发明名称 |
METHOD AND SYSTEM FOR AUTOMATICALLY RANKING PRODUCT REVIEWS ACCORDING TO REVIEW HELPFULNESS |
摘要 |
<p>A method and system for automatically ranking product reviews according to review helpfulness. Given a collection of reviews, the method employs an algorithm that identifies dominant terms and uses them to define a feature vector representation. Reviews are then converted to this representation and ranked according to their distance from a 'locally optimal' review vector. The algorithm is fully unsupervised and thus avoids costly and error-prone manual training annotations. In one embodiment a Multi Layer Lexical Model (MLLM) approach partitions the dominant lexical terms in a review into layers, creates a compact unified layers lexicon, and ranks the reviews according to their weight with respect to unified lexicon, all in a fully unsupervised manner. When used to rank book reviews, it was found that the invention significantly outperforms the user votes-based ranking employed by Amazon.</p> |
申请公布号 |
WO2009087636(A1) |
申请公布日期 |
2009.07.16 |
申请号 |
WO2009IL00039 |
申请日期 |
2009.01.11 |
申请人 |
YISSUM RESEARCH DEVELOPMENT COMPANY OF THE HEBREWUNIVERSITY OF JERUSALEM;RAPPOPORT, ARI;TSUR, OREN |
发明人 |
RAPPOPORT, ARI;TSUR, OREN |
分类号 |
G06F17/30 |
主分类号 |
G06F17/30 |
代理机构 |
|
代理人 |
|
主权项 |
|
地址 |
|