Research on the Impacts of Quantitative Factors on Sentimental Classification of Weibo of Different Topics
Ruoxi Zhang · 2015
Different numbers of terms and texts of training set are involved to optimise the performance of sentiment classification for Weibo repost.The experiment utilises CHI-square test to extract terms and uses support vector machine (SVM) to classify the different sentimental categories.By measuring the performance by F1-score for different training sets, the result illustrates the impacts of quantitative factors on the performance and the differences of the impacts between particular topics.