Target user identification method and device, equipment, storage medium and product

Abstract

The invention discloses a target user identification method and device, equipment, a storage medium and a product. The method comprises the following steps: acquiring first internet behavior characteristic data of a first user, wherein the first internet behavior characteristic data comprises an actual identification result of the first user; inputting the first internet behavior feature data into an association rule model, generating a first association rule of the first internet behavior feature data by using the association rule model, and determining a first confidence coefficient of the first association rule; screening out a second association rule of which the first confidence is greater than a confidence threshold from the first association rules; and the third association rule with the target recognition result in the actual recognition result as the rule result is screened out from the second association rules, so that the target user is recognized by using the third association rule, and the accuracy of PCDN user recognition is improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, device, equipment, storage medium and product for identifying a target user. Background Art

[0002] Peer to Peer Content Delivery Network (PCDN) services based on point-to-point technology are content distribution network services built by enterprises or individuals by renting a large amount of broadband from operators and exploiting the massive fragmented idle resources of edge networks. These businesses arbitrarily change the use of broadband and aggregate their own upstream bandwidth to form a large bandwidth for traffic management. This has led to the gradual transfer of traffic originally belonging to the Internet Data Center (IDC) computer room to home broadband or dedicated lines, causing many troubles for operators' supervision and causing serious impact on operators' normal business. Therefore, it is crucial to accurately identify PCDN users.

[0003] In the prior art, PCDN users are identified based on simple traffic feature data such as large uplink traffic and high uplink and downlink traffic ratio of PCDN users. However, relying solely on uplink and downlink traffic to identify PCDN users will result in a large number of false positives and false negatives, and the accuracy of identifying PCDN users is low. Summary of the invention

[0004] The embodiments of the present application provide a method, apparatus, device, storage medium and product for identifying a target user, which can solve the problem of low accuracy in PCDN user identification.

[0005] In a first aspect, an embodiment of the present application provides a method for identifying a target user, comprising:

[0006] Acquire first online behavior characteristic data of the first user, where the first online behavior characteristic data includes an actual identification result of the first user;

[0007] Inputting the first online behavior characteristic data into the association rule model, generating a first association rule of the first online behavior characteristic data using the association rule model, and determining a first confidence of the first association rule using relationship information between a preset association rule and a preset confidence;

[0008] Filtering out, from the first association rules, second association rules whose first confidence is greater than a confidence threshold;

[0009] From the second association rule, filter out the third association rule with the target recognition result in the actual recognition result as the rule result, which is used to identify the target user by using the third association rule. The target recognition result indicates that the user is a content delivery network (PCDN) user based on the point-to-point technology, and the target user is a PCDN user.

[0010] In a possible implementation, input the first Internet behavior feature data into the association rule model, generate the first association rule of the first Internet behavior feature data by using the association rule model, and determine the first confidence of the first association rule by using the relationship information between the preset association rule and the preset confidence, including:

[0011] Input the first Internet behavior feature data into the association rule model, filter out the second Internet behavior feature data with the occurrence times not less than the initial support threshold in the first Internet behavior feature data, generate the fourth association rule of the second Internet behavior feature data, and determine the second confidence of the fourth association rule by using the relationship information between the preset association rule and the preset confidence;

[0012] From the fourth association rule, filter out the fifth association rule with the second confidence greater than the initial confidence threshold;

[0013] From the fifth association rule, filter out the sixth association rule with the target recognition result in the actual recognition result as the rule result;

[0014] Use the first Internet behavior feature data to determine the accuracy of the sixth association rule. When the accuracy does not meet the training stop condition, adjust the initial support threshold and the initial confidence threshold to update the second Internet behavior feature data, the fourth association rule, the fifth association rule, and the sixth association rule, and update the accuracy by using the updated sixth association rule;

[0015] When the updated accuracy meets the training stop condition, obtain the support threshold and the confidence threshold;

[0016] Filter out the third Internet behavior feature data with the occurrence times not less than the support threshold in the first Internet behavior feature data, generate the first association rule of the third Internet behavior feature data, and determine the first confidence of the first association rule by using the relationship information between the preset association rule and the preset confidence.

[0017] In a possible implementation, obtain the first Internet behavior feature data of the first user, including:

[0018] Obtain the fourth Internet behavior feature data of the first user;

[0019] Based on the fourth Internet behavior feature data, construct the constructed feature data of the first user by using the preset feature construction relationship information;

[0020] Combine the target Internet behavior feature data in the combined construction feature data and the fourth Internet behavior feature data to obtain the fifth Internet behavior feature data;

[0021] Discretize each sub-Internet behavior feature data in the fifth Internet behavior feature data respectively to obtain the first Internet behavior feature data.

[0022] In a possible implementation embodiment, obtaining the fourth Internet behavior feature data of the first user includes:

[0023] Obtain the sixth Internet behavior feature data of the first user;

[0024] Display the sixth Internet behavior feature data;

[0025] Receive a selection input for the sixth Internet behavior feature data;

[0026] In response to the selection input, obtain the fourth Internet behavior feature data of the first user.

[0027] In a possible implementation embodiment, obtaining the sixth Internet behavior feature data of the first user includes:

[0028] Obtain the initial Internet behavior feature data of the second user;

[0029] Delete the initial Internet behavior feature data of the third user with missing values in the initial Internet behavior feature data to obtain the sixth Internet behavior feature data of the first user;

[0030] Wherein, the second user includes the first user and the third user.

[0031] In a possible implementation embodiment, in response to the selection input, obtaining the fourth Internet behavior feature data of the first user includes:

[0032] In response to the selection input, obtain the seventh Internet behavior feature data of the first user;

[0033] Calculate the correlation coefficient between any two Internet behavior feature data in the seventh Internet behavior feature data;

[0034] When the correlation coefficient is greater than the preset threshold, delete one of any two Internet behavior feature data to obtain the fourth Internet behavior feature data of the first user.

