CAJ | 학술논문

为了进一步提高文本分类的准确率，文中介绍了一种新的用于文本分类的概率分类器。该分类器首先通过自然语言处理技术对文本进行预处理，然后从训练集中读取文本信息从而产生正、负规则，并计算正负权重系数，最后计算正、负概率。文中给出了计算正负权重系数的算法，并根据计算出来的权重系数及正、负概率值对文本进行分类。将文中提出的概率分类器与SVM分类器进行对比实验，实验结果表明，文中设计的概率分类器对于文本分类效果较好。
위료진일보제고문본분류적준학솔，문중개소료일충신적용우문본분류적개솔분류기。해분류기수선통과자연어언처리기술대문본진행예처리，연후종훈련집중독취문본신식종이산생정、부규칙，병계산정부권중계수，최후계산정、부개솔。문중급출료계산정부권중계수적산법，병근거계산출래적권중계수급정、부개솔치대문본진행분류。장문중제출적개솔분류기여SVM분류기진행대비실험，실험결과표명，문중설계적개솔분류기대우문본분류효과교호。
In order to further improve the accuracy of text categorization,a new probabilistic classifier for text categorization is intro-duced. Firstly,it preprocesses the text using natural language processing technology. Then,read the documents from the training set and generate positive and negative rules,calculate the positive and negative weighted coefficient. Finally,compute the positive and negative probability. The algorithm of calculating the positive and negative weighted coefficient is presented,and categorize the text in accordance with the coefficient and the probability value. Take the experiment compared with SVM classifier,the experiments results show this classi-fier has better performance in text categorization.