Research On The Strategic Emerging Industry Technology Report Based On Thematic Model

Scientific and technological innovation is the core driving force for the development of strategic emerging industries,and scientific and technological innovation requires the support of scientific and technological information resources,which requires the transformation of scientific research results.Science and technology report is the main research achievement of science and technology project,it is important science and technology information resource.Technology report mainly describes the scientific research workers are engaged in scientific research,experiments and design process,progress and results,according to the regulations of the relevant format and writing and become,can truly reflect the complete technology content and experience of special engaged in scientific research literature.In 2014,with the official launch of the national science and technology reporting service system,it marked a new stage in China's science and technology reporting.As of April 2,2018,the national science and technology report included in the 123635 science and technology service system,users can search in the system,they are interested in the report of science and technology,it can satisfy the needs of users in a certain extent,but the user how to get from a large number of science and technology reports the report they want to become the resulting problems,referring users to meet the actual demand of science and technology become the problem to be solved next report.In view of this,the retrieval and reading a large number of related literature at home and abroad,on the basis of an in-depth study and research the methods as recommended,based on the theme of model put forward the following three recommendations:Firstly,based on the recommendation of LDA model,the topic extraction is completed by using LDA theme model,and the recommendation is completed by calculating the similarity between the document topics.This method not only realizes the dimensionality reduction of the word frequency space,but also reduces the difficulty of calculation,and can also find the potential relation between documents and the probability distribution of each topic within the document.Secondly,based on the recommendations from the CTM theme related model,the model in the LDA theme,on the basis of considering the potential association of different subjects in different documents,finally will subject correlation between larger documents recommended to the user,make up the theme the LDA model tend to recommend a topic document the scope of the problem,improve the effect of the recommended,and the advantages of CTM is verified by the empirical correlation model.Thirdly,based on the recommendations from the interest subject,through access to user behaviour on the historical record,the user interest feature extracting,building user documentation-the relationship between subject,then calculate the document theme and the degree of similarity between user interest subject,finally according to the similarity sorting before the similarity of the larger m document recommended to the user,so as to complete the recommended process.Based on the theme model,this method adds users'interest,which makes the recommendation result more fit to the actual needs of users,thus enhancing the recommendation effect.