RU2765483C2 - Method for recognizing and processing spam call - Google Patents

Abstract

FIELD: telephone communication.

SUBSTANCE: invention relates to telephone communication, in particular to methods of identifying, filtering and processing unwanted voice calls (spam calls) in cellular communication networks. In the method of recognizing and processing a spam call for use in cellular communication networks of GSM, IN, IMS standards, which, before connecting subscribers, provides for verification of the fact of connection or connection attempt of the calling and called subscribers for the period T, if such connection is not confirmed, the calling subscriber is routed and connected to the voice bot of the VMS server for recognition in the calling side of the voice bot or person, as well as advertising and/or fraudulent content.

EFFECT: technical result consists in simplification and improvement of reliability of spam-call recognition in signaling network of cellular communication operator.

1 cl, 3 dwg

Description

The invention relates to the field of telephone communications, in particular to methods for identifying, filtering and processing unwanted voice calls (spam calls) in cellular networks (clause 12, GOST R 53801-2010, 3GPP TS 23.002), a method without limiting the scope of legal protection can be used in GSM, IN, IMS networks.

Unwanted spam calls are known to be a significant inconvenience for cellular subscribers. Technically, spam calls are made automatically using a bot or manually by a call center employee. Economically, spam calls are for promotional or fraudulent purposes. To counter spam calls, subscribers use blacklists implemented at the operating system (OS) level of MS (Mobil Station) mobile phones. Cellular carriers are implementing "Who's Calling" services with caller information on the MS incoming call screen. There are mobile applications that have the "Who is calling" functionality. These services are powered by third-party databases (external analytics) that provide an API to retrieve data associated with a caller's MSISDN number. The functionality of these services is not sufficient, in any case, the subscriber is forced to independently make a decision to block a spam call and blacklist the incoming spam number. A generalized solution is needed that provides automatic reliable blocking of spam calls without the participation of the subscriber at the level of functional (network) objects of the cellular operator's network.

At the current level of technology, algorithms for recognizing a spam call in the operator's packet network are known by applying the rules for analyzing spam calls to the accumulated array of statistical data, including voice call events generated by the functional (network) objects of the operator's network when processing voice calls:

- identifier of the caller (calling) and called (called) subscribers;

- the reason for the failed call in the scenario called (called) subscriber is busy, unavailable, does not answer or dropped the call;

- the reason for the failed call in the script - the calling subscriber dropped the call, abnormal connection reset due to network errors;

- call duration;

- reason for terminating the call.

These algorithms mainly include spam call recognition features calculated from statistical data for period T:

- the ratio of the number of outgoing calls of subscriber A, the duration of which is greater than the specified one, to the total number of outgoing calls of subscriber A;

- the ratio of the number of outgoing calls of subscriber A to the number of incoming calls to subscriber A;

- the ratio of the number of incoming calls to subscriber B from subscriber A to the total number of incoming calls to subscriber B;

Technically, the main difficulty of this approach to recognizing a spam call is the formation of the necessary array of statistical data to calculate the signs of recognizing a spam call. To ensure sufficient accuracy, at best, such a data set should include the entire set of voice calls in a certain territory for a period T, and regardless of the telecom operator. But, for example, in the general case, the calling subscriber A and the called subscriber B can be registered in different networks. This means that the accumulation of sufficient statistics to calculate at least the first two features (in relation to subscriber A) for the operator of the called subscriber B is impossible. In this case, recognition of subscriber A's spam for the operator of subscriber B becomes at least difficult or, at worst, impossible. The complexity of accumulating an array of data on voice calls is exacerbated by the fact that spam subscribers often change numbers. This means that before the accumulation of statistical data for the period T and the determination of recognition signs from them, it becomes impossible to recognize a spam call.

Economically, the implementation of technical measures to recognize and block a spam call is complicated by the conflict of interests of the mobile operator associated with its processing. The inconsistency lies in the economic feasibility of combining an incoming spam call to generate income for its service and the commercial need to block a spam call under the terms of a subscriber's subscription to block it.

Taking into account the technical and economic conditions of the current state of the art, the authors propose a method for recognizing and processing a spam call, which simultaneously provides:

- a spam call recognition algorithm that does not require laborious accumulation and analysis of a statistical data array for a period T,

- increasing the economic performance of a cellular operator when executing the algorithm for recognizing a spam call and processing it.

