
multidimensional scaling (UMDS) and uniform class assignment (UCA) are
presented. The experiments on several real world datasets show the
effectiveness.
I will then present our work on breast cancer subtype discovery by
integrating different genomic datasets in TCGA. A multi-view
clustering approach with enhanced consensus (ECMC) is proposed for the
case when the consensus information shared by different views is
relatively weak. The survival analysis shows that our ECMC model
outperforms other methods when identifying cancer subtypes. By
Fisher’s combination test method, we found that three computed
subtypes roughly correspond to three known breast cancer subtypes
including luminal B, HER2 and basal-like subtypes.
|
|