Construction And Implementation Of An Evaluation Index System For Online Social Network Information Dissemination

The information dissemination index system is a key way to measure the dissemination of public opinion events in social networks.The research on the index system of information dissemination can help us understand the generation and dissemination mechanism of negative phenomena in information dissemination better,and provide us scientific basis for formulating corresponding policies and measures.At the same time,a scientific information dissemination index system can also help the public better understand the process and effect of information dissemination,improve the quality and efficiency of information dissemination,thereby enhancing the public’s cognitive ability and judgment on information.Because of the complexity,variability,information asymmetry,and evolution uncertainty of public opinion event information dissemination in social networks,the construction of traditional information dissemination index systems is overly dependent on manual intervention,with large deviations and limited scope of application.and other shortcomings,and cannot meet the requirements of an objective,all-round,and highly credible index system.The research of this paper aims to establish a reasonable and easy-to-use information dissemination evaluation index system,quantify the information dissemination situation per unit time,evaluate and calculate the situation,effectively identify dissemination risk factors,quantitatively evaluate the information dissemination situation level,and provide practical Actionable decision-making index system.The research points and innovation points of this paper mainly include the following points:(1)A multi-level and multi-granularity information dissemination index system construction method is proposed.This article starts with captured social network posting and news data,through the analysis of social network data and event influence,combined with the opinions of social network experts and ordinary users,it constructs perspectives from three perspectives: communication events,communication audiences,and communication media An index system with rich and comprehensive influencing factors.(2)Propose and implement a deep learning method using convolutional neural network to extract the characteristics of public opinion events in the process of information dissemination,and convert them into vector matrices,and use the SMOTE algorithm to process the sample word vector matrices of minority categories Amplification is achieved,and the problem of data imbalance is alleviated.Finally,the weight value of the information dissemination index system is trained through the designed convolutional neural network,and the corresponding index system is formed.Experimental results demonstrate that the used method outperforms other methods and performs better on domain-specific datasets.(3)Design and implement an online social network information dissemination evaluation system.Aiming at the characteristics of social media multi-source,multi-channel,spatiotemporal big data,a set of data acquisition architecture of "acquisition layer-acquisition management layer-main control layer-visualization layer" was built based on the Kubernetes platform,which realized the data collection in distributed Efficient collection in the scene.And using the information dissemination index system proposed in this paper and the trained prediction model,an online social network information dissemination evaluation system is designed and implemented.The system can conduct real-time and quantitative analysis on the dissemination of network events and online emergencies,and provide technical support and decision-making basis for the subsequent guidance and control of network information dissemination and the initiation of emergency plans.