Recognition of the state of mind using Hamming Swarm Net

Recognition of the state of mind has been attempted by many researchers using different computational techniques, but the achievement is not quite significant. In this paper, “whether a person can cause harm to others or not” has been considered as one of the states of mind of human being to be recognized. This paper presents Hamming Swarm Net (HSN) which is a hybridization the Hamming Net (HN) and the Particle Swarm Optimization (PSO) techniques. HN being a competitive network is usually used tasks like clustering, but in this work it is employed as a tool for pattern classification. HN considers fixed exemplar vectors, which restricts its mapping efficiency. The proposed technique evolves the optimal set of exemplar vectors by exercising intelligence of swarm and significantly improves the recognition capabilities. The moral database of UCI machine learning repository is taken here for recognition one of the states of mind. Simulation study shows that recognition capabilities of HSN is superior in comparison to many other techniques such as Multi Layer Perceptron (MLP), Functional Link Artificial Neural Network (FLANN), Polynomial Neural Network (PNN), Multiple Linear Regression (MLR), and Hamming Network (HN).