[0035] In a second aspect, an embodiment of the present application provides an identification device for a target user, including:

[0036] An acquisition module, configured to acquire the first Internet behavior feature data of the first user, where the first Internet behavior feature data includes the actual identification result of the first user;

[0037] A determination module, configured to input first Internet behavior feature data into an association rule model, generate a first association rule of the first Internet behavior feature data by using the association rule model, and determine a first confidence level of the first association rule by using relationship information between a preset association rule and a preset confidence level;

[0038] A screening module, configured to screen out a second association rule from the first association rules, where the second association rule has a first confidence level greater than a confidence level threshold;

[0039] The screening module is further configured to screen out a third association rule from the second association rules, where the third association rule has a target recognition result in an actual recognition result as a rule result, so as to use the third association rule to recognize a target user, the target recognition result indicates that the user is a PCDN user, and the target user is a PCDN user.

[0040] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0041] A processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the method for recognizing a target user according to any one of the above is implemented.

[0042] In a fourth aspect, an embodiment of the present application provides a computer storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the method for recognizing a target user according to any one of the above is implemented.

[0043] In a fifth aspect, an embodiment of the present application provides a computer program product, and when instructions in the computer program product are executed by a processor of an electronic device, the electronic device can execute the method for recognizing a target user according to any one of the above.

[0044] Method, apparatus, device, storage medium and product for identifying target users in embodiments of the present application. Obtain first Internet behavior characteristic data of a first user, where the first Internet behavior characteristic data includes an actual identification result of the first user; input the first Internet behavior characteristic data into an association rule model, use the association rule model to generate a first association rule for the first Internet behavior characteristic data, and use relationship information between a preset association rule and a preset confidence level to determine a first confidence level of the first association rule; from the first association rules, filter out second association rules whose first confidence level is greater than a confidence level threshold; from the second association rules, filter out third association rules with a target identification result in the actual identification result as the rule result, so as to use the third association rules to identify target users, where the target identification result indicates that the user is a content delivery network (PCDN) user based on the peer-to-peer technology, and the target user is a PCDN user. Based on the Internet behavior characteristic data of users, through association rule mining, determine association rules for identifying PCDN users, and use the association rules to identify PCDN users, improving the accuracy of PCDN user identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] To more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required to be used in the embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0046] Figure 1 It is a flowchart of a method for identifying target users provided by an embodiment of the present application;

[0047] Figure 2 It is a flowchart of a method for identifying target users provided by another embodiment of the present application;

[0048] Figure 3 It is a flowchart of a method for identifying target users provided by still another embodiment of the present application;

[0049] Figure 4 It is a flowchart of a method for identifying target users provided by yet another embodiment of the present application;

[0050] Figure 5 It is a schematic structural diagram of an apparatus for identifying target users provided by yet another embodiment of the present application;

[0051] Figure 6 It is a schematic structural diagram of an electronic device provided by yet another embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.

[0053] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0054] It should be noted that the acquisition, storage, use and processing of data in the embodiments of the present application are in compliance with the relevant provisions of national laws and regulations.

[0055] PCDN services are content distribution network services built by enterprises or individuals who rent a large amount of broadband from operators and exploit the massive fragmented idle resources of edge networks. These services arbitrarily change the use of broadband and aggregate their own upstream bandwidth to form a large bandwidth for traffic management. It allows users to directly obtain the required content from each other's computers without the need to transmit through a centralized server, reducing the need to access the central server. This has led to the gradual transfer of traffic originally belonging to the IDC computer room to home broadband or dedicated lines, causing many troubles for operators' supervision and causing serious impact on the normal business of operators. Therefore, it is crucial to accurately identify PCDN users.

[0056] In the prior art, PCDN users are identified based on simple traffic feature data such as large uplink traffic and high uplink and downlink traffic ratio of PCDN users. However, relying solely on uplink and downlink traffic to identify PCDN users will result in a large number of false positives and false negatives, and the accuracy of identifying PCDN users is low.

[0057] It should be noted that the Peer-To-Peer (P2P) network is a network structure for point-to-point transmission, which breaks the traditional client or server model. It is a content sharing method based on peer nodes' decentralized services. Each node (i.e., Peer) is in an equal state. An individual can act as both a client and a server, and resource sharing can be achieved between them. Simply put, P2P means that data transmission no longer goes through a server, but directly between network users. The biggest advantage of P2P is that it can improve the network utilization rate of network users. Since multiple nodes are interconnected, the network bandwidth where the user is located will be used to the greatest extent.

[0058] The core idea of the design of the Content Delivery Network (CDN) is that if the content is far from the user, the user may not be able to obtain a timely response. Therefore, as a content provider, one should find a way to cache the content closer to the user. The CDN publishes the content to the network edge closer to the user and schedules the user requests to the content node server closest to the content according to the user's location, enabling the user to obtain the content nearby, thereby reducing the network pressure on the central server and improving the user access response speed.

[0059] The core design idea of the PCDN technology is to introduce a new P2P content network autonomous domain at the edge node layer of the traditional CDN network, and utilize the massive fragmented edge user resources to build a low-cost content delivery network service. Through a single or multiple user edge node terminal devices and the end users covered by them as peer content entities, a P2P content delivery service layer is jointly formed. The PCDN service makes the content closer to the user, the transmission more reliable, and the cost of content sharing and distribution lower. When the PCDN provides external services, the CDN is regarded as a super seed, similar to the central super node in the traditional Peer to Server&Peer (P2SP) architecture. At the same time, it utilizes P2P, and there are also many P2P nodes providing services. Although the ability of a single P2P node is far weaker than that of the CDN, relying on a large number of P2P nodes can better ensure the reliability of the entire network, and it is stronger than the traditional P2P service in terms of reliability.

[0060] PCDN currently supports various typical business scenarios such as video on demand, video live broadcast, and large file download. The specific application scenarios are as follows:

[0061] Video on demand: Long video on demand, short video on demand with concentrated popularity;

[0062] Video live broadcast: Including various large-scale evening parties, sports events live broadcasts, local network station live broadcasts, interactive live broadcasts, etc.;

[0063] Large file download: mainly applied to file distribution where the files are large and the popularity is relatively concentrated, such as in mobile application stores, online audio download and playback services, etc.

