论文部分内容阅读
[目的/意义]进一步完善电子商务交易网站的评论体系,提升用户的在线购物体验。[方法/过程]从评论内容、评论者特征和商家特征3个维度确定了在线商品评论可信度的10个影响因素指标,并在此基础上提出了基于DDAG-SVM的在线商品评论可信度分类模型。最后基于MATLAB和LIBSVM,利用淘宝平台近5000条数据集对该模型的准确度进行对比测试。[结果/结论]实验结果达到了93.687%的平均分类准确率,具有较高的准确率和一定的可行性。[局限]分类器预测的准确性一定程度上依赖于人工标注的评论数据集。
[Purpose / Significance] To further improve the commentary system of e-commerce transaction websites and enhance the users’ online shopping experience. [Method / Process] The 10 influencing factors of the credibility of online product reviews are determined from the three dimensions of comment content, commenter characteristics and business characteristics, and based on this, a DDAG-SVM-based online product review credible Degree classification model. Finally, based on MATLAB and LIBSVM, Taobao platform using nearly 5000 data sets to compare the accuracy of the model test. [Result / Conclusion] The experimental results achieved an average classification accuracy of 93.687%, with high accuracy and certain feasibility. [Limitations] The accuracy of classifier predictions relies to some extent on manually annotated comment datasets.