In the context of the application, the following definitions were adopted: HomeNet subscriber - a subscriber registered in the network of a cellular operator NP (Network Provider). OffNet subscriber - a subscriber not registered in the NP network. Subscriber A (Calling) - the calling subscriber or calling party, subscriber B (Called) - the called subscriber or called party. In the general case, it is assumed that subscriber A is an offNet subscriber; in a particular case, subscriber A can be both homeNet and offNet subscribers. Incoming call - an incoming voice call made by subscriber A to subscriber B. Spam call or spam call - an incoming call recognized by the NP operator (without the participation of subscriber B's homeNet) as unwanted for subscriber B. Voice bot - a software algorithm for computer analysis and synthesis of texts in natural languages having an audio (voice) interface, including in an interactive form. Subscriber ID - unique standardized data used to process connections on the network nodes of the NP operator (SSP, SCP, HLR, VLR, BSS, etc.) and related to the provision of services and charging (MSISDN, IMEI, IMSI, etc.), in including PSTN (Public Switched Telephone Network) subscriber numbers. Call connection establishment of a signaling SS7 connection for making a conversation between subscribers A B. Signaling network NP - CS (Circuit Switched Domain) a set of objects (including ISUP, SIP, CAP protocols) that support SS7 signaling (SS7 - Common Signaling Channel No. 7 or Signaling System No. 7) and processed on functional (network) objects standardized 3GPP, ETSI - network nodes SSP, SCP, HLR, IP, etc. Packet network NP - PS (Packet Switched Domain) set of objects (GPRS network packets) with independent routing processed on standardized 3GPP, ETSI network nodes PDN GW, S-GW, MME, SGSN, etc. Core or core network NP - CN (Core Network) combining signaling and packet network NP (clause 3.2.3 GPP TS 23.002 ). IN (Intelligent Network) network - network architecture of the NP operator focused on the integration of additional communication services (ITU-T Q.1200 Rec.). IMS (IP Multimedia Subsystem) network - NP operator's network architecture based exclusively on IP networks (3GPP IMS). Separately, it is noted that the IMS standard excludes the CS signaling network from the network architecture of cellular communications, uses the SIP protocol to control communication sessions, and does not support the ISUP protocol. However, since in the IMS of the SIP network, traffic is processed at functional entities (network nodes), associated with MSISDN, IMEI, IMSI, and QoS requirements, in the context of the SIP application, IMS network traffic is assigned to the signaling network protocol.

Patent RU 2722685 is known from patent sources, according to which the mobile operator creates a list of advertising identifiers of subscribers A approved by the operator NP, and a list of advertising identifiers of subscribers A authorized by subscriber B. Such lists are generated by the NP operator over all switching channels (Call, SMS, Web) with subscriber B. The decision to connect the voice call of subscribers A and B is made by the operator only if the identifier (MSISDN) of subscriber A is in both lists. Thus, the operator achieves the connection of the subscriber only with useful voice calls and reliably blocks any spam calls. The solution appears to be reliable and useful. Nevertheless, within its homeNet, subscriber B must enter into the management system an entry about the categories of goods and services of interest to him in order to form a list of allowed advertising identifiers. Those. the subscriber indirectly participates in measures to block unwanted calls. For the operator NP, the inconvenience of this solution lies in the need to manually enter entries in the table of advertising identifiers A.

The closest in technical essence known patent US 10708416 which is chosen by the authors for the prototype. As described in the patent, the method includes receiving a first amount of spam call data from external analytics databases and a second amount of voice call data from a subscriber's MS call log. After that, the machine learning algorithm determines the signs of recognition of a spam call (parameters in terms of US patent 10708416) based on the first amount of data. From the entire list, the authors reveal two parameters:

- the ratio of unique (single or at least rare) incoming calls to subscriber B to the total number of incoming calls to subscriber B;

- the ratio of outgoing calls of subscriber A, the duration of which is greater than the specified to the total number of outgoing calls of subscriber A (success rate in terms of application US 10708416).

After these parameters have been generated by a machine learning algorithm, they are applied to a second amount of data for recognizing a spam call (identification in terms of US 10708416). To clarify the identification of a spam call, the authors of US 10708416 propose to generate a request to the subscriber's MS device to obtain information about the telephone number. Additionally, the authors of US 10708416 use the geodata of the MS of the caller to refine the classification.

