Modeling of TST Using Biological Signals and KL-Hausdorff Distance Similarity Measure for Gene Based Parkinson’s Disease Recognition

Abstract

Genetic research experienced drastic transformation since past decades, which benefits the biological area eventually for the detection of neurodegenerative ailment like Parkinson’s disease (PD). Recently, rigorous investigate had been conceded out for of PDs detection instigated through-sequence -and recessive auto-somal-of dominant-genes such as PARK2, LRRK2, SNCA, PARK7 and PINK1. Several inherent based similarity degree representations such as Cosine similarity and Hamming Distance model were introduced for the detection of these genes. However, these representations detect 2 to 3 gene sequence barely by maximum Root Mean Square Error (RMSE) and minimum accuracy rate. The ratio of misclassification is too great for prevailing scheme. To perceive PD through low RMSE and high accuracy a Kullback-Leibler Hausdroff distance (KL-H) similarity measure model is proposed so as to discover the affected patient pattern efficiently. It works in two phases, in first, protein sequence of amino acid is determined with the use of model transcription, splicing and translation (TST). The second stage in turn distinguish PD that depends on the model of similarity measure which comprise assessment of template sequence and specified sequence with the use of Hausdorff distance and KL-distance process. The property of nucleotide density in KL distance measure algorithm was employed. The result analysis and comparative study were presented among the proposed and existing system. We attained maximum accuracy of 88%, with sensitivity 67.86%, specificity 93.81%, precision 76%, F1 score 71.69%, minimum RMSE (12%) and FPR (6.19%)in comparison to the prevailing similarity measurement model.