A Novel Meta-heuristic Hybrid Machine learning Model for disseminating the factors corresponding to Type II Diabetes with devised clinical trials

ACO that stands for Ant Colony Optimization solves the problem of finding the best path by increasing the number of ants in a weighted graph. In ACO, the ants are moved randomly in the graph to build solution. The movement of the ants in the graph is biased by a pheromone model with the set of parameters and its components in runtime. This research work’s main goal is to develop a decision support model with modified ACO by pheromone generation at each level for feature selection process. Features that are selected are fed to J 48 classifier model as input upon hybridization. During the construction phase of the ants, it evaluates the solution by modification of the trial values and the pheromone for the search of future ants. Experimental results show that the models correctness was found to be $94 \%$ with precision and recall to be $93.51 \%$ and $93.73 \%$ respectively. The model has been evaluated and validated using Fisher’s F-Test statistic with correlation among the attributes to be 0.86 and the observed P-value to be $\lt 0.001$ and ROC as 0.85 accordingly. The developed model can be aided by experts for the determining major risk factors which corresponds to a specific region upon disease management service. Also, the dependency among the risk factors is subjected to likelihood, location and dietary habits of the people.