Font Size: a A A

Research On Robot Literature Subject Classification Based On Improved GSDMM Clustering Model

Posted on:2018-02-07Degree:MasterType:Thesis
Country:ChinaCandidate:L LiFull Text:PDF
GTID:2428330569985170Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
With the advent of the big data era,it has become an important means of auxiliary consulting research to dig large data and research the future development of technology in various fields.In recent years,with the application field of robot technology constantly expanding and deepening,the market demand in robot technology is huge so that it has become the focus of national economic development.Therefore,the analysis of robot technology and technical foresight to ensure its scientific development is an important part.In the process of text subject division,there are existing problem of timeliness,lacking of completeness,complex subject of literature and difficulty to determine the number of topics in the traditional method such as citation clustering.Based on the study of the robot 's subjective needs,a large-scale text subject partition system based on robot literature data is constructed.And it is the focus of the study to establish a suitable machine learning model to obtain accurate clustering results.In this paper,the text clustering model of GSDMM(aquapsed Gibbs Sampling algorithm for the Dirichlet Multinomial Mixture model)is introduced into the process of document subject division.By incorporating the text category label information into the model,the disadvantages of GSDMM have been improved which including low classification accuracy,bad classification performance,introduction problem of prior knowledge.This paper proposes a novel GSDMM semi-supervised text clustering model with additional class tags,which improves the performance of the model.The clustering effect verification experiment is designed to compare the GSDMM model before and after the improvement,and the effectiveness of the improved algorithm is verified.Compared with the benchmark K-means and other models,the validity and reliability of the GSDMM model for text clustering are verified.Based on the above research,we can obtain the literature data of large-scale robot field and realize the division of robot literature technology field,which further supports the research on the science and technology strategy of related engineering in the field of robot.
Keywords/Search Tags:machine learning, text clustering, topic partition, robot, GSDMM
PDF Full Text Request
Related items