LDA model-based customer demand identification method for power supply service

Abstract

The invention relates to a pattern recognition method, in particular to a power supply service customer claim recognition method based on an LDA model. The recognition method comprises the following steps: text preprocessing: text segmentation and text quantization are carried out on the text in the work order, and the long content of the text is mainly segmented according to certain rules; LDA topic generation: randomly select samples as training samples and test samples in the total sample, through the training of training samples to obtain performance parameters, and then test the test samples to obtain performance parameters with high recognition rate, and use the performance parameters to generate recognition text to identify the customer claims contained in the work order; Automatictext categorization. The invention has the advantages that the LDA subject model is utilized in identifying a plurality of customer demands in a full-quantity customer service work order, and text data is fully mined and applied to practical work.

Description

Technical field: [0001] The invention relates to a method for pattern recognition, in particular to a method for recognizing demands of power supply service customers based on an LDA model. Background technique: [0002] With the rapid development of information technologies such as Internet +, big data, and cloud computing, most information has transitioned from paper carriers to electronic carriers, and most of these information are unstructured or semi-structured text information. How to effectively manage, mine, and analyze the information contained in massive unstructured data has become a major challenge in the field of big data. In unstructured data, text data occupies an important position. For enterprises with a large amount of text data, how to effectively use this part of data resources determines the future development of the enterprise. In the data of the customer service center in the electric power industry, how to process the work order data, so as to accur...

AI Technical Summary

Problems solved by technology

[0004] In the work orders of electric power customers, the customer demands of each work order are not all single demands, and it is particularly important to accurately identify all the demands of each work order; in the text classification technology of machine learning classification, the machine learning classification algorithm can only identify A single appeal cannot satisfy the situation that a work order contains multiple demands, and the LDA model in the topic model can automatically and accurately identify multiple topics in the text, so the invention innovatively combines the LDA model with the work order data to achieve accurate Identify all customer claims in each work order

Images