Abstract : | In this thesis, we focus on finding and implementing new mixture models for clusteringcancer genomes. Having received real data from 98 cancer patients who have the same typeof cancer (B-cell chronic lymphocytic leukemia also known as chronic lymphoid leukemia(CLL)), we group these cancer patients based transition tables, namely by type of mutations thatmade them. In this way, we manage to find specifics on mutations of some patients, giving thechance for further research and helps doctors to explore in greater depth the problem of cancerand in the future they find a cure. Based on the results that we received from the experiments,90% of cancer patients are grouped in a common cluster, the remaining 10% presents specificson mutations. The implementation does not only affect this problem. This idea can be appliedto many problems involving data sequence. classification.
|
---|