Application Of Neural Network Trained By Particle Swarm Optimization Algorithm

Data mining technology is used to help people finding the information and knowledge in great capacity of data. It has become the core technology of the intelligence in recent decades. It has also been widely used in many areas and drawn the attention of the whole academe. It is very important to analyze the learning capability of students in education area. The learning-capability analysis based on the using of classification in data mining technology has practical meaning to improve teaching. Data classification is a partitioning process according to the character of a set of data. It has been widely used in such areas as statistics , machine learning system , neural network and expert system . In fact , classification can be divided into two steps . The first step is constructing a model which can describe specific data set or conception set . The second step is assessing the prediction accurate rate of model . If the accurate rate can be accepted , the model will be used in classification . Generally , model can be expressed by classification rule , decision tree and mathematical formula . Some classification rule algorithms in data mining are commonly used at present, such as neural network algorithm, genetic algorithm and decision tree algorithm . In this paper , as a method in data classification mining , neural network algorithm has been mainly used to study its applications in teaching .BP (Back Propagation) algorithm is the most popular training algorithm in applications for its non-linear mapping approach capability , robustness and easy realization . However , it is known to have some defects , such as complex gradient calculation , converging slowly and immersing in local vibration frequently . Considering the great capacity of data in education area and the character of neural network used in classification ,the thinking of particle swarm optimization (PSO) algorithm has been used to improve BP (PSOBP) algorithm to train neural network in this paper . PSO algorithm essentially is an algorithm to seek superior answer through probability calculation . It has no limit to objective function and adapts to the processing of large number of data . PSOBP algorithm has no gradient calculation in traditional BP algorithm . It seeks the best weights matrix through particle swarm iteration . It has been proved by experiment that PSOBP algorithm not only can find the smallest value in overall situation but also can greatly enhance the convergence rate . Its classification effect is superior to traditional BP algorithm .