Automatic work order classification method for electricity marketing service hot spot 95598

Abstract

The invention discloses an automatic work order classification method for the electricity marketing service hot spot 95598.The concept of 'increment' and 'normalization' is introduced into the TF-IDF method, 'increment' and 'normalization' are considered as a variable in the calculation process of feature vector weight, feature selection is conducted on the basis for work orders of the electricity marketing service hot spot 95598, and a training set is formed with the service hot spot as the base point; compared with various text mining and classifying algorithms, an algorithm with the optimal classification effect is selected, a classifier model is formed, work orders of 95598 are classified, service hot spot events of electricity marketing can be mined in time, and lean management of electricity marketing is supported.

Description

technical field [0001] The invention relates to an automatic classification method for 95598 work orders of electric power marketing service, in particular to an automatic classification method for 95598 work orders of electric power marketing service based on TF-IDF algorithm, and belongs to the technical fields of information collection, information filtering, information classification and the like. Background technique [0002] With the popularization of the application of computer technology, a large amount of text information began to exist in the form of computer readable and writable, thus computer automatic text classification technology came into being. Currently commonly used text classification methods mainly include: decision tree classification method, support vector machine method (SVM method), naive Bayesian classification method, K-nearest neighbor method (KNN), neural network method, fuzzy classification method, Rocchio Classification methods and Boosting a...

AI Technical Summary

Problems solved by technology

However, the TF-IDF algorithm and many improved TF-IDF algorithms often regard each word component as a static feature item, ignoring the impact of part of the short text information and the increase or decrease of feature items on the classification accuracy, which cannot meet the requirements for the classification accuracy. Classification Requirements for 95598 Work Orders of Power Marketing Service Hotspots

Images