Method for discovering sensitive phone call in phone bill information based on SVM classification algorithm

Abstract

The invention discloses a method for discovering a sensitive phone call in phone bill information based on a SVM classification algorithm. The method includes a first step of finding out, from phone bill big data, phone numbers having a series of sensitive features and fundamental differences from general and ordinary phone calls in a phone bill of a user; a second step of performing mining and analysis of the phone bill big data of a telecom operator, and extracting call record feature information of different numbers in the phone bill big data of the user; a third step of using the feature information as eigenvectors for training a SVM classification model, selecting a certain amount of known phone bill data from the phone bill big data, and training the SVM classification model; and a fourth step of using the SVM classification algorithm model to distinguish the possible sensitive phone numbers in the phone bill from the normal numbers and finding out suspicious phone calls.

Description

technical field [0001] The invention belongs to a pattern recognition method, in particular to a method for discovering sensitive phones in bill information based on an SVM classification algorithm. Background technique [0002] As artificial intelligence becomes more and more mature, it has been more practically applied in production and life. As the most core algorithm in artificial intelligence technology, machine learning technology has further attracted people's attention and has become a research hotspot in pattern recognition and classification algorithms. Among them, the statistical learning theory has been widely used since it was proposed, and the classification algorithm based on SVM has received extensive attention and good development because of its simple structure, strong generalization ability, short learning and prediction time, and the ability to achieve global optimal performance. . In particular, the SVM classification algorithm has great advantages in ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the inaccurate identification of traditional sensitive calls, a sensitive call discovery method based on SVM classification algorithm bill information is proposed, by mining the fundamental difference between sensitive calls and other ordinary call bills in the big data of call bills As a feature vector, based on the SVM classification algorithm, the sensitive phone numbers in the dialogue big data are distinguished from the normal phone numbers, so that the obtained SVM classifier model has better classification effect and higher classification accuracy

Images