A NOVEL AND HYBRID WHALE OPTIMIZATION WITH RESTRICTED CROSSOVER AND MUTATION BASED FEATURE SELECTION METHOD FOR ANXIETY AND DEPRESSION.

BACKGROUND

Anxiety and depression are two leading human psychological disorders. In this work, several swarm intelligence-based metaheuristic techniques have been employed to find an optimal feature set for the diagnosis of these two human psychological disorders.

SUBJECTS AND METHODS

To diagnose depression and anxiety among people, a random dataset comprising 1128 instances and 46 attributes has been considered and examined. The dataset was collected and compiled manually by visiting the number of clinics situated in different cities of Haryana (one of the states of India). Afterwards, nine emerging meta-heuristic techniques (Genetic algorithm, binary Grey Wolf Optimizer, Ant Colony Optimization, Particle Swarm Optimization, Artificial Bee Colony, Firefly Algorithm, Dragonfly Algorithm, Bat Algorithm and Whale Optimization Algorithm) have been employed to find the optimal feature set used to diagnose depression and anxiety among humans. To avoid local optima and to maintain the balance between exploration and exploitation, a new hybrid feature selection technique called Restricted Crossover Mutation based Whale Optimization Algorithm (RCM-WOA) has been designed.

RESULTS

The swarm intelligence-based meta-heuristic algorithms have been applied to the datasets. The performance of these algorithms has been evaluated using different performance metrics such as accuracy, sensitivity, specificity, precision, recall, f-measure, error rate, execution time and convergence curve. The rate of accuracy reached utilizing the proposed method RCM-WOA is 91.4%.

CONCLUSION

Depression and Anxiety are two critical psychological disorders that may lead to other chronic and life-threatening human disorders. The proposed algorithm (RCM-WOA) was found to be more suitable compared to the other state of art methods.