Crank call identification method and crank call identification device based on deep artificial neural network

Abstract

The invention discloses a crank call identification method and a crank call identification device based on a deep artificial neural network, wherein the method comprises the following steps: collecting stranger call records to establish a training set; extracting behavior information of each stranger call record in the training set to generate a multi-dimensional vector; establishing the deep artificial neural network, and training the deep artificial neural network with the stranger call records in the training set; acquiring the behavior information of a stranger call number to generate themulti-dimensional vector and inputting the multi-dimensional vector into the trained deep artificial neural network, and judging whether the stranger call number is a crank call according to a characteristic value obtained by an output layer of the deep artificial neural network. Thus, manually extracting feature codes of the crank call through consuming resources is not needed, and a problem of identification of the crank call is solved effectively.

Description

technical field [0001] The invention examples of this application relate to machine learning and data mining methods in the field of computers, and in particular to a method and device for identifying harassing calls based on a deep neural network. Background technique [0002] Machine learning classification algorithms can be used to predict categories or single instances of categorical data, where the goal of binary classification is to predict one of two outcomes, for example: an email filter uses binary classification to determine whether an email is spam ; the other is multi-class classification, where the goal is to predict one of many outcomes; while the output of a classification algorithm is called a classifier, which can be used to predict the label of a new (unlabeled) instance. The advancement of machine learning technology in recent years has led to a wide range of applications, such as recommendation engines, targeted advertising, medical diagnosis, natural lan...

AI Technical Summary

Problems solved by technology

[0004] At present, most of the relevant technologies are to directly collect relevant suspicious call numbers to create a black and white list to intercept them. However, due to the advancement of caller number forgery technology, this method obviously cannot intercept and block in real time; tree, random forest and other calculation methods, hoping to achieve real-time blocking, but this is only applicable to specific application scenarios in a few countries. It is obviously more difficult to deal with the increasing number of harassing calls in various countries and regions around the world. insufficient

Images