Research And Implementation Of Enterprise Innovation Capability Evaluation System

Since China's implementation of the innovation-driven development strategy,innovation has become increasingly important in various fields.For an enterprise,innovation capability is an important manifestation of an enterprise's comprehensive strength and one of the main ways to enhance its comprehensive strength.For the assessed company,through regular monitoring and evaluation,using appropriate guidance methods to make the company's innovation activities benign development,thereby enhancing the company's innovation ability,and then enhance its comprehensive strength.Most indexes in the evaluation index system established by the predecessors are extremely difficult to obtain and are not suitable for the evaluation of new capabilities of current enterprises.In the era of big data,enterprise data is distributed all over the Internet,such as business data,patents,and scientific and technical literature.These data reflect the innovation ability of enterprises to a certain extent.Based on these data,this article has done the following research:(1)Constructed a BERT-based enterprise entity relationship extraction model.In view of the missing key data,this article extracts these missing enterprise data from the company profile text.Due to the complexity of the Chinese text,simple rule matching cannot accomplish this task well.This paper uses BERT to extract the entity relationship of the enterprise,and the final result is in line with It is expected that the extraction accuracy will be above 80%.(2)Designed and implemented enterprise data collection system.Aiming at the distribution of enterprise data in different websites and knowledge databases,the design and implementation of a data collection system based on the Scrapy crawler framework,crawling the enterprise list of industrial and commercial information,patent information and journal paper information and other innovative capacity evaluation index system required For data,the average coverage of the final data collection is above 90%.(3)On the basis of observing the SMART guidelines,the three items of input,output efficiency and output quality are regarded as the first-level indicators,and the financial input,manpower input,the proportion of invention patents and the proportion of core journals are regarded as the second-level indicators.Under the premise of minimizing subjective factors,the close value method is selected as the method of evaluating the model,and the final evaluation result is calculated.(4)Design and implement the evaluation system.In order to intuitively display the data in the database and facilitate the management of data,Java web-related technologies are used to visually display and manage the enterprise data existing in the database.