Font Size: a A A

Research On Key Technology Identification In The Field Of CNC Machine Tools Based On Semi-Supervised Machine Learning

Posted on:2020-09-10Degree:MasterType:Thesis
Country:ChinaCandidate:Z ChenFull Text:PDF
GTID:2381330599459268Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
In 2015,Chinese Academy of Engineering launched the “China Engineering Science and Technology 2035 Development Strategy Research” project,aiming to propose technical list of Chinese engineering science and technology fields for 2035,which will provide reference for formulating China's development strategy in the next 20 years.Identification of key technology is a core step of formulating technical list.At present,the key technology identification methods based on expert experience and bibliometric analysis have problems such as long identification period,large subjectivity,unclear technology field,and incomplete technical identification.In order to solve these problems,this paper has made a series of improvements.First identified the key technology of CNC machine tools through the paper data based on semi-supervised interactive topic model that incorporated expert knowledge,and the problem of “long identification period and large subjectivity” was solved.In order to improve quality of the key technology identification using only interactive topic model,this paper proposed a semi-supervised TM-GSDMM text clustering algorithm that combined text-level must-link knowledge,and completed the division of the papers of CNC machine tools,obtained five sub-areas of numerical control system,process system,intelligence,network,and machine tool components,and the problem of “unclear technology field” was solved.In order to identify key technologies in sub-field papers,this paper proposed a semi-supervised WM-ATM author-topic model that integrated word-level must-link knowledge,and identified key technologies in five sub-areas,the problem of “incomplete technical identification” was solved.Finally,based on the methods proposed in this paper,the process of identifying key technologies in all field,division of technical field and identifying key technologies in sub-field was summarized to product technical list,and the “Clustering tool for generating technical list” was built accordingly.Taking the paper data of the whole field of CNC machine tools as input,the technical list of CNC machine tools was exported and recognized by experts.
Keywords/Search Tags:Technical list, Identification of key technology, Division of technical field, Semi-supervised text clustering, Semi-supervised topic model, CNC machine tool
PDF Full Text Request
Related items