Thema der Dissertation:
Computational analysis of cancer transcriptomes: drug response prediction in colorectal cancer and gene regulatory networks and long non-coding genes in medulloblastoma Thema der Disputation:
Feature selection in computational analysis of cancer transcriptomes
Computational analysis of cancer transcriptomes: drug response prediction in colorectal cancer and gene regulatory networks and long non-coding genes in medulloblastoma Thema der Disputation:
Feature selection in computational analysis of cancer transcriptomes
Abstract: In cancer research, machine learning is applied to solve relevant research questions and tasks by the computational analysis of high-dimensional (big) molecular data of tumors obtained via high-throughput omics technologies such as next-generation sequencing and arrays. Here, cancer transcriptome data offer a rich resource that can be used to address, for example, biomarker detection, molecular subtype identification, and the inference of gene regulatory networks. However, the successful application of machine learning methods highly depends on the selection of informative and expressive features.
During my talk, I will provide an introduction on feature selection methods focusing on methods applied for supervised machine learning. This introduction includes an overview on five types of feature selection methods: filter, embedded, wrapper, hybrid, and ensemble methods. Advantages and disadvantages of the five method types will be discussed.
During my talk, I will provide an introduction on feature selection methods focusing on methods applied for supervised machine learning. This introduction includes an overview on five types of feature selection methods: filter, embedded, wrapper, hybrid, and ensemble methods. Advantages and disadvantages of the five method types will be discussed.
Zeit & Ort
08.02.2022 | 10:00