Crank call identification method based on signaling data

Abstract

The invention discloses a crank call identification method based on signaling data, and the method specifically comprises the following steps: S1, collecting the signaling data of a calling user and a crank call blacklist of a telecommunication network end in real time, and presetting a threshold value X; s2, constructing a judgment and recognition model, and preprocessing and analyzing the signaling data of the calling subscriber based on the judgment and recognition model to obtain key features; and S3, calculating a threshold value M of the signaling data according to the key features, and if the threshold value M is greater than or equal to the threshold value X, identifying the calling user as a crank call by the judgment and identification model. The invention belongs to the technical field of crank call identification, and aims to solve the problem of poor identification effect on crank calls when signaling data is less in the prior art. The technical effects are that the problems of crank call identification lagging and limited identification range can be effectively solved according to the signaling data, and the crank call identification effect is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of harassing call identification, and in particular to a harassing call identification method based on signaling data. Background Art

[0002] With the development of communication technology, communication users often receive a large number of harassing calls from unfamiliar calls. Some illegal companies or individuals spread harassing information by making illegal calls. A large number of illegal calls not only occupy precious communication resources, but also directly lead to problems such as a decrease in connection rate and congestion of network equipment, greatly reducing the experience of legitimate mobile users and greatly affecting the work and life of mobile communication users. In order to avoid being disturbed by harassing calls, many users choose to manually or use harassing call identification technology to determine whether the incoming call number is a harassing call and reject it. However, it is easy to make mistakes in judgment and reject the call, which may cause users to miss important calls.

[0003] Signaling data is an important part of harassing call identification technology. It contains a large amount of records generated by the communication between mobile phones and mobile network base stations, which can provide information about user location, mobile behavior and communication patterns. By collecting signaling data, harassing call identification technology can obtain call information of user numbers, and then analyze call characteristics to identify harassing calls. The telecommunications network uses cloud-based interception technology to add the calling user to the harassing call blacklist.

[0004] Most of the existing harassing call identification technologies mainly rely on the comprehensive use of a large amount of data such as the call records, location information and behavior of the calling user in the signaling data to identify the harassing calls. However, when the harassing caller consciously controls the dialing time interval or makes fewer calls in the early stage during the dialing process, resulting in less signaling data of the calling user, there is a problem of recognition lag and limited recognition range, which reduces the recognition effect of the harassing calls. Therefore, the present invention proposes a harassing call identification method based on signaling data, aiming to improve the timeliness of recognition and expand the scope of recognition. Summary of the invention

[0005] To this end, the present invention provides a method for identifying harassing calls based on signaling data to solve the above-mentioned problems in the prior art.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] According to a first aspect of the present invention, a method for identifying harassing calls based on signaling data specifically comprises the following steps:

[0008] S1: collect the signaling data of the calling user and the harassing call blacklist of the telecommunications network in real time, and preset the threshold value X;

[0009] S2: Construct a judgment and recognition model, and pre-process and analyze the signaling data of the calling user based on the judgment and recognition model to obtain key features;

[0010] S3: Calculate the threshold M of the signaling data based on the key features. If the threshold M ≥ the threshold value X, the identification model will be used to identify the calling user as a nuisance call. If the threshold M < the threshold value X, the calling user is judged to be a number to be verified.

[0011] S4: constructing a first node graph of the number to be verified based on the judgment and recognition model;

[0012] S5: Match the historical dialing data of the calling user with the nuisance call blacklist, find all the records that have been in contact with the blacklist numbers and mark them, filter out the marked blacklist numbers and their related dialing information, and generate a feature set;

[0013] S6: Compare the feature set with the calling user signaling data, insert it into the first node graph, draw a second node graph, and input the connection result of the first node graph and the second node graph into the judgment and recognition model to identify whether it is a harassing call.

[0014] Furthermore, the calculation formula of the threshold M is:

[0015] M=A+T×Y

[0016] Among them, A is the mean of the key feature, T is the time period, and Y is the preset number of calls in the time period.

[0017] Furthermore, the calculation formula of the mean value A of the key feature is:

[0018]

[0019] Where X is the number of key features and n is the number of calls.

[0020] Furthermore, the signaling data in S1 includes the calling frequency of the calling party, the called party information and the calling party's geographical location, and the called party information includes the current called party information and the historical called party information.

