A Statistical Approach To Classify The Leukemia Patients From Generic Gene Features

Classifying the Leukemia patient is very important in clinical diagnosis and further treatment. Leukemia cancer is classified based on how fast it is spreading and has four different types. A gene is the most vital physical and functional factor of the human body. Genes are formed up of DNA. It differs in size for individuals from a few hundred Chromosome bases to higher than 2 million bases. Every person has two copies of each gene, one inherited from each parent. Most of them are same in human, but a small number of genes about 1 percent of the total are a bit different. The study of Gene feature series of the different patient helps to understand the risk that is associated with the various types of cancer. In this paper, we used the classification algorithm, Naive Bayes, to classify the Leukemia patient from gene feature and predicting if a patient is positive or negative. We used NCBI GEO dataset and normalized the data set using our proposed methodology. Which works based on similarity score to find the fuzzy table. Then using the graph approach and our statistical model we were able to successfully predict the result which matches our actual data set with a high accuracy rate.