本文選題：評(píng)論有用性評(píng)估 + 書(shū)寫(xiě)行為�。� 參考：《計(jì)算機(jī)工程與應(yīng)用》2016年06期

【摘要】：針對(duì)電子商務(wù)網(wǎng)站充斥著大量有用性較低的評(píng)論,提出一種基于用戶書(shū)寫(xiě)行為與語(yǔ)義特征的中文評(píng)論有用性評(píng)估方法。方法通過(guò)在Web客戶端實(shí)時(shí)監(jiān)聽(tīng)評(píng)論文本框值的變化,識(shí)別出句尾插入、非句尾插入、句尾刪除、非句尾刪除等書(shū)寫(xiě)行為,在服務(wù)器端根據(jù)書(shū)寫(xiě)行為、評(píng)論的語(yǔ)義特征建立的線性評(píng)估模型計(jì)算用戶評(píng)論的有用性。實(shí)驗(yàn)結(jié)果表明該方法能夠較為準(zhǔn)確地識(shí)別有用性較高的評(píng)論。
[Abstract]:In view of the fact that e-commerce websites are filled with a large number of comments with low usefulness, a method for evaluating the usefulness of Chinese comments based on user's writing behavior and semantic features is proposed. Methods by monitoring the changes of the value of comment text box in Web client in real time, we can identify the writing behaviors such as insert at the end of sentence, insert at the end of the sentence, delete the end of the sentence, delete the end of the sentence and delete the end of the sentence according to the behavior of writing on the server side. The linear evaluation model based on the semantic features of comments calculates the usefulness of user comments. The experimental results show that the method can accurately identify useful comments.
【作者單位】： 上海應(yīng)用技術(shù)學(xué)院計(jì)算機(jī)科學(xué)與信息工程學(xué)院;阜陽(yáng)師范學(xué)院計(jì)算機(jī)與信息工程學(xué)院;
【基金】：國(guó)家自然科學(xué)基金(No.61471232)
【分類號(hào)】：TP391.1
，

本文編號(hào)：1859436