A patent keywords extraction method using TextRank model with prior public knowledge

Abstract For large amount of patent texts, how to extract their keywords in an unsupervised way is a very important problem. In existing methods, only the own information texts analyzed. this study, improved TextRank model proposed, which prior public knowledge effectively utilized. Specifically, two following points are first considered: (1) network constructed for each text, (2) based on dictionary data, edges represent interpretation relationship among all words entries. Then, node rank value evaluation formula designed networks introduced. Finally, can be extracted by finding top-k with higher values. our experiments, text clustering task used examine performance proposed method, wherein several comparison experiments executed. Corresponding results demonstrate that, new method markedly obtain better than methods extraction way.