Applied Research On Artificial Fishswarm-Back Propagation Neural Networkalgorithm In Text Classification

With the vast amounts of information and the explosive growth of newinformation, making it difficult for users to get what they need in the deluge ofinformation; text categorization is very usefull to solve the problem of informationclutter at a greater extent, classficate the amounts of text into different categories,allows users to get access to the information that they need in a quickly and effectiveway. Text classification is one of the main branch of data mining, it is a typicalconcrete application that is based on natural language processing and machinelearning algorithms,the research on a variety of efficient classification algorithm inthe application of text classification is one of the important research topics, it is theproblem that is urgent to be solved.This paper describes the research status of text classification and the currentproblems about it, describes the general process in the text classification, discussionthe related technologies, analyzed and studied the text preprocessing, the textrepresentation, feature selection and other important steps and Commonly used textclassification algorithm.This article describes the artificial fish swarm algorithm,and the BP neuralnetwork principle in a systematic way, it is separately discussed on the problemsabout them in the text classification and pointed out that the flaws and Inadequate inthe text classifier of the traditional BP neural network.BP neural network algorithm is more stable and immunity than others, works wellin text classification, but it has many shortcomings such as the learning efficiency isstill low, the convergence rate is not fast enough, easy to fall into local minimum andso on.In comparison,artificial fish swarm algorithm Overcomes the local optimization,obtain the global extremum with strong searching capability, it is not only lessdemanded on the initial values and parameters but also not require a sensitive to theheuristic function, it shows good Performance to solve the more complexcombinatorial optimization problems. So we combine the respective advantages ofartificial fish swarm algorithm and BP neural network,use the artificial fish swarmalgorithm to optimize the text classifier on BP neural network algorithm,constructed amodel based on artificial fish swarm_BP neural network. On this basis, we designedand implemented a text classification system based on artificial fish swarm algorithm and BP neural network, and it is proved that the proposed algorithm has betterclassification results in text classification through comparing the experimental results.