Research On The Group Opinions Evolution Model In Knowledge Collaboration Under Complex Networks

In the era of knowledge economy,enterprises integrate and optimize knowledge resources through knowledge collaboration,which is a more advanced management method,to realize enterprise innovation.However,the process of knowledge collaboration is a continuous evolutionary process of coordinating conflicts and moving from disorder to order,and is characterized by complexity,randomness and dynamism.Therefore,how to monitor and ensure the smooth progress of the knowledge collaboration process and keep it in a "controlled" state is a problem that should be solved by the open knowledge collaboration system.In this paper,we introduce opinion dynamics theory and stochastic process theory,construct a group opinion evolution model by using system modeling method,and deeply analyze the interaction process and opinion evolution mechanism in knowledge collaboration layer by layer,so as to better understand the self-organization behavior of users in the absence of organizational control and formal collaboration mechanism.First of all,stochastic process theory is used to establish group opinion evolution model,which is helpful to explain and predict the opinion evolution process in the knowledge collaboration process.However,the actual interactive network is complex and large scale,so it is very difficult to directly construct the transition probability matrix,which limits the application of this method.In order to reduce the difficulty of modeling and calculation,the method of spectral clustering algorithm are used to decompose the network into subgroups of different levels.Finally,Markov chain is introduced to establish a one-step transition probability matrix and its recursive formula for various rule networks.This method can not only describe the evolution process of subgroup opinion,quantitatively analyze and model the evolution law of group opinion,but also extend to large groups with similar structure.Secondly,the simulation experiment method and the empirical analysis method were used to verify the rationality and validity of the model.The simulation experiment method found that although the relative errors of the theoretical model and some data of the simulation experiment were large,the overall average error and the average relative error were small,which indicated that the model established could basically simulate the evolution process of the group opinion and the model was valid.Through empirical analysis of the real collaborative interaction data in "Baidu Tieba platform",the results showed that after all discussions in the theoretical model,the proportion of group selection in support of BYD’s "Han EV" opinion was 68.7%,which was very close to the actual proportion of 70.6%,which again verified that the model established in this paper could better predict the state of group opinion evolution in collaboration.Finally,aming at the phenomenon of frequent interactions between individuals with different opinions within a real network group,an index is presented to quantify the differences of opinions within a subgroup.Then,based on the model and empirical data,the influence of different opinion differences on the evolution of group opinion is compared.Different degree of opinion difference essentially changes the interactive network structure,and then changes the way and efficiency of opinion exchange.The results show that when the degree of difference of opinions is higher,people are more active,the discussion enthusiasm is higher,the atmosphere is more heated,and the efficiency of the exchange of opinions is higher.Therefore,the model proposed in this paper help to explain the evolution mechanism of opinions dominated by individual interaction in knowledge collaboration under complex networks,and lay a foundation for ensuring the smooth development of knowledge collaboration activities and improving the quality of knowledge service.