Application of unsupervised learning techniques to DNA microarray databases: a comparative case study

We present the application of the unsupervised machine learning clustering techniques K-Means and Fuzzy C-Means to DNA microarray gene expression data from breast cancer tissues. The mathematical basis and the development of the algorithm of the techniques are shown, and the Expectation Maximization technique is presented as an alternative for the assignment of missing data. Results show that the best method for the clustering of the genetic expression data under study is the Fuzzy C-Means algorithm.