[0064] By configuring the CDN, the operator caches the content on the servers in each region, enabling nearby users to obtain the content without connecting to the main server. In this way, the operator can serve more users without expanding the bandwidth.

[0065] With the emergence of PCDN, the idle upstream network bandwidth of each broadband user is used as a micro CDN distribution service node, enabling other users to obtain a nearby acceleration experience in scenarios such as downloading, live streaming, and gaming. There are two key issues here. One is that it requires occupying the upstream bandwidth, and the other is that it basically no longer requires the operator's CDN for video acceleration, which brings many troubles to the operator's supervision and seriously affects the operator's normal business. To solve the problems of the existing technology, the embodiments of the present application provide a method, device, equipment, storage medium, and product for identifying target users. Based on the user's Internet behavior characteristic data, through association rule mining, the association rules for identifying PCDN users are determined, and the PCDN users are identified using the association rules, improving the accuracy of PCDN user identification.

[0066] The following introduces the method for identifying target users provided by the embodiments of the present application. As Figure 1 shown, the method for identifying target users provided by the embodiments of the present application includes the following steps S110 to S140.

[0067] S110. Obtain the first Internet behavior characteristic data of the first user, where the first Internet behavior characteristic data includes the actual identification result of the first user.

[0068] Among them, the first Internet behavior characteristic data can be data after discretization processing. It should be noted that the Internet behavior characteristic data includes characteristic fields and sub-Internet behavior characteristic data, and the sub-Internet behavior characteristic data is the field value of the characteristic field.

[0069] In some embodiments, the first Internet behavior characteristic data after discretization processing is represented by range identifiers. For example, the different sub-Internet behavior characteristic data of the first target characteristic field are represented as A1, A2, A3, A4; the different sub-Internet behavior characteristic data of the second target characteristic field are represented as B1, B2, B3, B4.

[0070] As an example, the first target characteristic field is the proportion of traffic in the early morning, and the range identifiers A1, A2, A3, A4 respectively correspond to values in the ranges of 0-a, a-b, b-c, c-1. Among them, a, b, c are numbers between 0 and 1, and a < b < c.

[0071] In another example, A2 represents a value of the processed early morning traffic proportion data within the range of (0.19, 0.38], and A3 represents a value of the processed early morning traffic proportion data within the range of (0.38, 0.69]; B3 represents a value of the processed upstream and downstream traffic ratio coefficient data within the range of (0.458, 0.65]; C3 represents a value of the processed UDP traffic proportion data within the range of (0.658, 0.85]; E2 represents a value of the processed IP session count data within the range of (0.354, 0.456]; F2 represents that the destination IP is a home broadband address; Y represents being a PCDN user.

[0072] It should be noted that specific characteristic fields are not specifically limited herein.

[0073] In some embodiments, the first Internet access behavior characteristic data includes the actual recognition result of the first user, and the actual recognition result includes the target recognition result, and the target recognition result characterizes that the user is a PCDN user.

[0074] In some embodiments, the actual recognition result may further include a result characterizing that the user is a non-PCDN user.

[0075] S120: Input the first Internet access behavior characteristic data into the association rule model, generate the first association rule of the first Internet access behavior characteristic data by using the association rule model, and determine the first confidence level of the first association rule by using the relationship information between the preset association rule and the preset confidence level.

[0076] Among them, the association rule model can be an Apriori association rule model.

[0077] In some embodiments, the model mainly includes an input part, an algorithm processing part, and an output part. Among them, the input part includes the input of the first Internet access behavior characteristic data and the input of model parameters. The algorithm processing part adopts the Apriori association rule algorithm. The output part is the association rule. Among them, the model parameters can be set in advance.

[0078] The specific implementation steps of the association rule model are as follows: First, set the support threshold and confidence threshold of the model parameters, input the first Internet access behavior characteristic data, and then use the Apriori association rule algorithm to analyze the first Internet access behavior characteristic data, and output the association rule with the support threshold and confidence threshold set by the model parameters as conditions.

[0079] As an example, under the conditions of the support threshold and the confidence threshold, 8 first association rules are output. Each first association rule can be expressed as f(x)→y. Where f(x) represents the rule composed of various Internet behavior feature data, and Y represents a PCDN user. For example, A2---C3---E2→Y, which represents A2 AND C3 AND E2=>Y. Similarly, A3---B3---Y→F2 represents A3 AND B3 AND Y=>F2.

[0080] It can be understood that by using the association rule model, the relationship between PCDN and each Internet behavior feature data can be obtained. Explore the relationship between PCDN and each Internet behavior feature data, so as to find the combination of a few key Internet behavior feature data for determining PCDN. The association rule model is mainly used to find the association relationship between Internet behavior feature data. Based on the statistical law of Internet behavior feature data, association rule mining can reveal the unknown relationship between features. According to the mined association relationship, when the confidence reaches a certain threshold, it is possible to infer the information of one feature from the information of another feature.

[0081] In some embodiments, the preset association rules include first association rules, and the preset confidence levels include first confidence levels. Using the relationship information between the preset association rules and the preset confidence levels, calculate the first confidence level of each first association rule. The relationship information between the preset association rules and the preset confidence levels is shown in formula (1):

[0082]

[0083] Where support(X∪Y) is the support of the union of X and Y, and support(X) is the support of X. Among them, each first association rule can be expressed as f(x)→y, X is f(x), and Y is y.

[0084] S130. From the first association rules, screen out the second association rules whose first confidence level is greater than the confidence threshold.

[0085] In some embodiments, screen out the second association rules whose first confidence level is greater than or equal to the confidence threshold as the association rules output by the model.

[0086] S140. From the second association rules, screen out the third association rules whose rule results are the target recognition results in the actual recognition results, so as to use the third association rules to identify the target users. The target recognition result represents that the user is a content delivery network PCDN user based on the point-to-point technology, and the target user is a PCDN user.

[0087] In some embodiments, not all association rules output by the model are meaningful. Only the association rules with Y as the rule result need to be selected. Y represents a PCDN user.

[0088] In some embodiments, to simplify the PCDN user identification rules, only the top two association rules with the highest confidence can also be selected. For example, A2---C3---E2→Y and A3---B3---F2→Y.

