Genetic Algorithm-Based Feature Selection and Self-Organizing Auto-Encoder (SOAE) For SNP Genomics Data Classification

SNPs (single nucleotide polymorphisms) are the most common biomarkers, and they have a wide range of applications, especially in human genetics. The high cost of genotyping many SNPs is one of the most significant obstacles in SNP analysis. Selection and classification of characteristics are particularly tough due to these challenges. Therefore, a method for detecting significant SNPs and distinguishing healthy and ill samples has been developed. Finding the (sub) optimal subset of attributes is a Nondeterministic Polynomial Time (NP) hard problem. To begin, the nominal SNP data is transformed to numeric using an intelligence method called Mean Encoding. After that, a Genetic Algorithm (GA)-based feature selection is applied, which selects the important traits while discarding the redundant ones. Finally, the recommended Self-organizing Auto-Encoder (SOAE) is utilized to categorize, allowing it to create its structure autonomously based on the input thyroid cancer. The proposed strategy is applied to a thyroid cancer dataset from the Gene Expression Omnibus (GEO) database to see how well it works. When compared to other methodologies using measurements like accuracy, recall, and F-measure, the results suggest that it performed well.