Research On Key Information Extraction And Automatic Classification For Scientific And Technological Demands Of Enterprises

Nowadays,the collaborative innovation model is a new model of technological development.Enterprises,universities,and scientific research teams can give full play to their advantages in the model.It combines complementary resources,transfers a large number of technological achievements,and keeps industrial technology innovation constantly.The platform of collaborative innovation technology services is an important carrier of the model.However,it is difficult for enterprises to select scientific research teams proactively,accurately and flexibly.With the increasing demand in the platform,the platform will face a large amount of noise data inevitably,and the problem of information overload will occur,which is not conducive to matching the demands of enterprises with scientific research institutes or universities accurately.Therefore,the paper extracts key information for scientific and technological demands of enterprises and classifies it automatically.In this way,the demands of the enterprise can be intelligentzed,and the matching problem between the technological demands of the enterprise and the research team can be solved effectively.Based on the texts of technological demands of small and medium-sized enterprises(SMEs),the paper builds a model for extracting key information on the technological demands of SMEs base on LDA model.Then the paper takes the scientific and technological demands of software projects as an example,and constructs the COSD-RCNN model.Finally,the paper designs and implements a software technology service platform for the collaborative innovation model,which recommends a suitable research team for the enterprise to solve the demands.The main research contents of the paper are as follows:(1)Key information extraction method for technological demands of SMEs based on LDA.Aiming at a variety of information on technological demands,the paper based on the text of the technological demands of SMEs,and proposes a method for keywords extraction based on LDA.Firstly,natural language processing technology is used to preprocess text.Then,the paper combines multiple feature weighting and LDA model to extract keywords.Finally,the method chooses the best keywords through the post-processing.The method was tested on Chinese Scientist In dataset.The experimental results show that the F-value is higher than that of TF-IDF and LDA.It also shows that the method can extract key information of technological demands for enterprises effectively,and improve the intelligence of the scientific and technological collaboration platform.(2)Automatic classification of software project demand texts based on RCNN.With the increase of the application software on the internet,the demands of software projects are more diverse,and the classification and management of demand texts have become more difficult for us.Based on the traditional machine learning classification method,the paper uses the deep learning text classification method,and proposes an automatic classification method of software project demand texts based on RCNN.After analyzing the software project requirements text,the software project is divided.Then the paper collects and preprocesses large-scale demand text data,and uses the trained COSD-RCNN model to classify the title part of the demand texts.After experimental comparison,the average F-value of the model on the test set is 93.02%,which is higher than other baseline algorithms.The experimental results show that the model has strong classification ability.(3)Design and implementation of software technology service platform for collaborative innovation.The platform is based on the B/S architecture and follows the PHP specification.The paper designs platform architecture,functions,business processes,and databases.Then the paper implements functions including the release of software project requirements,the recommendation of scientific research teams,the display of successful case results,the display of scientific and technological information,and the communication of applied technology.By processing the software project demands of the enterprise,the platform recommends suitable scientific research teams to solve the demands of users actively and accurately,and improves the collaborative innovation technology service capability effectively.