Research on Tanimoto Coefficient Similarity Based Mean Shift Gentle Adaptive Boosted Clustering for Genomic Predictive Pattern Analytics

Gene expression data clustering is a significant problem to be resolved as it provides functional relationships of
genes in a biological process. Finding co-expressed groups of genes is a challenging problem. To identify
interesting patterns from the given gene expression data set, a Tanimoto Coefficient Similarity based Mean Shift
Gentle Adaptive Boosted Clustering (TCS-MSGABC) Model is proposed. TCS-MSGABC model comprises
two processes namely feature selection and clustering. In first process, Tanimoto Coefficient Similarity
Measurement based Feature selection (TCSM-FS) is introduced to identify relevant gene features based on the
similarity value for performing the genomic expression clustering. Tanimoto Coefficient Similarity Value
ranges from ‘ ’ to ‘ ’ where ‘ ’ is highest similarity. The gene feature with higher similarity value is taken to
perform clustering process. After feature selection, Mean Shift Gentle Adaptive Boosted Clustering (MSGABC)
algorithm is carried out in TCS-MSGABC model to cluster the similar gene expression data based on the
selected features. The MSGABC algorithm is a boosting method for combining the many weak clustering
results into one strong learner. By this way, the similar gene expression data are clustered with higher accuracy
with minimal time. Experimental evaluation of TCS-MSGABC model is carried out on factors such as
clustering accuracy, clustering time and error rate with respect to number of gene data. The experimental results
show that the TCS-MSGABC model is able to increases the clustering accuracy and also minimizes clustering
time of genomic predictive pattern analytics as compared to state-of-the-art works.