Automated patents search through semantic similarity

Patent search is a complex process involving ample time and manpower. The search process is filled with complexity and results are often ambiguous. This research has worked on patent abstract and has found similar patents with the help of expanding query key words. Word Net is used as an external medium for query expansion. Wu-Palmer similarity model is applied in retrieving the output from the database. The output shows that expanded query produces better results in terms of qualitative & quantitative rather than unexpanded query. Query expansion increases the number of keywords which help in fetching similar patents. The results of this model can be used for patent prior art search.