Research On The Life Cycle Of Technical Topics Based On Patent Data

With the development of economic globalization,technological innovation has become an important driving force for the development of today's world.It is important for enterprises and Governments to grasp the opportunity of technological innovation in time.The potential of technological innovation emphasizes the current stage of development in the technical field and the future development trend and development space.Patent literature is one of the achievements and manifestations of technological innovation,it is important to apply text mining to study the life cycle of technology in patent data.Firstly,the patent text data was transformed into structured patent topic data.Analyze the content and characteristics of patent data and select the text data that can be used for technical topics in patent data.Segment the patent text data and disable the useless words in the word segmentation results.Through the vector space model,the TF-IDF model,and the latent dirichlet allocation,the structured patent topic data is exported to provide data for the patent technology life cycle detection method.Secondly,define the life cycle of the technical topics and introduce the detection methods of the topic life cycle.A study of methods that use multiple indices to detect life cycles defines three phases of the technical topic life cycle: potential phase,emerging phase,and maturity phase.According to the characteristics of patent data of this paper,we optimize the novelty index and propose the technical range index using the number calculation of classification numbers.Use the detection method using the published volume index and use the technical range index to detect the technical topic's life cycle.Thirdly,for the problem of using an index that is too monotonous for the method of using published volume index,an issued volume index is used instead of the published volume index.Based on this,technical range index are added to expand and supplement.Use the issued volume index and the technical range index to improve the detection method using the published volume index.Finally,the test takes the patent data of the automobile field as an example to verify the detection method.Extracting technical topics from the automotive field and judging the life cycle stages of each topic.This test compares and analyses the results of the method of using the issued volume index,the detection method using the technical range index and the improved methods which are combined with the issued volume index and the technical range index.