Abstract:

The artificial neural network is an effective method of text categorization. However, the uncertainty of the network makes it difficult to find a suitable network. This paper uses the particle swarm optimization algorithm to optimize neural network, makes it adaptive to adjust its connection weights and network structure in the evolutionary process. First, represents the text set as a vector space, and then uses the information gain algorithm to select feature items, uses the term frequencyinverse document frequency to calculate its weight.Finally, uses the evolutionary neural network to categorize the Chinese text automatically. Experimental results show that comparing with the original BP neural network, the classification of evolutionary BP neural network is better.

Key words: text categorization, information gain, term frequencyinverse document frequency, neural network, particle swarm optimization algorithm