An improved text classification algorithm is presented to improve the accuracy and efficiency of the public opinion classification. The algorithm filters the part of speech before feature extraction to decrease the useless feature and then classifies text according to the calculated weight. The experimental results show that the feature extraction of the improved algorithm is more efficient than the previous ones, and the text classification results in different feature dimensions are more accurate, especially in the lower dimensions. Therefore, it has important significance for text classification by analyzing the weight of the part of speech to extract feature and calculate weight before classification.
展开▼