Research And Application Of Text Classification Technology Based On BC-ACO

With the rapid development of the medical technology, CT, MRI data generated by growing more and more, resulting in more and more clinical workload of doctors. Therefore, medical research Diagnosis System is very meaningful. In this paper, for medical images into text data characterizing classify help medical image processing and fast and accurate diseased tissue respectively.This article studies Naive Bayesian Classification Technology, at the same time, also studies AODE and AAPE, improved Naive Bayes Algorithm. AODE and AAPE has the same rules in classification that is they all ensure the relevance between attributes, but AODE classification sets each attributes as parent attributes, then calculate the average. While AAPE sets parent attributes using random ways, ignoring the importance of the property which can make the poor accuracy of classification. This article analyses the naive Bayesian Algorithm, shortcomings of AODE classification and AAPE classification, extracts the parent attribute by using ant colony algorithm. Below are the researches of this article.1 This article puts forward the concept of BC-ACO(Bayes Classification-Ant Colony Optimization) classification Module, by absording characteristics of the AODE and AAPE classification, combined with the Ant Colony Algorithm, aimed at the medical image processing system, with Naive Bayesian Algorithm. BC-ACO classification guarantees the correlation between property at the same time, also optimizes the parent selection, rather than selecting parent attributes randomly, and averaging the parent attributes, using the ant colony algorithm to select the parent properties improves the accuracy of the final classification results.2 This article achieves Medical Image Processing System based on BC-ACO, this system sets BC-ACO as the classification module, first using the Naive Bayesian Algorithm carries on the preliminary classification, then interrupts system, if there is no interrupt instruction, this system will enter the second stage, using ant colony algorithm for further classification. And it can also configure other parameters in ant colony algorithm to achieve multi-adaptability goal.