Analysis Of Intelligent Transfer Of Appeals Based On Text Topic Mining

With the progress of society and the rapid development of the market economy,government appeals have gradually become one of the important ways for the government to interact with the public.At the same time,public appeals are showing a trend of diversification with the increasing number,and the problems received by the government and enterprises are more diversified.Therefore,how to transfer government affairs quickly and accurately so as to improve the service efficiency of government and enterprises has become an important topic.The current appeal transfer methods mainly focus on manual judgment and appeal transfer systems which are mostly based on traditional data analysis methods.With the increasing of appeal date,these methods are tending to be inefficient and inaccurate.Therefore,the intelligent transfer of government appeals based on text knowledge mining and other related technologies have been extensively developed.This paper explores these issues on the basis of text knowledge mining algorithms and topic model algorithms.Aiming at the problems that the existing methods only conduct knowledge mining on single-scale text data,and cannot comprehensively analyze the characteristics of specific domain corpus from multiple levels,this paper proposes a text knowledge mining method based on multi-scale clustering.Firstly,apply word2 vec algorithm to train the preprocessed text data for word vector;secondly,use the fast clustering algorithm based on local density to cluster the result set of different scales of text word vector;finally,perform on the results of different scales of text clustering to carry out the knowledge mining to achieve multi-angle characterization of text data at multiple different scales.Experiments show that this method can dig out a variety of appeal themes from user appeal data,and specific social problems in the user-defined scale appeal data,and provide data and technical support for managers to make decisions.In view of the increasingly diversified problems of government appeals,and the waste of resources and inefficiency caused by traditional appeal transfer methods due to limited functions,this paper proposes an intelligent appeal transfer method based on the topic model.First,a method for constructing a claim topic model based on the TI-LDA algorithm is proposed.This method calculates the word frequency and inverse document frequency of each claim keyword after data preprocessing based on the historical claim data,and builds the core word bag model of each department;and uses the bag of words to filter The keywords of each department are constructed by constructing the word frequency matrix of each department’s appeal text to generate the topic model of each department.Secondly,the method also builds a forecasting model for the appeal department based on the actual situation of the government appeal,and accurately predicts the administrative department to which the appeal belongs.Experiments show that the method can carry out the intelligent transfer of government affairs demands quickly and accurately,and adjust the corresponding models with the adjustments of department function,which greatly improves the service quality and efficiency of government and enterprises.