An Improved Genetic with Particle Swarm Optimization Algorithm Based on Ensemble Classification to Predict Protein–Protein Interaction

Protein–protein interaction plays an important role in biological function. Though protein interaction and non-interaction is a broad field, PPI is considered more important than PPNI. The concentration of dataset with PPNI is also used to predict the protein–protein interaction. False negatives of non-interaction data have to be identified in the non-proven negative genetic interactions. A learning approach of ensemble selection is a “build and select” strategy, where multiple classifiers have to be trained. Diversity and accuracy of the multi-classifier have to be selected to find the solution. In this paper, PPNI datasets are identified from PPI dataset. Three levels of development have been considered such as, Dataset construction carried out by Negatome, Random pair and Recombine pair methods. Feature extraction and feature selection performance can be carried out by the way of N-Gram techniques. Ensemble classification is done by utilizing the classifiers such as Support Vector Machine, Decision Tree, Neural Network and Naive Bayes. For the enhanced optimization algorithm expressed through the search operation, Genetic-PSO algorithm is proposed. The result exposes the reduced false negatives with the process of dataset construction and the execution of the random pair dataset effectively.