Font Size: a A A

Experts Recommend Model Optimization And Its Application In Telecom Project Management

Posted on:2021-05-26Degree:MasterType:Thesis
Country:ChinaCandidate:J H XuFull Text:PDF
GTID:2428330602976850Subject:Computer technology
Abstract/Summary:PDF Full Text Request
Project review is an important part of project comprehensive management,and the selection and recommendation of evaluation experts is the core part of project review.In order to improve the timeliness and accuracy of the selection and recommendation of project review experts,the study and application of expert intelligent recommendation method has become one of the key problems to be solved in the field of project comprehensive management.For this reason,this paper carries out optimization research on the selection and recommendation model of project evaluation experts,and designs and implements a telecom project comprehensive management system.Main work of this paper:1.A Time Decay Author Topic Model(TD ATM)was proposed to introduce a Time Decay Author Topic Model(TDATM).Its main feature is that by introducing the time"attenuation" factor,it measures the specific weight relation of experts' literature feature words in different time periods,which can not only capture the hidden relation between experts' literature and experts,but also reduce the topic feature weight of experts' previous research.The experimental results of Chinese DBLP corpus and crawler grabbing data sets from literature retrieval database show that compared with LDA,Doc2Vec,tf-idf and other theme models,TDATM model has better performance of text classification recommended by experts.2.The project evaluation expert scoring model is optimized and reconstructed.In order to improve the accuracy and selecting experts recommend,applying analytic hierarchy process(AHP),from the target layer,criterion layer,layer three layers integrated adjustment affect the value of expert grading index system,optimize the reconstructed project evaluation experts score model,the example analysis shows that the proposed model is effective and feasible,and accords with the practice of project selecting evaluation experts recommend.3.Based on the TDATM model and the optimized reconstruction of the expert rating model,a new project evaluation expert selection recommendation scheme is proposed.The program firstly calculates the similarity between the expert literature and the evaluation items through the TDATM model,and then generates the expert candidate set link.Then the expert candidate set is optimized and selected through the optimized and reconstructed expert scoring model,and the result of the expert recommendation is formed.The experimental results on the actual data set of telecom show that the results recommended by the evaluation experts for the selection of the scheme are in line with the requirements of telecom project management and are effective and feasible.4.Based on the new scheme recommended by the review experts,a telecom project integrated management system is developed by SSM framework.The system includes four functional modules,such as declaration management,review management,expert management and fine management,which can realize the intelligent selection and recommendation of experts and fine management of project information by combining with the business requirements of telecom project management.The test results show that the system achieves the expected function and performance,and has good stability.Main contributions:this paper proposes an author's theme model which introduces time "attenuation" factor,optimizes and reconstructs the project evaluation expert rating model,and presents a new project evaluation expert recommendation scheme.A telecom project integrated management system is developed,which can realize the intelligent selection and recommendation of experts and the fine management of project information.
Keywords/Search Tags:Expert selection recommendation, Text representation, Time Decay factor, Score model, Analytic hierarchy process
PDF Full Text Request
Related items