Design And Implementation Of A Content Operation Assistance System For Wechat Official Accounts Based On Text Analysis

With the rapid rise of the content operation industry and the growing popularity of the WeChat Official Account Admin Platform,the official accounts are not only opinion positions of traditional media and self-media,but also windows of almost all enterprises and government departments for marketing and publicity.Consequently,the third-party platform to serve operation workers came into being.As a result of the relatively closed nature of the WeChat ecosystem and the subjective nature of the topic selection,the majority of the existing platforms just served the basic data collection and statistics,leaving the core issues unsolved,such as the integration of resources and selection of topics.In view of the above difficulties,this paper employs text analysis technology to design and implement a content operation assistance system for WeChat official accounts,which provides the functions of sorting materials out and recommending topics,improving the efficiency and quality of operation.This system mainly includes three modules: data acquisition,data processing and system management.To enlarge the source range of the topics,the data acquisition module adopts the methods of real time acquisition and timing acquisition and collects the data of many kinds of channels,such as browsing records,prevailing channels and official accounts,and so on.The data processing module uses the methods of text clustering and keyword extraction to analyze and process the data from different sources.In order to help the operation workers to select the topic,this paper puts forward the concept of "inspired word",and designs the methods of inspired-word extraction based on the algorithms of keyword extraction and new word recognition.By combining the two techniques of text clustering and inspired-word extraction,we can get the inspired words of different dimensions to assist the operation workers to find new starting points.The system management module provides the Web background service for the functions oftopics recommendation by integrating the data processing results and controlling the system’s procedure.In consideration of the diversification of data types,this system realizes the data storage scheme based on MongoDB and Elasticsearch,and makes full use of the advantages of the efficient query of the former and the real-time search of the latter,which effectively supports the data acquisition and processing.In this paper,a number of comparative experiments are designed for the process of feature selection,feature dimension,feature model and inspired-word extraction.The experimental results show the effectiveness of the text analysis algorithms used in this system.Online practice and user feedback prove that the system can help users to integrate resources and select topics,with high practical value.At present,the system is performing well in the company’s online project.