The disadvantages of the solution US 10708416 follow from the above objective technical difficulties of recognizing a spam call:

- the need for accumulation and complex processing of an array of data for analysis,

- the need for software development of MS - installation of an anti-virus component,

- the impossibility of analyzing network events generated by the functional (network) objects of the signaling CS network, since the operation of the method is limited to the packet PS network,

- the inability to control the processing of spam calls in the signaling network, technically the method does not have access to the CS of the operator's signaling network.

Taking into account the objective technical difficulties of recognizing a spam call in the general set of voice calls and the shortcomings of technical solutions known from the current state of the art, the authors propose a method for recognizing and processing a spam call, which provides:

- a reliable spam call recognition algorithm in the conditions of variability of outgoing spam call numbers;

- no need for time-consuming accumulation and analysis of a data array to apply the spam call recognition algorithm;

- cost-effective processing of a spam call without the participation of the called (Called) subscriber B;

The method provides the claimed characteristics due to:

- operation of the method on the technical means of functional (network) objects of the signaling CS network of the operator;

- implementation of a method based on the author's spam call recognition algorithm, which is the know-how of the applicant.

The technical result of the method for recognizing and processing a spam call is to simplify and increase the reliability of recognizing a spam call in the signaling CS network of a cellular operator and to increase the economic performance of a cellular operator when processing a recognized spam call.

The technical result is achieved by a method for recognizing and processing a spam call, according to which the cellular operator receives at the switching node a SIP message, ISUP protocols to establish an incoming voice call connection from the calling party to the called party, routes the incoming voice call connection to the network node of an intelligent peripheral that performs the functions of VMS voice mail, and before routing the connection of an incoming voice call to the node of intelligent peripherals, the cellular operator, using the identifiers of the calling and called parties using the SIP, ISUP, CAP, DIAMETER protocols, checks the fact that the connection of the calling and called parties has been established for a period T equal to 30 days, if there is no confirmation of the fact of the connection of the calling and called parties for the period T, the cellular operator connects the incoming call to the program logic of computer analysis and synthesis of texts in natural languages of the network node of intelligent peripherals having a voice audio interface, and the program logic of computer analysis and synthesis of texts in natural languages of the intelligent peripheral node, which has a voice audio interface, generates statements in the affirmative, interrogative and imperative form and searches for patterns in the statements of the caller to recognize the voice bot or person in the caller , as well as promotional or fraudulent content.

The invention is illustrated by drawings:

In FIG. 1 shows a simplified NP signaling CS architecture with an integrated intelligent spam call recognition platform.

In FIG. 2 shows a generalized block diagram of a method for recognizing and processing a spam call.

In FIG. 3 shows a generalized dialog of functional network nodes when executing the method for recognizing and processing a spam call.

In the figures, for simplicity, the network nodes IN and IMS networks performing similar functions in the context of the application are made in single functional blocks and are denoted by a fraction. The IN and IMS signaling connections of the network are indicated by dashed lines.

In FIG. Figure 1 shows the generalized architecture of the signaling network CS of the operator NP in whose network the called homeNet subscriber MS B is registered, served by the current switching node SSP (Service Switching Point) CSCF / VMSC, which performs the functions of switching signaling connections using the signaling protocols SIP, ISUP. In the IMS network, SSP functions are performed by a standard SIP server that performs P/I/S-CSCF (Call/Session Control Function) SIP session management functions. In the IN network, the SSP functions are performed by the VMSC (Visitor Mobile Switching Center) network node.

SCP PPS (Service Control Point Prepaid Pay System) is a platform for charging subscribers with advance billing, in IN and IMS networks it performs the functions of an Online Charging System (OCS) billing.

The IP VMS (Intelligent Peripheral Voicemail System) platform is represented by a standard voice mail network node - a system for registering, storing and forwarding telephone voice messages. In the IMS network, the functions of the IP VMS network node are performed by the classic SIP AS (Application Server) network node. In the IN network, the IP VMS implements the VAS (Value Added Services) logic. Within the framework of the claimed method, IP VMS is a technical tool that performs the logic of computer analysis and synthesis of texts in natural languages with an audio (voice) interface, including an interactive one. To reproduce utterances, the text synthesis program logic contains a list of pre-recorded utterances, including those that stimulate the interlocutor to respond. The natural language analysis software logic includes templates and keywords for recognition on the calling side A of a voice bot or person, as well as advertising or fraudulent content. The spam call recognition software algorithm based on the analysis and synthesis of texts in natural languages with a voice audio interface is the know-how of the applicant and performs the recognition of a voice bot, advertising and fraudulent content of statements with a given reliability.

