A method for actively predicting emergencies in the field of mobile customer service based on singlepass algorithm to realize unlabeled corpus

Abstract

The invention proposes a method based on the SinglePsss algorithm to realize the active prediction of emergencies in the field of mobile customer service with unmarked corpus. It mainly includes six stages: preprocessing, constructing sentence vectors, clustering based on the SinglePass algorithm, finding unexpected words, calculating the correlation of unexpected words, and returning emergency work orders. In the preprocessing stage, the work order data in the customer service field is preprocessed such as word segmentation, stop word removal, and word vector training. In the stage of constructing the sentence vector, the vector addition method is used to add the word vectors of each word in the sentence to obtain the sentence vector. In the clustering stage based on the SinglePass algorithm, based on the sentence vector, the SinglePass algorithm is used to cluster the work order data to obtain the topic of the emergency. In the stage of finding sudden words, the sudden words under each topic are obtained based on the time series algorithm (Kleinberg). In the phase of calculating the relevance of sudden words, mutual information is used to obtain the correlation between sudden words and the popularity of sudden topics. In the stage of returning emergency work orders, return the work orders containing the most relevant sudden word in the most popular text of the sudden topic, that is, get the sudden event.

Description

technical field

[0001] The invention relates to a method for proactively predicting emergencies in the field of mobile customer service in the field of natural language processing. It specifically relates to a method based on the SinglePass algorithm to realize unlabeled corpus to actively predict emergencies in the field of mobile customer service. Mainly related to patent classification number G06 Calculation; Calculation; Counting G06F Electrical digital data processing G06F17 / 00 Digital computing equipment or data processing equipment or data processing methods especially suitable for specific functions G06F17 / 30 Information retrieval; and its database structure. Background technique

[0002] With the explosive growth of data and the rapid development of information technology, how to discover the potential intentions of users through massive data and actively predict emergencies has become the focus of attention. Emergencies refer to events that occur suddenly and attr...

Images