[0021] Furthermore, the key features include the calling frequency, calling time, number of calls and base point location of the calling user.

[0022] Furthermore, the preprocessing and analysis require first cleaning the signaling data, removing duplicate, missing and abnormal signaling data, retaining the most discriminating distinguishing features, and then extracting key features from the distinguishing features, normalizing the key features through normalization processing technology, and converting all data values ​​​​of each feature to ensure that the integrity and consistency of the key feature data are maintained during the conversion process.

[0023] Furthermore, the first node graph regards the calling user and the called user as a node respectively, draws the first node graph through the programming library, connects every two nodes with a line segment to clearly show the relationship between the nodes, and regards the base point position of the calling user as a separate node.

[0024] Further, the dialing information includes the base location and contact frequency of the blacklisted number;

[0025] The base point location can provide accurate location data of the marked blacklist number;

[0026] The analysis of the contact frequency helps to identify whether the calling user frequently contacts the blacklisted number, thereby revealing a closer association or a higher risk of harassment.

[0027] Furthermore, the feature set of the marked blacklist number is regarded as a node. If the marked blacklist number is respectively connected with the calling user or the called user, and the contact time exceeds 5 seconds, each two nodes in the second node graph and the first node graph are connected by a line segment.

[0028] The present invention has the following advantages:

[0029] 1. Preliminary identification of harassing calls is performed by preprocessing and analyzing the signaling data of the calling user. If the signaling data is small, the historical dialing data of the calling user is matched with the harassing call blacklist, and the first node graph and the second node graph are constructed respectively. According to the first node graph and the second node graph, the connection between the calling user and the called user and the connection between the calling user and the harassing call blacklist can be effectively judged, and whether the calling user is a harassing call can be effectively judged. This effectively solves the problem of recognition lag and limited recognition range caused by the small amount of signaling data of the calling user, and improves the recognition effect of harassing calls. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A flowchart of a method for identifying harassing calls based on signaling data is provided for some embodiments of the present invention. DETAILED DESCRIPTION

[0031] The following is a description of the implementation of the present invention by specific embodiments. People familiar with the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] like Figure 1 As shown, a method for identifying harassing calls based on signaling data in the present invention specifically includes the following steps:

[0033] S1: collect the signaling data of the calling user and the harassing call blacklist of the telecommunications network in real time, and preset the threshold value X, wherein the signaling data includes the calling frequency, the called user information and the calling geographic location, and the called user information includes the called user information and the historical called user information;

[0034] S2: construct a judgment and recognition model, and pre-process and analyze the signaling data of the calling user based on the judgment and recognition model to obtain key features, where the key features include the calling user's call frequency, call time, number of calls, and base point location;

[0035] Preprocessing and analysis requires cleaning the signaling data first, removing duplicate, missing, and abnormal signaling data, retaining the most distinguishing features, extracting key features from the distinguishing features, normalizing the key features through normalization processing technology, and converting all data values ​​of each feature to ensure the integrity and consistency of the key feature data during the conversion process;

[0036] S3: Calculate the threshold M of the signaling data based on the key features. If the threshold M ≥ the threshold value X, the identification model will identify the calling user as a spam caller. If the threshold M < the threshold value X, the calling user is identified as a number to be verified. The threshold M represents the similarity between the calling user's behavior and the spam call behavior.

[0037] The calculation formula of the threshold M is:

[0038] M=A+T×Y

[0039] Where A is the mean of the key feature, T is the time period, and Y is the preset number of calls in the time period;

[0040] The calculation formula for the mean A of the key features is:

[0041]

[0042] Where X is the number of key features and n is the number of calls;

[0043] S4: constructing a first node graph of the number to be verified based on the judgment and recognition model, where the first node graph is used to display the relationship data and historical dialing data of the calling user;

[0044] The first node graph regards the calling user and the called user as a node respectively. The first node graph is drawn by the programming library, and each two nodes are connected by a line segment so as to clearly show the relationship between the nodes. Meanwhile, the base point position of the calling user is regarded as a separate node.