[0089] Based on the user's Internet behavior characteristic data, through association rule mining, the embodiments of the present application determine the association rules for identifying PCDN users, and use the association rules to identify PCDN users, improving the accuracy of PCDN user identification.

[0090] Based on this, in some embodiments, as Figure 2 shown, the above S120 may specifically include S121 to S126.

[0091] S121. Input the first Internet behavior characteristic data into the association rule model, screen out the second Internet behavior characteristic data with the occurrence times not less than the initial support threshold in the first Internet behavior characteristic data, generate the fourth association rule of the second Internet behavior characteristic data, and determine the second confidence of the fourth association rule by using the relationship information between the preset association rule and the preset confidence.

[0092] Among them, the initial support threshold is set in advance. The relationship information between the preset association rule and the preset confidence is shown in formula (1). The preset association rule includes the fourth association rule, and the preset confidence includes the second confidence. The fourth association rule can all be expressed as f(x)→y, where X is f(x) and Y is y.

[0093] S122. Screen out the fifth association rule with the second confidence greater than the initial confidence threshold from the fourth association rules.

[0094] Among them, the initial confidence threshold is set in advance.

[0095] In some embodiments, screen out the fifth association rule with the second confidence greater than or equal to the initial confidence threshold as the association rule output by the association rule model.

[0096] S123. Screen out the sixth association rule with the target recognition result in the actual recognition result as the rule result from the fifth association rules.

[0097] In some embodiments, not all association rules output by the model are meaningful. Only the association rules with the target recognition result as the rule result need to be selected. The target recognition result indicates that the user is a PCDN user.

[0098] S124. Determine the accuracy of the sixth association rule using the first Internet behavior feature data. When the accuracy does not meet the training stop condition, adjust the initial support threshold and the initial confidence threshold for updating the second Internet behavior feature data, the fourth association rule, the fifth association rule, and the sixth association rule, and update the accuracy using the updated sixth association rule.

[0099] In some embodiments, determine the accuracy of the sixth association rule using the actual recognition result corresponding to the first Internet behavior feature data. When the accuracy does not meet the training stop condition, adjust the initial support threshold and the initial confidence threshold for updating the second Internet behavior feature data, the fourth association rule, the fifth association rule, and the sixth association rule, and update the accuracy using the updated sixth association rule. When the updated accuracy does not meet the training stop condition, readjust the initial support threshold and the initial confidence threshold.

[0100] In some embodiments, the first Internet behavior feature data can be divided into a training set and a test set. Determining the accuracy of the sixth association rule using the first Internet behavior feature data may specifically include: determining the accuracy of the sixth association rule using the test set.

[0101] S125. When the updated accuracy meets the training stop condition, obtain the support threshold and the confidence threshold.

[0102] In some embodiments, the initial support threshold and the initial confidence threshold need to be set using empirical values, and then during the model training process, they are adjusted multiple times to obtain the support threshold and the confidence threshold that conform to the business. For example, the support threshold is 6% and the confidence threshold is 70%.

[0103] S126. Screen out the third Internet behavior feature data in the first Internet behavior feature data whose occurrence times are not less than the support threshold, generate the first association rule of the third Internet behavior feature data, and determine the first confidence of the first association rule using the relationship information between the preset association rule and the preset confidence.

[0104] It can be understood that when performing the test set verification, there are no obvious changes in the support and the confidence, achieving the purpose that the association rule also has generalization ability on the test set.

[0105] In the embodiments of the present application, the accuracy of the association rule determined by the model for identifying the target user is used to adjust the initial support threshold and the initial confidence threshold, so that the finally determined association rule is more accurate, improving the accuracy of PCDN user identification.

[0106] Based on this, in some embodiments, as Figure 3 shown, the above S110 may specifically include S111 to S114.

[0107] S111. Obtain the fourth Internet access behavior characteristic data of the first user.

[0108] Among them, the fourth Internet access behavior characteristic data may include, but is not limited to, uplink bandwidth, downlink bandwidth, total uplink traffic, total downlink traffic, early morning traffic, HyperText Transfer Protocol (HTTP) video traffic, User Datagram Protocol (UDP) traffic, daily Domain Name System (DNS) resolution volume, Internet Protocol (IP) session number, destination IP, number of Media Access Control (MAC) addresses under a Passive Optical Network (PON) port, number of broadband accounts under a PON port, whether ports 80 and 443 are opened, and whether the user is a PCDN user (i.e., the actual identification result of the first user).

[0109] Next, perform data transformation on the fourth Internet access behavior characteristic data. The data transformation mainly uses feature construction and data discretization for processing.

[0110] S112. Based on the fourth Internet access behavior characteristic data, construct the constructed characteristic data of the first user by using the preset feature construction relationship information.

[0111] Among them, the preset feature construction relationship information is set in advance.

[0112] In some embodiments, feature construction is used to obtain more concise and meaningful feature data. Use the existing fourth Internet access behavior characteristic data to construct new constructed characteristic data, further explore and discover the mutual connection between feature data, and add it to the fifth Internet access behavior characteristic data.

[0113] As an example, there are 4 newly constructed characteristic data:

[0114] Early morning traffic ratio = early morning traffic (0:00 - 6:00) / (total uplink traffic + total downlink traffic);

[0115] Uplink / downlink traffic ratio coefficient = total uplink traffic / total downlink traffic;

[0116] UDP traffic ratio = UDP traffic / (total uplink traffic + total downlink traffic);

[0117] Account / MAC ratio coefficient = number of broadband accounts under a PON port / number of MAC addresses under a PON port.

[0118] S113. Combine the target Internet behavior feature data in the combined construction feature data and the fourth Internet behavior feature data to obtain the fifth Internet behavior feature data.

[0119] Among them, the fourth Internet behavior feature data includes the target Internet behavior feature data.

[0120] In some embodiments, the target Internet behavior feature data in the construction feature data and the fourth Internet behavior feature data are combined. For example, after retaining 2 original feature data, namely the number of IP sessions and whether the target IP is a home broadband address, and excluding the remaining feature data, it is combined with the construction feature data to form the fifth Internet behavior feature data.

