The Quantitative Analysis Methodology For The Influencing Factors Of Intertwined Public Complaints And The Development Of System Based On Public Complaints Data

With the rapid development of the economy,intertwined public complaints are becoming more and more common in letters-and-visits work.The problems of intertwined public complaints are important but difficult.Because they always last for a long time,derive from complex reason and even affect social order.With the advent of the information age,data mining technology has been applied in many fields to solve specific problems.However,the letters-and-visits services lags behind the development of society.The existing information technology cannot effectively use the valuable public complaints data resources.The situation in the academics is also not satisfying.The current research on the issue of intertwined public complaints only stays on the qualitative analysis.The research results are relatively incomplete in explaining the causal relationship to the intertwined public complaints.The theory is difficult to guide practice.The application of data mining methods is limited to specific areas such as medical care,education,finance,business,and microblog.Based on the above background,this paper puts forward the goal of “Selecting and improving data mining techniques for the letters-and-visits field to analyze the origin of intertwined public complaints and establish a long-term analysis mechanism”.In terms of methodology,this article pioneered the improvement of the LDA topic mining model and the K-means clustering analysis algorithm in the field of letters-and-visits.Based on the perplexity,a traditional method,A method is proposedto select the number of LDA topics with low computational complexity in this paper.With the discretenessof the public complaints data,a new clustering algorithm combined with K-means clustering and WOE theory,is more accurate.The above research enriches the theoretical system of clustering algorithm and LDA topic number selection algorithm,and has certain theoretical value.In practical applications,this paper as for the land public complaints of X provice,we analyze the provincial data using the topic mining and clustering methods,and mining groups with high intertwined public complaints,and propose four factors as the key influencing factors for intertwined public complaints,includingthe compensation standards for land acquisition,the influence of village officials,the level of letters-and-visits services,and the attitude of petitioners.Based on local conditions,suggestion has been proposed for the letters-and-visits work from four aspects including petition work,compensation policy,incorrupt government,and the attitude of the petitioner.In addition,in order to establish a long-term mechanism for supporting petition data mining,a petition data mining system is designed and inlomented in this paper.The system will be of great significance for improving the level of informatization of the letters-and-visits department and the efficiency of service work,and also for reducing the generation of tangled visits.Therefore,it has certain practical value.