Kernel Methods for Genomic Data
- Friday, May 8, 2020 from 10:00am to 11:00am
Kernel methods are always used in machine learning to map the data to new feature space for classification purposes. For genomic data, kernels can be used to extract information from the existing formats. The use of kernel methods on genomic data is not very frequent, but there are some studies with promising outcomes. Some of the kernels used for genomic data are string kernels, graph kernels, and tree kernels. Out of them, many studies use string kernel to compare genomic sequences as string kernels can capture the similarities between sequences that cannot be found with sequence alignment methods. Due to this growing interest, this report summarizes some of the promising methods that use kernel methods for genomic data. It is apparent after the analysis that there are more opportunities to expand further and develop kernel methods for genomic data.
- Gianforte School of Computing