STAS/XSCP is a service control network node, performs SCF (Service Control Function) and data control functions for providing SDF (Service Data Function) services. The SDF function for checking the fact of connection, calling (Calling) and called (Called) subscribers for the period T can use the records of the internal database about the call history and / or interact with the records about the call history of the operator CDR (Charging Data Record). Within the framework of the claimed method, STAS/XSCP, together with the IP VMS node, performs the functions of recognizing a spam call. In the IN network, it is represented by the XSCP intelligent platform. In the IMS network, it is represented by an AS application server STAS (Smart Telecom Application Server).

MS A is the calling subscriber, in FIG. 1, 3 is designated as offnet, without loss of generality of the method and scope of legal protection, it can be a homeNet subscriber A or a subscriber of the PSTN network. In a private implementation of the method, without limiting its essence and scope of legal protection, the functions of the STAS/XSCP intelligent platform can be performed by the switching node SSP VMSC or the CSCF node for IMS networks. Such an implementation is less preferable because it violates the principle of separating the functions of switching connections and providing services.

In FIG. 2 shows a generalized block diagram of the algorithm for recognizing and processing spam calls.

At stage 1, STAS/XSCP performs preliminary recognition of a spam call according to the CAPCHA condition for whether the call belongs to a spam call. The CAPCHA condition in the context of the application provides for checking the fact of connection or attempt to connect subscribers A and B for the period T. In various software implementations, the period T can vary from a week to six months. The authors consider the optimal period T to be 30 days. This value is determined statistically and, in the general case, makes it possible to form an assumption about the unknown calling (Calling) subscriber A for the called (Called) subscriber B due to the absence of communication between them and / or attempts to communicate during this period. When confirming the fact of connection between A and B subscribers in period T, the software algorithm STAS / XSCP performs a regular call connection with subscriber B 6.

In a particular case, such as when calling from a delivery service, the CAPCHA prefix is neither informative nor sufficient. Therefore, the method provides for the implementation of the main necessary stages of recognition 2-4 and processing 5 of a spam call on the IP VMS. To do this, the STAS / XSCP software algorithm performs call routing to IP VMS, which launches algorithms for recognizing voice bot 2 or person 3, as well as advertising and / or fraudulent content (2 and 3) and performs a software algorithm to keep a spam call on the connection for the maximum possible time 5 At stages 2-3, the IP VMS software algorithm for synthesizing texts in natural languages, having an audio interface from the list of pre-recorded utterances, reproduces utterances to the Calling subscriber A in interrogative and incentive forms. At the same time, the methods for recognizing a voice pain and a person are different. For example, in the simplest case, voice bot recognition is performed by playing the utterance during the playback of the audio recording of the voice bot and confirming that the recording of the voice bot utterance continues to play continuously, i.e. the lack of a human reaction of the interlocutor to the interruption of the utterance.

After recording and converting to text the response statements of caller A, the algorithm performs an analysis that includes searching for patterns and keywords in the statements to recognize voice bot 2, person 3, and in both cases, advertising and / or fraudulent content. According to the result of the analysis carried out at stages 2, 3, the software algorithm at stage 4 determines whether the incoming call belongs to a spam call. In the case of membership (spam), the subsequent utterance generation and parsing algorithm in step 5 reproduces utterances in affirmative, interrogative, and imperative form to stimulate the caller's response and aims to keep the spam call on the connection for the maximum possible time tmax to generate operator revenue for servicing the charging call max. If the software algorithm in step 4 does not find patterns and keywords in the statements of the calling party (Calling) that are signs of advertising and / or fraudulent content, then the incoming call is recognized as valid (no spam) and the IP VMS software algorithm routes the call to connect with the called subscriber AT 6.