[0121] It should be noted that the fourth Internet behavior feature data includes the target IP. The target IP can be compared with the home broadband address to generate sub-Internet behavior feature data on whether the target IP is a home broadband address.

[0122] In some embodiments, in order to eliminate the dimensionality difference between feature data and ensure the stability and convergence of model training, after combining the construction feature data and the target Internet behavior feature data in the fourth Internet behavior feature data, it is scaled to a specific range. For example, using Min-Max normalization processing, the data is linearly mapped to the range [0, 1] to normalize the sub-Internet behavior feature data corresponding to each feature field and obtain the fifth Internet behavior feature data. The calculation formula is: the normalized data = (the original data - the minimum value of the data) / (the maximum value of the data - the minimum value of the data).

[0123] S114. Discretize each sub-Internet behavior feature data in the fifth Internet behavior feature data respectively to obtain the first Internet behavior feature data.

[0124] In some embodiments, the clustering algorithm is used to discretize each sub-Internet behavior feature data in each feature field of the fifth Internet behavior feature data to form the first Internet behavior feature data.

[0125] It can be understood that in order to convert the original data into a format suitable for modeling, the data needs to be discretized. For example, the clustering algorithm is used to discretize the sub-Internet behavior feature data corresponding to each feature field, and the sub-Internet behavior feature data corresponding to each feature field is clustered into 4 categories.

[0126] The embodiments of the present application use the existing Internet behavior feature data to construct new feature data, further explore and discover the mutual connections between feature data, combine it with the original feature data, and perform discretization processing after dimensionality reduction, so as to help the model mine deeper pattern knowledge and improve the efficiency and accuracy of data mining.

[0127] Based on this, in some embodiments, such as Figure 4 shown, the above S111 may specifically include S210 to S240.

[0128] S210. Obtain the sixth Internet access behavior characteristic data of the first user.

[0129] Among them, the sixth Internet access behavior characteristic data may include, but is not limited to, account, city, Broadband Remote Access Server (BRAS), interface, Network Address Translation IP (NAT IP), Optical Line Terminal (OLT), Passive Optical Network (PON), Media Access Control (MAC), Service Virtual Local Area Network (SVLAN), Customer Virtual Local Area Network (CVLAN), upstream bandwidth, downstream bandwidth, total upstream traffic, total downstream traffic, early morning traffic, HTTP video traffic, UDP traffic, daily DNS resolution volume, IP session number, destination IP, number of MAC addresses under the PON port, number of broadband accounts under the PON port, whether the 80 and 443 ports are opened, and whether it is a PCDN user.

[0130] In some embodiments, the sixth Internet access behavior characteristic data may be collected from systems such as a network management system, a DNS system, a PON performance network management, a Remote Authentication Dial In User Service (RADIUS) system, and a wired user access platform.

[0131] S220. Display the sixth Internet access behavior characteristic data.

[0132] In some embodiments, in order to reduce the complexity of the data and improve the efficiency of data mining, it is necessary to delete irrelevant or redundant characteristic data to reduce the dimension of the data, thereby simplifying the data mining task. The sixth Internet access behavior characteristic data is displayed through a display interface for the user to screen the characteristic fields of the Internet access behavior characteristic data.

[0133] S230. Receive the selection input for the sixth Internet access behavior characteristic data.

[0134] As an example, the sixth set of Internet behavior characteristic data collected includes a total of 24 characteristic fields. To more effectively mine it, the redundant characteristic fields that are significantly irrelevant to the mining task are manually removed. Fourteen key characteristic fields are selected to obtain the fourth set of Internet behavior characteristic data. At the same time, for the removed characteristic fields, the corresponding sub-Internet behavior characteristic data is deleted.

[0135] S240. In response to a selection input, obtain the fourth set of Internet behavior characteristic data of the first user.

[0136] In some embodiments, in response to a selection input, obtain the fourth set of Internet behavior characteristic data of the first user. The fourth set of Internet behavior characteristic data may include, but is not limited to, uplink bandwidth, downlink bandwidth, total uplink traffic, total downlink traffic, early morning traffic, HTTP video traffic, UDP traffic, daily DNS resolution volume, IP session number, target IP, number of MAC addresses under the PON port, number of broadband accounts under the PON port, whether ports 80 and 443 are opened, and whether the user is a PCDN user (i.e., the actual identification result of the first user).

[0137] The embodiments of the present application reduce the dimension of the data by deleting irrelevant or redundant characteristic data, reduce the complexity of the data, and improve the efficiency of data mining.

[0138] Based on this, in some embodiments, the above S210 may specifically include:

[0139] Obtain the initial Internet behavior characteristic data of the second user;

[0140] Delete the initial Internet behavior characteristic data of the third user with missing values in the initial Internet behavior characteristic data to obtain the sixth set of Internet behavior characteristic data of the first user;

[0141] Wherein, the second user includes the first user and the third user.

[0142] In some embodiments, use the user accounts that have been verified and determined to be PCDN users as the objects of data mining. Match these user accounts on systems such as the network management system, DNS system, PON performance network management, RADIUS system, and wired user access platform, and associate the Internet behavior characteristic data of these user accounts. Combine and organize the collected Internet behavior characteristic data into initial Internet behavior characteristic data.

[0143] As an example, the Internet usage behavior feature data obtained from each system includes: account, city, BRAS, interface, NAT IP, OLT, PON, MAC, SVLAN, CVLAN, upstream bandwidth, downstream bandwidth, total upstream traffic, total downstream traffic, early morning traffic, HTTP video traffic, UDP traffic, daily DNS resolution volume, IP session count, destination IP, number of MAC addresses under the PON port, number of broadband accounts under the PON port, whether ports 80 and 443 are opened, and whether it is a PCDN user.

[0144] In some embodiments, data cleaning is performed on the initial Internet usage behavior feature data. To facilitate model analysis, it is necessary to process missing values, duplicate values, and error values in the initial Internet usage behavior feature data. The processing methods are as follows:

[0145] For the missing values of numerical fields such as total upstream traffic, total downstream traffic, HTTP video traffic, and UDP traffic, the median is used for filling;

[0146] For fields lacking data such as BRAS, interface, NAT IP, OLT, PON, MAC, SVLAN, and CVLAN, direct deletion operations are performed.

