1. Dr. R. KAVITHA - Associate Professor, Department of Computer Science and Engineering, Dr. MGR Educational and
Research Institute, Chennai.
2. Dr. R. RAJESWARI - Associate Professor, Department of Computer Science, Dr. MGR Educational and Research Institute, Chennai.
3. Dr. P. G. SIVAGAMINATHAN - Professor, Department of Computer Science and Engineering, Koneru Lakshmaiah Educational Foundation,Vaddeswaram, Andhra Pradesh.
4. Dr. A. NITHYA - Assistant Professor, Department of B.Com (CA), PSGR Krishnammal College for Women, Coimbatore.
Neurological disorders are clinically specified as infirmity that affects the mind,spinal cord. and the nerves at other parts the human body. The Structural, biochemical or electrical abnormalities in the mind, spinal cord or other nerves can result in quite a number of signs and symptoms. Regardless of its numerous perspectives,side effects, signs and effects, various studies have been made to distinguish the main cause of the problem. Genetic Association Studies have been a popular approach for assessing the association between common Single Nucleotide Polymorphisms (SNPs) and complex diseases. However, other genomic data involved in the Genetic Association Studies, for example, gene expressions, are usually neglected in these association studies. Recently, many Machine Learning algorithms have been utilized to identify the significant of SNP. The curseof dimensionality is the main challenge. On the otherhand, the number of samples is decidedly smaller than the number of SNPs. In addition, the number ofhealthy and patient samples can be unequal. These challenges make the feature selection and classification very difficult. Therefore, an efficient method is proposed to identify significant SNPs andclassify healthy and patient samples. Searching for the (sub) optimal subset of features is a Nondeterministic Polynomial Time (NP) hard problem. In this regard, firstly, the Mean Encoding, as an intelligent method, is utilized to convert the nominal SNP data to numeric. Then a Binary Swallow Swarm Optimization (BSSO) method is used for feature selection, which removes the irrelevant and redundant features. The binarization of the continuous swallow swarm meta-heuristics iscarried out using a special function. Finally, theproposed Deep AutoEncoder Based DataClustering (DAEDC) algorithm is employed to classify so that it can construct its structure based on input data, automatically. To evaluate, apply the proposed approach to mental retardation SNPdataset, which obtained from the Gene Expression Omnibus (GEO) dataset. The proposed method has given higher results in terms of precision, recall, F-measure, and accuracy in mental retardation, and autism. The results indicate that it has succeeded with high efficiency, comparedwith other classifiers.
Single Nucleotide Polymorphism (SNP), Feature selection, Complex diseases, BinarySwallow Swarm Optimization (BSSO), Deep learning, and Deep Auto-Encoder Based Data Clustering (DAEDC).