Classication of SNPs for breast cancer diagnosis using neural-network-based association rules

Medical diagnosis is a major area of current research in Machine Learning and Data Mining. Single Nucleotide Polymorphisms (SNPs) are an important source of the human genome variability and have thus been implicated in several human diseases, including cancer. Breast Cancer is the most common malignant tumour for women and has known a large spread during the past twenty years. To separate the tumorous samples from the normal ones, various techniques have been applied on SNP data. One of the major problems related to SNP data is the high number of features which makes the task of classification complex. In this paper, a new hybrid intelligent technique based on Association Rule Mining (ARM) and Neural Networks (NN) which uses an Evolutionary Algorithm (EA) is proposed to deal with the dimensionality problem for the diagnosis of breast cancer. ARM optimized by Grammatical Evolution (GE) is used to select the most informative features and reduce the dimensionality by extracting associations between SNPs, while NN is used for efficient classification. The proposed NN-GEARM approach has been applied on breast cancer SNP dataset obtained from the NCBI Gene Expression Omnibus (GEO) website. The created model has reached an accuracy of up to 90%.