A Variation of the Semisupervised Self-organizing Map for Gene Expression Classification

S.-I. Wu and B. Colvin (USA)


Neural Networks, Data Mining, TimeFrequency Analysis, Computing in Science


We have implemented several neural networks in solving classification problems. We have tried the Multi Level Perceptron (MLP), Self-Organization Map (SOM), and Adaptive Subspace Self-Organization Map (ASSOM) systems. We have applied these systems in classification of gene expression, hand-written digits, and artificially generated data sets. In general, ASSOM performs better than MLP and SOM in the sense that ASSOM is more accurate and efficient. In this paper, we focus on reporting the results in mining the gene classification from the time series of gene expressions.

Important Links:

Go Back