Design And Implementation Of Patent Evaluation Platform

In recent years,the level of national innovation has continued to increase,and the number of patent invention applications has continued to rise.It has ranked first in the world for many years,far exceeding other countries.With the substantial increase in the number of patent applications,the demand for patent value assessment has also increased tremendously.Most of the behaviors of patents in the patent life cycle require patent evaluation,such as in patent transfer,patent infringement,etc.Due to the complicated process and long evaluation cycle of traditional patent value assessment methods,it is difficult for small businesses,small groups,and individuals to have a reasonable understanding of the patent value they own,which is not conducive to the healthy development of intellectual property rights.Using BP neural network to predict the value of patents can greatly shorten the evaluation cycle and provide patent evaluation services to the public conveniently.In the process of project development,the author first participated in the requirements analysis of the project,analyzed the functional requirements of the platform,and clarified the overall goals of the project.In the outline design,the author participated in designing the overall architecture of the platform,the implementation method of each functional module,and the database table according to the analysis of platform requirements.According to the needs of the platform,the platform is divided into a data acquisition module,a text analysis module,a value evaluation module,a patent recommendation module,and a user center module.The data acquisition module mainly writes crawler scripts for new data sources and improves existing crawler solutions,enabling it to handle new anti-crawling methods and continue to perform data acquisition tasks;introducing multi-coroutines and dynamic http tunnels into the crawler script Solution,increase the single success rate of data crawling and concurrent execution performance.For the text analysis module,the TF-IDF algorithm is used to calculate the representative feature words of each patent text.In the design of the value evaluation module,the author searched a large number of literatures on intangible asset evaluation.According to the nature of patents and factors considered in traditional patent evaluation,he chose reasonable data and BP neural network models to realize the value evaluation function of patents.When designing a text recommendation module,different patents are recommended for users in different platform operations,and similarity is calculated from different angles of patents and users to implement the recommendation function.The patent value evaluation platform using BP neural network can provide users with convenient patent value evaluation functions.Users can easily understand the value of patents and related information using the platform;the recommendation function provided by the platform to calculate user behavior,Can provide users with patent services that better meet user needs.At present,the platform has passed the functional test and non-functional test to reach the on-line standard,and is waiting to be deployed.