Figure 3 illustrates a generalized dialogue of functional network nodes within the implementation of the claimed method for IN and IMS networks using the SIP protocol as an example. In the IN network, when the VMSC (SSP) SIP receives a request to establish an incoming voice call 1 INVITE (Calling A, Called A), the T-CSI (Terminating CAMEL Subscription Information) trigger for processing incoming calls IN is triggered: T-CSI trigger init. The trigger initiates the routing of an incoming call to an XSCP address. In the IMS network, call 1 INVITE (Calling A, Called A) is routed from the CSCF (SSP) node to STAS according to the IFC (Initial Filter Criteria) subscriber's rule subscribed to the IMS spam call blocking service: IFC STAS rout.

In the IN network, the XSCP exchanges with the SCP 3 INITIAL DP to verify the subscriber's subscription to the spam call blocking service. In the case of a 4 APPLY CHARGING subscription confirmation, the STAS/XSCP software algorithm performs preliminary recognition of a spam call according to the CAPCHA init condition (flow diagram of FIG. 2, step 1). When the CAPCHA condition is passed, the NO SPAM DETECT script is executed and the call is connected to the called (Called) subscriber B 10 INVITE (Calling A, Called A). If the CAPCHA condition is not passed, the DETECT SPAM script is executed, which provides for routing the call to the IP VMS to perform the basic steps of recognizing a spam call. IP VMS performs a 6 200 OK, 7 ACK voice call connection with Calling Subscriber A and launches a spam call recognition algorithm based on computer analysis and synthesis of texts in natural languages, having an audio bot startup voice audio interface (flow diagram of Fig. 2 steps 2-4). According to the result of the main steps of the spam call recognition algorithm (block diagram of Fig. 2 steps 2-4), a connection is made with the called (Called) subscriber B 11 INVITE (Calling A, Called A) if the call is recognized as appropriate (Fig. 2 step 4 no spam, step 6). If the call is recognized as a spam call, the charging max startup algorithm is executed to hold the call 8 VOICE PHASE on the connection for the maximum possible time tmax until the message 9 BYE is received from the calling subscriber A.

The proposed method for processing a recognized spam call may include notification of subscriber A's homeNet about processed spam calls made without his participation, via any switching channels: voice mail messages, SMS, PUSH notification. For a detailed report, the method may include transmitting a transcript of the recognition dialog on the IP VMS node to the messenger and/or homeNet social network of subscriber A.

The technical effect of the method disclosed above for recognizing and processing a spam call in the above implementations, in contrast to the known solutions, provides:

- improving the accuracy of spam call recognition due to the insensitivity of the method to the variability of outgoing numbers of spam calls;

- removal of the need to generate and analyze a large amount of data to apply the spam call recognition algorithm;

- recognition of automatic calling by a voice bot and manual calling with the participation of a person.

- generating additional revenue for the operator by connecting and keeping a spam call on the connection for the maximum possible time without the participation of subscriber B's homeNet.

The claimed method fits well into the network architecture of the IN network based on the network service control node SCP and the network architecture of the IMS network based on the standard AS SIP server. The method has been tested by the applicant and provides the claimed technical result with a given reliability and meets the requirements of quality of service QoS in cellular networks of GSM, IN, IMS standards.

Claims (2)

1. A method for recognizing and processing a spam call, according to which the cellular operator receives at the switching node a message to establish an incoming voice call connection from the calling party to the called party, routes the connection of the incoming voice call to the network node of intelligent peripherals, characterized in that up to routing the connection of an incoming voice call to an intelligent peripheral node, the cellular operator, using the identifiers of the calling and called parties, checks the fact of establishing a connection between the calling and called parties for the period T, if the fact of the connection between the calling and called parties is not confirmed for the period T, the cellular operator connects the incoming call to the program logic computer analysis and synthesis of texts in natural languages of the network node of the intellectual periphery with a voice audio interface, and the program logic of computer analysis and synthesis of texts in natural languages of the node of the intellectual periphery The series, which has a voice audio interface, generates utterances in affirmative, interrogative, and imperative forms and searches for patterns in the utterances of the caller to recognize the caller as a voice bot or person, as well as advertising or fraudulent content. 2. A method for recognizing and processing a spam call according to claim 1, characterized in that the establishment of an incoming voice call connection is a message of the SIP, ISUP protocols, the intelligent peripheral node is a VMS voice mail network node, checking the fact of the connection of the calling and called parties for the period T is performed using the SIP, ISUP, CAP, DIAMETER protocols, and the verification period T is 30 days.

Images