[0045] S5: Match the historical dialing data of the calling user with the nuisance call blacklist, find all the records that have been in contact with the blacklist number and mark them, filter out the marked blacklist number and its related dialing information, generate a feature set, and the dialing information includes the base point location and contact frequency of the blacklist number;

[0046] The base point location can provide the precise location data of the marked blacklisted numbers;

[0047] Analysis of contact frequency helps identify whether the caller frequently contacts blacklisted numbers, revealing a closer connection or higher risk of harassment;

[0048] S6: Compare the feature set with the calling user signaling data, insert it into the first node graph, draw a second node graph, and input the connection result of the first node graph and the second node graph into the judgment recognition model to identify whether it is a harassing call; regard the feature set of the marked blacklist number as a node, if the marked blacklist number is in contact with the calling user or the called user respectively, and the contact time exceeds 5 seconds, then

[0049] Every two nodes in the second node graph and the first node graph are connected by a line segment;

[0050] And determine whether the base point location of the feature set is within 200 meters of the base point location of the calling user. If so, connect the feature set node with the node of the base point location of the calling user, otherwise, the node is not connected; if the marked blacklist number nodes are connected to the nodes of the calling user and the called user respectively, and the nodes of the base point location are connected, then the calling user is judged to be a harassing call and added to the blacklist, otherwise, the calling user is judged to be a normal call.

Claims

1. A method for identifying harassing calls based on signaling data, characterized in that: The specific steps include: S1: collect the signaling data of the calling user and the harassing call blacklist of the telecommunications network in real time, and preset the threshold value X; S2: Construct a judgment and recognition model, and pre-process and analyze the signaling data of the calling user based on the judgment and recognition model to obtain key features; S3: Calculate the threshold M of the signaling data based on the key features. If the threshold M ≥ the threshold value X, the identification model will be used to identify the calling user as a nuisance call. If the threshold M < the threshold value X, the calling user is judged to be a number to be verified. S4: constructing a first node graph of the number to be verified based on the judgment and recognition model; S5: Match the historical dialing data of the calling user with the nuisance call blacklist, find all the records that have been in contact with the blacklist numbers and mark them, filter out the marked blacklist numbers and their related dialing information, and generate a feature set; S6: Compare the feature set with the calling user signaling data, insert it into the first node graph, draw a second node graph, and input the connection result of the first node graph and the second node graph into the judgment and recognition model to identify whether it is a harassing call.

2. According to a method for identifying harassing calls based on signaling data according to claim 1, it is characterized in that: The calculation formula of the threshold value M is: M=A+T×Y Among them, A is the mean of the key feature, T is the time period, and Y is the preset number of calls in the time period.

3. According to a method for identifying harassing calls based on signaling data according to claim 2, it is characterized in that: The calculation formula of the mean value A of the key feature is: Where X is the number of key features and n is the number of calls.

4. According to a method for identifying harassing calls based on signaling data according to claim 1, it is characterized in that: The signaling data in S1 includes the calling frequency of the calling party, the called party information and the calling party's geographical location. The called party information includes the current called party information and the historical called party information.

5. According to a method for identifying harassing calls based on signaling data according to claim 1, it is characterized in that: The key features include the calling frequency, calling time, number of calls and base point location of the calling user.

6. According to a method for identifying harassing calls based on signaling data according to claim 1, it is characterized in that: The preprocessing and analysis requires first cleaning the signaling data, removing duplicate, missing and abnormal signaling data, retaining the most discriminating distinguishing features, and then extracting key features from the distinguishing features. The key features are normalized through normalization processing technology, and all data values ​​of each feature are converted to ensure that the integrity and consistency of the key feature data are maintained during the conversion process.

7. According to a method for identifying harassing calls based on signaling data according to claim 1, it is characterized in that: The first node graph regards the calling user and the called user as a node respectively. The first node graph is drawn through a programming library, and each two nodes are connected by a line segment to clearly show the relationship between the nodes. At the same time, the base point position of the calling user is regarded as a separate node.

8. According to a method for identifying harassing calls based on signaling data according to claim 1, it is characterized in that: The dialing information includes the base location and contact frequency of the blacklisted number; The base point location can provide accurate location data of the marked blacklist number; The analysis of the contact frequency helps to identify whether the calling user frequently contacts the blacklisted number, thereby revealing a closer association or a higher risk of harassment.

9. According to a method for identifying harassing calls based on signaling data according to claim 1, it is characterized in that: The feature set of the marked blacklist number is regarded as a node. If the marked blacklist number is respectively connected with the calling user or the called user, and the contact time exceeds 5 seconds, each two nodes in the second node graph and the first node graph are connected by a line segment.

Images