[0147] For outliers, the method of mean replacement is used for processing.

[0148] In the embodiment of the present application, by deleting the initial Internet usage behavior feature data of the third user with missing values in the initial Internet usage behavior feature data, data cleaning is performed, avoiding the interference of abnormal data and improving the accuracy of association rule determination.

[0149] Based on this, in some embodiments, the above S240 may specifically include:

[0150] In response to the selection input, obtain the seventh Internet usage behavior feature data of the first user;

[0151] Calculate the correlation coefficient between any two Internet usage behavior feature data in the seventh Internet usage behavior feature data;

[0152] When the correlation coefficient is greater than the preset threshold, delete one of any two Internet usage behavior feature data to obtain the fourth Internet usage behavior feature data of the first user.

[0153] Among them, the preset threshold is set in advance.

[0154] In some embodiments, data dimensionality reduction can also be performed by means of Pearson correlation coefficient, Principal Component Analysis (PCA), etc. Calculate the correlation coefficient between any two feature fields in the seventh Internet behavior feature data. If the correlation coefficient is greater than a preset threshold, delete one of the any two feature fields to obtain the fourth Internet behavior feature data of the first user.

[0155] In the embodiments of the present application, the correlation coefficient of the Internet behavior feature data is used to perform dimensionality reduction on the data again, further reducing the complexity of the data and improving the efficiency of data mining.

[0156] In the embodiments provided in the present application, for the target users screened out by using the third association rule, that is, users with the uplink bandwidth utilization rate exceeding a preset value (such as 70%), restrict their uplink bandwidth. For users whose uplink bandwidth utilization rate does not exceed the preset value, display the user accounts for further on-site verification, improving the accuracy of PCDN user identification and also being able to improve user satisfaction.

[0157] In the embodiments of the present application, by mining to find out the key combinations of Internet behavior feature data, PCDN user identification is realized, with simple operation and high accuracy. That is, find out which feature data has a strong correlation with PCDN, so that only through the combination discrimination of these small amounts of key feature data, PCDN users can be accurately identified. It solves the deficiencies of the traditional solutions, including subjective screening of features and the need for a large amount of manual data analysis. In the embodiments of the present application, through the combination of a small amount of key feature data, it is possible to simply, efficiently and accurately judge PCDN users, thereby improving the PCDN identification efficiency.

[0158] The main ideas based on data mining include applying feature engineering, modeling the data, etc., as follows:

[0159] 1. Take the user accounts whose status as PCDN users has been verified and determined as the objects of data mining, and randomly split these user accounts into two parts, 70% as training data and 30% as test data;

[0160] 2. Query the Internet behavior characteristic data of these user accounts on systems such as the network management system, DNS system, PON performance network management, RADIUS system, and wired user access platform, including city, BRAS, interface, NAT IP, OLT, PON, MAC, SVLAN, CVLAN, upstream bandwidth, downstream bandwidth, total upstream traffic, early morning traffic, total downstream traffic, HTTP video traffic, UDP traffic, daily DNS resolution volume, IP session number, destination IP, number of MAC addresses under the PON port, number of broadband accounts under the PON port, whether ports 80 and 443 are opened, etc.;

[0161] 3. Statistically record the account and its related characteristic data in a form as the original data for subsequent data mining;

[0162] 4. Preprocess the above original data, including operations such as data cleaning, feature reduction, data transformation, as well as feature construction and data discretization, and finally form a modeling data set;

[0163] 5. Model construction. The goal of model construction is to explore the relationship between PCDN users and these characteristic data, and to mine out a few combinations of characteristic data with the highest correlation with PCDN users;

[0164] 6. Model testing. Use the test set to verify the determined model to verify whether this combination of a few characteristic data can directly identify PCDN users and requires a relatively high accuracy;

[0165] 7. Model application. After building the model, perform model screening on all accounts and their related characteristic data in the live network to identify all suspected PCDN users and accurately take measures such as speed limiting and shutdown.

[0166] The process of determining the final association rules in the embodiments of this application includes feature data construction, obtaining multiple association rules representing the relationship between PCDN users and each characteristic data, and determining the final association rules. The biggest feature of the embodiments of this application is that it finds a simple and efficient method for identifying PCDN users, and it is obtained based on data mining statistics, with a relatively high confidence level, high reliability, and small misjudgment rate. Specifically, the embodiments of this application find a feature combination that is fast, simple, and effective for identifying PCDN users. To find this feature combination, the embodiments of this application use data mining technology, apply feature engineering, and use association algorithms. Thus, it is possible to find the feature combination with the greatest correlation (i.e., the highest confidence level) with PCDN, and use this feature combination to judge PCDN violation services, thereby simplifying the complexity of traditional solutions, using the simplest few association features to identify PCDN users with the highest probability, greatly simplifying the workload while improving the identification accuracy, and enhancing the discovery and supervision capabilities of broadband user violation operations.

[0167] Based on the method for identifying a target user provided in the above embodiments, correspondingly, the present application also provides a specific implementation manner of an apparatus for identifying a target user. Please refer to the following embodiments.

[0168] Refer to Figure 5 , the apparatus 300 for identifying a target user provided in the embodiments of the present application includes:

[0169] An obtaining module 310, configured to obtain first Internet access behavior feature data of a first user, where the first Internet access behavior feature data includes an actual identification result of the first user;

[0170] A determining module 320, configured to input the first Internet access behavior feature data into an association rule model, generate a first association rule of the first Internet access behavior feature data by using the association rule model, and determine a first confidence level of the first association rule by using relationship information between a preset association rule and a preset confidence level;

[0171] A screening module 330, configured to screen out a second association rule from the first association rules, where the second association rule has a first confidence level greater than a confidence level threshold;

[0172] The screening module 330 is further configured to screen out a third association rule from the second association rules, where the third association rule uses a target identification result in the actual identification result as a rule result, so as to identify a target user by using the third association rule. The target identification result indicates that the user is a PCDN user, and the target user is a PCDN user.

