| With the development of smart grids,power grid companies have accumulated a large proportion of unstructured data,which is an important part of power big data.The processing and analysis of text data is of great significance to the construction of smart power grids.Nowadays,the application of intelligent text mining technology in the power field is mainly focused on the application of power defect texts and dispatch texts,which is still in its infancy.There are few researches on power dialogue texts,and this part of the text contains a lot of power equipment suppliers.The information has mining value.At the same time,the satisfaction evaluation of power equipment suppliers of power grid companies currently mainly relies on manual statistics and index calculations,and its accuracy is greatly affected by the evaluators and evaluation content.This article starts with text intelligent mining technology and its application in the electric power field.It takes dialogue texts involving power equipment suppliers as an example to study the topic of dialogue texts and research on power equipment supplier evaluation models,which solves the problem of one-sided evaluation of power grid companies.The main research content of this paper includes the following points:1.Combined with the text structure and characteristics of the electric power field,the text intelligent mining technology and its application and prospects are analyzed.First the key technology of text intelligent mining and its application in various fields are introduced.Then the application status of text intelligent mining in the electric power field is analyzed.Compared it with other applications in various fields,the key points and difficulties of text intelligent mining in the electric power field are pointed out.The paper looks forward to the future development direction of the technology in the power field.2.In order to solve the problems of power dialogue text,such as irrelevant and redundant content,diverse syntactic forms,implicit evaluation objects and crossinterruption,a topic induction method for power dialogue text suppliers based on text intelligent mining technology is proposed.A single-round dialogue text next sentence predictive analysis method based on Transformer-based two-way encoder next sentence prediction and cosine similarity weighting is proposed,and a dialogue interruption cross-processing process and supplier identification rules are established to realize the topic induction of the power dialogue based on the equipment supplier.Through calculation examples,the proposed method for topic induction of electric dialogue text can effectively divide the topic of dialogue text,and its topic induction results lay the foundation for the establishment of subsequent evaluation models.3.Aiming at the semantic richness and complexity of power dialogue text,a supplier evaluation model based on text sentiment analysis is proposed.On the basis of expanding the entries and attributes of the existing electric power ontology dictionary,the dialogue sentiment analysis rules are established,and the supplier evaluation model based on sentiment analysis is adopted to realize the normalized evaluation of power equipment suppliers.The calculation example shows that the power equipment supplier evaluation based on the intelligent mining of dialogue text is feasible and effective,which can be used as a useful supplement to the current supplier evaluation method. |