Hybrid artificial neural network for data classification problem
Jaspreet Kaur, Ashima Kalra · 2017
The benchmarking databases for artificial neural network (ANN) include several datasets from several different domains. All datasets exhibit feasible problems which could be called diagnosis jobs and all but one contain genuine world data. Two such standard problems, for categorization are taken in this paper to analyze the capability of intelligent water drop (IWD), particle swarm optimization (PSO) and hybrid IWD-PSO with ANN. In this work, SI algorithm is tested on a set of two benchmark functions. Further a comparison is made between Swarm intelligence algorithm-ANN in terms of sum square error, Elapsed time. The research is chosen for finding primary weights and biases for an artificial neural network. The amalgamation of swarm intelligence (SI) optimization and ANN greatly help in quick convergence of ANN in classification to various benchmark problems. The result shows that utilization of swarm intelligence minimizes the classification error.