Automatic Keyword Extraction Method for 3GPP Technical Standard
Huan Li, Xueqiang Lv, Liping Xu · 2017 International Conference on Computer Systems, Electronics and Control (ICCSEC) · 2017
In the patent infringement testing work, keyword extraction for the 3GPP technology standard has been both the key point and the difficulty in the examination of communication patent. In order to improve the retrieval quality of 3GPP technical standard and reduce the labor cost. Based on the analysis of the characteristics of 3GPP technical standard, Firstly, a method based on word frequency-document entropy to extract the universal words in the 3GPP technical standard is proposed, and then the algorithm of extracting the candidate keywords based on the dependency syntax analysis tree is proposed. After filtering the candidate keywords, the position feature, the word co-occurrence feature and the context semantic feature are used to calculate the weight score of the candidate keywords, and the dynamic thresholds are calculated according to the weight score range of the candidate keywords in each technical standard. Finally, dynamic thresholds are used to determine the result. Experiments show that the method has achieved higher correct rate and recall rate than the traditional keywords extraction method, and also proves the validity of three different features when extracting different numbers of keywords.