Abstract: This research is constructed on the theoretical basis of text mining technology and explores to establish a patent function analysis approach which mainly focuses on the specific Action-Object structure in patent text documents after summarizing research progress in patent text mining home and abroad. This research has realized to identify potential field of patent technology industrialization from patent text documents through defining, extracting and analyzing patent function so that to connect patent with real industry. Importantly, this research presents the S-curve and S-index of patent technology function efficacy and improves the evaluation model of patent technology industrialization applicability by defining the three evaluating indexes which are S-index, Absolute Importance of the Function (AI) and Relative Importance of the Function (RI). Finally, this research takes patents that bear the function of “reduce PM2.5” as an example to verify the feasibility and significance of the approach which helps to identify and evaluate patent technology industrialization by patent function analysis and thus provides a new thinking on the path choice of patent technology industrialization in China.

Key words:  patent function analysis, patent technology industrialization, applicability, text mining, PM2.5