The Research Of The Emerging Technology Weak Signal Recognition Based On Patent

Whether the countries, companies or research institutions, accurately controllingthe technological development direction, seize the initiative for occupying adominant position will play a very important role in orgnizations’ developmentstrategy decision-making. However,emerging technologies have the feature of highlyuncertain, creative destruction, explosive growth, discontinuous change, to achievethe goal of keeping track of the industry or technical direction, need time to monitorglobal changes in technology forefront of the dynamic,and capture the track of theemerging technologies accurately, take action as soon as possible.Since the emerging technology often appears in the margins outside themainstream technology,and the scientific knowledge base, technical architecturestandards, functional gain is not clear, which is easily to be overlooked, so thesignificant signs of weak signals often hidden behind a lot of noise signal, competitionsignals or mainstream technical signals.How to carry out weak signals fromcomplex signals, identify potential technology trends, capture signal featuretechnology, is the main purpose of this thesis.In the research, we use Derwent patent information(DII) as the data source,Weuse business intelligence tools for data integration for the purpose of constructingpatent dataset with high quality,and we use text mining, information extraction,pattern recognition theory and techniques to construct structural basis SAO-Csemantic feature extraction algorithm, as well as the deeply annotation algorithm forthe extraction of the semantic representation of the patent information and theconstruction of the knowledge tree according to the domain. What’s more, we use thenumber of the patents corresponding the content of the tree to identify the potentialweak signals, and to assess them with other patent semantic features.Innovation of the dissertation are mainly in the following three points: first,semantic annotation algorithm based on the SAO-C structure is proposed; second,knowledge tree construction method based on the semantic feature extractiontechnique is presented; third, building the weak signals recognition method accordingto the semantic information, text mining technology and the knowledge tree.