Classification of Primary Open Angle Glaucoma through Genetic and Demographic Data

Primary Open Angle Glaucoma (POAG) is considered to be one of the leading causes of irreversible blindness. More than 66 million people lie in this category from which 50% of people unaware of its adverse effect. To prevent its adverse effects like blindness there is an imperious need for automated technique to be developed for its detection. Recently, a genetic data has been explored with machine learning techniques for the detection and prevention of POAG. The genes along with other demographic data sets give an evident base for the detection of this disease. In this paper both the genetic and demographic data is used for the detection of this disease. Here an algorithm is proposed for preprocessing. The feature sets comprises of genes (i.e. MYOC, CYP1B1, NTF4, OPTN), SNP alleles, risk alleles, chromosomes, family history, race, age and gender of patients. For this paper, we use 590 patients’ genetic and demographic data sets from various online repositories. For the performance evaluation of the proposed approach we have applied different types of classifiers (Naive Bayes, J48, SMO, LWL, K*). The classifiers were evaluated to understand their ability of predicting the desire results on sensitivity, accuracy and specificity parameters. The results revealed that Support Vector Machine (SMO) classifier meet high classification accuracy i.e. 98%.