CAJ | 학술논문

支持向量机是在统计学习理论基础上发展起来的一种机器学习方法，广泛应用于文本分类领域。利用信息增益法进行文本特征选取，运用TF-IDF进行特征权重设置。对支持向量机不同核函数，通过网格与交叉验证组合法优化参数选择，比较支持向量机在不同核函数下文本分类性能，得出一些有价值的结论。
지지향량궤시재통계학습이론기출상발전기래적일충궤기학습방법，엄범응용우문본분류영역。이용신식증익법진행문본특정선취，운용TF-IDF진행특정권중설치。대지지향량궤불동핵함수，통과망격여교차험증조합법우화삼수선택，비교지지향량궤재불동핵함수하문본분류성능，득출일사유개치적결론。
Support vector machine is one of the machine learning methods based on statistical learning theory. It is widely used in the field of text classification. The text features were chosen with the information gain method, and the feature weight was set based on TF-IDF method. K-cross validation and grid search were combined to optimize the support vector machine (SVM) kernel function’s parameters. Finally, some helpful results were obtained by comparing text classification performance of the SVM under different kernel functions.