Data monitoring and analyzing system of company image improving system

Abstract

The invention discloses a data monitoring and analyzing system of a company image improving system. The system comprises a user data service and fusion system module, a social media propagation optimization module, a multi-source massive social media data real-time acquisition module, a heterogeneous social media key information extraction module and a stream computing technology module for supporting massive social media data monitoring, processing and analysis, wherein the multi-source massive social media data real-time acquisition module comprises a micro-service global distributed data coordination module, a data acquisition and capture service module, a data acquisition and processing service module and a data persistence service module, and the micro-service global distributed datacoordination module is provided. The data monitoring and analyzing system of the company image improving system can fuse company customer service data and complex social network data and construct a data fusion system architecture, and has the advantages of being more efficient and reliable in data analysis and processing, wide in application field, high in practical value and low in expert cost.

Description

technical field [0001] The invention relates to the field of monitoring and analysis systems, in particular to a data monitoring and analysis system for a company image promotion system. Background technique [0002] Existing social media big data is mainly composed of heterogeneous webpage texts with multiple sources and low information density. During the pilot verification process, there are both real-time computing and processing requirements and offline large-scale analysis requirements; there are both semi-structured webpages and For unstructured documents, there are also great differences in data processing, analysis tasks, data update cycle, access delay, call frequency, etc. for different data processing such as potential customer demand analysis, emotion recognition, and social media communication monitoring and analysis. Traditional rules Extraction techniques have low recall and insufficient robustness to cover comprehensive structured fields and key indicators. ...

AI Technical Summary

Problems solved by technology

[0002] Existing social media big data is mainly composed of heterogeneous webpage texts with multiple sources and low information density. During the pilot verification process, there are both real-time computing and processing requirements and offline large-scale analysis requirements; there are both semi-structured webpages and For unstructured documents, there are also great differences in data processing, analysis tasks, data update cycle, access delay, call frequency, etc. for different data processing such as potential customer demand analysis, emotion recognition, and social media communication monitoring and analysis. Traditional rules Extraction techniques have low recall and insufficient robustness, making it difficult to cover comprehensive structured fields and key indicators

[0003] The traditional rule extraction technology has a low recall rate and insufficient robustness, and it is difficult to cover comprehensive structured fields and key indicators. First, social media content exists in the form of web pages, which contain a large number of key fields that need to be analyzed, mined, and calculated. Information, such as the number of retweets, comments, and account information on Weibo, the number of reading articles on WeChat official accounts, etc.

Second, the web pages of different social media have different structures and indicators are also different, and they are often changing. Only using rule extraction technology cannot meet the updates and changes of social media

Third, whether it is natural language processing or wrapper technology, it is often aimed at a specific information source, a specific topic or field. Every time a new information source appears, a set of extraction procedures needs to be reconstructed, resulting in poor scalability of the system.

At the same time, the current structure-based methods of various rule extractions rely too much on the web page structure, and the content-based methods require complex language knowledge and natural language processing technology

[0004] However, the existing data monitoring and analysis system has certain disadvantages in use. It is difficult to integrate the company's customer service data and complex social network data to build a data fusion system architecture. The data analysis and processing efficiency is low, the practical value is not strong enough, and the expert cost higher

Images