[0173] Based on this, in some embodiments, the determining module 320 may specifically be configured to:

[0174] Input the first Internet access behavior feature data into an association rule model, screen out second Internet access behavior feature data in the first Internet access behavior feature data that appears no less than an initial support threshold, generate a fourth association rule of the second Internet access behavior feature data, and determine a second confidence level of the fourth association rule by using relationship information between a preset association rule and a preset confidence level;

[0175] Screen out a fifth association rule from the fourth association rules, where the second confidence level of the fifth association rule is greater than an initial confidence level threshold;

[0176] Screen out a sixth association rule from the fifth association rules, where the sixth association rule uses a target identification result in the actual identification result as a rule result;

[0177] Determine the accuracy of the sixth association rule by using the first Internet access behavior feature data. If the accuracy does not meet the training stop condition, adjust the initial support threshold and the initial confidence level threshold to update the second Internet access behavior feature data, the fourth association rule, the fifth association rule, and the sixth association rule, and update the accuracy by using the updated sixth association rule;

[0178] When the updated accuracy meets the training stop condition, a support threshold and a confidence threshold are obtained;

[0179] Filter out the third Internet behavior feature data whose occurrence times in the first Internet behavior feature data are not less than the support threshold, generate the first association rule of the third Internet behavior feature data, and determine the first confidence of the first association rule by using the relationship information between the preset association rule and the preset confidence.

[0180] Based on this, in some embodiments, the obtaining module 310 may specifically be used for:

[0181] Obtain the fourth Internet behavior feature data of the first user;

[0182] Based on the fourth Internet behavior feature data, construct the constructed feature data of the first user by using the preset feature construction relationship information;

[0183] Combine the target Internet behavior feature data in the constructed feature data and the fourth Internet behavior feature data to obtain the fifth Internet behavior feature data;

[0184] Discretize each sub-Internet behavior feature data in the fifth Internet behavior feature data respectively to obtain the first Internet behavior feature data.

[0185] Based on this, in some embodiments, the obtaining module 310 may specifically be used for:

[0186] Obtain the sixth Internet behavior feature data of the first user;

[0187] Display the sixth Internet behavior feature data;

[0188] Receive the selection input for the sixth Internet behavior feature data;

[0189] In response to the selection input, obtain the fourth Internet behavior feature data of the first user.

[0190] Based on this, in some embodiments, the obtaining module 310 may specifically be used for:

[0191] Obtain the initial Internet behavior feature data of the second user;

[0192] Delete the initial Internet behavior feature data of the third user with missing values in the initial Internet behavior feature data to obtain the sixth Internet behavior feature data of the first user;

[0193] Wherein, the second user includes the first user and the third user.

[0194] Based on this, in some embodiments, the obtaining module 310 may specifically be used for:

[0195] In response to the selection input, obtain the seventh Internet behavior characteristic data of the first user;

[0196] Calculate the correlation coefficient between any two Internet behavior characteristic data in the seventh Internet behavior characteristic data;

[0197] In the case where the correlation coefficient is greater than the preset threshold, delete one of any two Internet behavior characteristic data to obtain the fourth Internet behavior characteristic data of the first user.

[0198] Each module of the target user identification device provided in the embodiments of the present application can implement the functions of each step of the above-provided target user identification method and achieve its corresponding technical effects. For the sake of brevity, it will not be elaborated here.

[0199] Based on the same inventive concept, the embodiments of the present application also provide an electronic device.

[0200] Figure 6 The figure shows a schematic hardware structure diagram of the electronic device provided in the embodiments of the present application.

[0201] The electronic device may include a processor 401 and a memory 402 storing computer program instructions.

[0202] Specifically, the above-mentioned processor 401 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0203] The memory 402 may include a mass storage for data or instructions. By way of example and not limitation, the memory 402 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disc, a magneto-optical disc, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. In a suitable case, the memory 402 may include a removable or non-removable (or fixed) medium. In a suitable case, the memory 402 may be internal or external to the integrated gateway disaster recovery device. In a specific embodiment, the memory 402 is a non-volatile solid-state memory.

[0204] The memory may include a read only memory (ROM), a random access memory (RAM), a magnetic disk storage media device, an optical storage media device, a flash memory device, an electrical, optical, or other physical / tangible memory storage device. Thus, in general, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to one aspect of the present disclosure.

[0205] The processor 401 reads and executes the computer program instructions stored in the memory 402 to implement the identification method of any target user in the above embodiments.

[0206] In one example, the electronic device may further include a communication interface 403 and a bus 410. Among them, as Figure 6 shown, the processor 401, the memory 402, and the communication interface 403 are connected through the bus 410 and complete communication with each other.

[0207] The communication interface 403 is mainly used to implement communication between the modules, devices, units, and / or devices in the embodiments of the present application.

[0208] Bus 410 includes hardware, software, or both, and couples components of an electronic device to each other. By way of example and not limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (Peripheral Component Interconnect-X, PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VESA Local Bus, VLB) bus, or other suitable bus or a combination of two or more of these. Where appropriate, bus 410 may include one or more buses. Although embodiments of the present application describe and illustrate specific buses, the present application contemplates any suitable bus or interconnect. The electronic device can execute the method for identifying a target user in the embodiments of the present invention, thereby implementing the above-mentioned method for identifying a target user.

[0209] In addition, in combination with the method for identifying a target user in the above embodiments, an embodiment of the present application can be implemented by providing a computer storage medium. Computer program instructions are stored on the computer storage medium; when the computer program instructions are executed by a processor, any one of the methods for identifying a target user in the above embodiments is implemented.

[0210] The present application also provides a computer program product. When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device is caused to execute each process of implementing any one of the above embodiments of the method for identifying a target user.

[0211] It should be clear that the present application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order between steps after understanding the spirit of the present application.

