Feature Selection in Microarray Data Using Entropy Information

Abstract

Researchers in biological sciences and genetics are faced with high-dimensional data, such as the microarray data, and the analysis and proper interpretation of these data are very important in bioinformatics and systems biological sciences. In such types of data, the number of variables, for example, the genes, is many times greater than the number of samples. Therefore, the dimension of the data must be reduced at the primary point. Then, the analysis, for example, clustering, is performed on the compacted data. This process is called data summarization. There are various ways to summarize high-dimensional data, which depends on the nature of the data. The aim of data summarization is to remove unnecessary features so that the data are classified more accurately. Shannon' s entropy information is a common method for clustering genes in microarray data and selecting a set of disease-related genes. This chapter introduces and illustrates statistical inference concepts of entropy in microarray data clustering to select a set of the most important genes associated with a disease.

An Information Entropy Based to Identify Dominant Species for No Biological Data Genes

“…To prevent common errors in statistical modeling, many methods have been introduced, and entropy is one of the most widely used concept in medical and genetic sciences. Entropy was introduced by Nicholas Georgescu-Roegen in 1971 and later developed by scientists based on the principles established by Shannon [8]. Shannon had a major role in introducing entropy information, which has been widely used in high-dimensional studies [9].…”

“…Entropy was introduced by Nicholas Georgescu-Roegen in 1971 and later developed by scientists based on the principles established by Shannon [8]. Shannon had a major role in introducing entropy information, which has been widely used in high-dimensional studies [9]. One of the advantages of entropy is that calculation of values is based on theoretical forms, not the empirical and personal concepts.…”

“…One of the advantages of entropy is that calculation of values is based on theoretical forms, not the empirical and personal concepts. These values give small or large weights, proportional to the small or large actual values [10].…”