A neural network guided clustering algorithm for unsupervised microarray data analysis
MetadataShow full item record
Microarrays are a powerful technology for the genome-wide analysis of gene expression and behavior, opening up new horizons in the research of molecular and cellular biology. There has been enormous amount of data generated by researchers in an attempt to understand the complex mechanism of life. The applications of this technology in cancer research has been phenomenal in identifying new cancer types, newer drug targets, effects of drugs and in determining the underlying causes of tumor. The overload of information and the complexity of the data have led microarray technologists, biologists, statisticians, data analysts and computer scientists to come together and devise efficient models to interpret the data. Sophisticated computational power availability has further contributed the quest for improvised methods to mine data. Owing to recent advancement in microarray technology, the whole human genome structure has been unraveled within a short span of time. In effect, gene annotations and functional genomics were available and a lot of research has been carried out in the area of supervised or annotated data analysis. Much of the current research is still unsupervised or exploratory, given the complexity of biological design, without a priori information available. Clustering is usually the answer in excavating useful conclusions from this class of data and is a very important aspect in current microarray analysis. This thesis work is an attempt to improve clustering by using artificial neural networks, a machine learning technique. Artificial Neural Networks have been implemented to improvise the classic K -Means Clustering Algorithm.