[0212] The functional blocks shown in the above-described block diagrams can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a functional card, and so on. When implemented in software, the elements of the present application are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium or transmitted via a data signal carried in a carrier wave on a transmission medium or a communication link. A "machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, read-only memory (ROM), flash memory, erasable read-only memory (EROM), floppy disks, compact disc read-only memory (CD-ROM), optical discs, hard disks, fiber optic media, radio frequency (RF) links, and so on. The code segment can be downloaded via a computer network such as the Internet, an intranet, and so on.

[0213] It should also be noted that the exemplary embodiments mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above steps, that is, the steps can be executed in the order mentioned in the embodiments, or different from the order in the embodiments, or several steps can be executed simultaneously.

[0214] Aspects of the present disclosure have been described above with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each block in the flowchart and / or block diagram, and the combinations of blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing apparatus enable the implementation of the functions / operations specified in one or more blocks of the flowchart and / or block diagram. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field programmable logic circuit. It should also be understood that each block in the block diagram and / or flowchart, and the combinations of blocks in the block diagram and / or flowchart, can also be implemented by dedicated hardware performing the specified functions or operations, or by a combination of dedicated hardware and computer instructions.

[0215] The above is only the specific implementation manner of the present application. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, modules, and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be elaborated herein. It should be understood that the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of various equivalent modifications or substitutions within the technical scope disclosed in the present application, and these modifications or substitutions should all be covered by the protection scope of the present application.

Claims

1. A method for identifying a target user, characterized in that: include: Acquire first online behavior characteristic data of a first user, where the first online behavior characteristic data includes an actual identification result of the first user; Inputting the first online behavior characteristic data into an association rule model, generating a first association rule of the first online behavior characteristic data using the association rule model, and determining a first confidence of the first association rule using relationship information between a preset association rule and a preset confidence; Filtering out, from the first association rules, a second association rule whose first confidence is greater than a confidence threshold; A third association rule is screened out from the second association rule, with the target identification result in the actual identification result as the rule result, so as to identify the target user using the third association rule, wherein the target identification result characterizes that the user is a PCDN user of a content distribution network based on point-to-point technology, and the target user is a PCDN user.

2. The target user identification method according to claim 1, characterized in that: The step of inputting the first online behavior characteristic data into an association rule model, generating a first association rule for the first online behavior characteristic data by using the association rule model, and determining a first confidence of the first association rule by using relationship information between a preset association rule and a preset confidence level includes: Inputting the first online behavior characteristic data into an association rule model, screening out second online behavior characteristic data whose occurrence times in the first online behavior characteristic data is not less than an initial support threshold, generating a fourth association rule for the second online behavior characteristic data, and determining a second confidence of the fourth association rule by using relationship information between a preset association rule and a preset confidence; Filtering out, from the fourth association rules, a fifth association rule whose second confidence is greater than an initial confidence threshold; Filtering out a sixth association rule having a target recognition result in the actual recognition result as a rule result from the fifth association rule; Determine the accuracy of the sixth association rule by using the first Internet behavior feature data, and when the accuracy does not meet the training stop condition, adjust the initial support threshold and the initial confidence threshold to update the second Internet behavior feature data, the fourth association rule, the fifth association rule and the sixth association rule, and update the accuracy by using the updated sixth association rule; When the updated accuracy satisfies the training stop condition, obtaining the support threshold and the confidence threshold; The third online behavior characteristic data whose occurrence times are not less than the support threshold are screened out from the first online behavior characteristic data, a first association rule for the third online behavior characteristic data is generated, and a first confidence of the first association rule is determined by using the relationship information of the preset association rule and the preset confidence.

3. The target user identification method according to claim 1 or 2, characterized in that: The obtaining of the first online behavior characteristic data of the first user includes: Acquire fourth online behavior characteristic data of the first user; Based on the fourth online behavior characteristic data, constructing the structure characteristic data of the first user using preset characteristic construction relationship information; Combining the construction characteristic data with the target online behavior characteristic data in the fourth online behavior characteristic data to obtain fifth online behavior characteristic data; Discretization processing is performed on each sub-surfing behavior characteristic data in the fifth surfing behavior characteristic data to obtain first surfing behavior characteristic data.

4. The target user identification method according to claim 3, characterized in that: The obtaining of fourth online behavior characteristic data of the first user includes: Acquire sixth online behavior characteristic data of the first user; Displaying the sixth online behavior characteristic data; Receiving a selection input of the sixth online behavior characteristic data; In response to the selection input, fourth online behavior characteristic data of the first user is obtained.

5. The target user identification method according to claim 4, characterized in that: The obtaining of the sixth online behavior characteristic data of the first user includes: Acquire initial online behavior characteristic data of the second user; Deleting the initial online behavior characteristic data of the third user having missing values ​​in the initial online behavior characteristic data, to obtain the sixth online behavior characteristic data of the first user; The second user includes the first user and the third user.

6. The target user identification method according to claim 4, characterized in that: The step of obtaining fourth online behavior characteristic data of the first user in response to the selection input includes: In response to the selection input, obtaining seventh online behavior characteristic data of the first user; Calculating a correlation coefficient between any two Internet surfing behavior characteristic data in the seventh Internet surfing behavior characteristic data; When the correlation coefficient is greater than a preset threshold, one of the two online behavior characteristic data is deleted to obtain fourth online behavior characteristic data of the first user.

7. A target user identification device, characterized in that: include: An acquisition module, configured to acquire first online behavior characteristic data of a first user, wherein the first online behavior characteristic data includes an actual identification result of the first user; a determination module, configured to input the first online behavior characteristic data into an association rule model, generate a first association rule of the first online behavior characteristic data using the association rule model, and determine a first confidence of the first association rule using relationship information between a preset association rule and a preset confidence; A screening module, configured to screen out, from the first association rules, second association rules whose first confidence is greater than a confidence threshold; The screening module is further used to screen out a third association rule with a target identification result in the actual identification result as a rule result from the second association rule, so as to identify a target user using the third association rule, wherein the target identification result indicates that the user is a PCDN user, and the target user is a PCDN user.

8. An electronic device, characterized in that: The device comprises: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the target user identification method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the method for identifying a target user according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device is enabled to execute the target user identification method as described in any one of claims 1 to 6.

Images