CN114880522A - Method and device for realizing ID Mapping based on graph database - Google Patents

Abstract

The present invention provides a method and device for implementing ID Mapping based on a graph database. The method includes: obtaining the relationship between an ID node appearing on the T-th day and an ID node appearing on the T-day from a source ID data record; The ID node of the ID node, the ID node relationship that occurs on the T day, and the ID relationship network corresponding to the T-1 day are identified and connected, and the first ID relationship network corresponding to the T day is obtained; According to the activity of the ID node in the first ID relationship network The degree and the activity of the ID node relationship are used to clean up the first ID relationship network, and obtain the second ID relationship network corresponding to the T-th day. The present invention cleans up expired IDs by cleaning the ID nodes whose activity is lower than the threshold, and realizes the cleaning of the weak association of ID nodes by disconnecting the ID node relationship whose activity is lower than the threshold, thereby improving the reliability of the user ID relationship network. Accuracy and stability.

Description

Method and device for implementing ID Mapping based on graph database

technical field

The invention relates to the technical field of big data, and in particular, to a method and device for implementing ID Mapping based on a graph database.

Background technique

In real life, users can obtain services provided by enterprises through a variety of devices and from various portals; enterprises can also develop a variety of business lines, form a variety of products, and can provide users from different channels. Provide services, so the data of the same user comes from different data sources, and the data is diverse and scattered in various locations.

The big data platform physically solves the problem of "data islands" in which data is scattered in various locations, but logically, it is difficult to establish associations between data from different sources, and the data is still fragmented. Constructing a one-sided portrait of a user with a single or very little data is equivalent to "blind man touching an elephant", and it is difficult to provide complete information of a user.

A common feature of data from different sources is that they come from the same user, and usually the data records information identifying the user's identity, which is collectively referred to as an "identity document (ID)" here. A user ID is a series of serial numbers representing a user entity, such as ID number, mobile phone number, email, WeChat ID, device number, Cookie ID, and Media Access Control (MAC) address. Therefore, establishing the relationship between the user IDs can establish a connection between the data. Among them, the process of building the relationship between user IDs is the main process of ID Mapping.

Generally speaking, ID Mapping is to use various technical means to identify user IDs in multiple data from different sources as the same subject and generate a unified identity identifier (one-ID) that identifies the unique identity of the user. According to the data structure used in the processing process, the ID Mapping implementation methods can be roughly summarized into three categories: dictionary method, table method and graph method, but the current specific implementation methods in these three methods mainly focus on building ID relationships and do not pay attention to them. The problem of ID expiration and reuse and the solution of the complex relationship between IDs seriously affect the reliability, accuracy and stability of the constructed user ID relationship network. Among them, the problem of ID expiration and reuse is caused by the different life cycle and precision of user IDs. For example, a person usually has only one ID number in his life, while the mobile phone number, email address and device number will change frequently; the complex relationship between IDs The problem is caused by complex real-world scenarios, such as multiple accounts on the same device, multiple devices with the same account, multiple accounts, multiple data sources, and abnormal data.

SUMMARY OF THE INVENTION

The invention provides a method and device for realizing ID Mapping based on a graph database, which is used to solve the defects of low reliability, low accuracy and low stability of the user ID relation network in the prior art, and realizes the detection of ID expiration, ID reuse and effective processing of complex ID relationships improve the reliability, accuracy and stability of the user ID relationship network.

The present invention provides a method for implementing ID Mapping based on a graph database, comprising:

From the source ID data record, obtain the relationship between the ID node that occurs on the T-th day and the ID node that occurs on the T-th day;

The ID node that occurs on the T day, the ID node relationship that occurs on the T day, and the ID relationship network corresponding to the T-1 day are identified and connected, and the first ID relationship network corresponding to the T day is obtained;

The first ID relationship network is cleaned up according to the activity degree of the ID nodes in the first ID relationship network and the activity degree of the ID node relationship, and the second ID relationship network corresponding to the T-th day is obtained.

Optionally, before cleaning the first ID relationship network according to the activity of the ID nodes and the activity of the ID node relationship in the first ID relationship network, the method further includes:

respectively update the attribute of the ID node and the attribute of the ID node relationship;

Perform feature extraction on the updated attribute of the ID node and the updated attribute of the ID node relationship, respectively, to obtain the feature value of the ID node and the feature value of the ID node relationship;

Obtain the activity of the ID node according to the feature value of the ID node and the weight corresponding to the feature value of the ID node;

The activity of the ID node relationship is acquired according to the feature value of the ID node relationship and the weight corresponding to the feature value of the ID node relationship.

Optionally, clean up the first ID relationship network according to the activity of the ID nodes in the first ID relationship network and the activity degree of the ID node relationship, and obtain the second ID relationship network corresponding to the T-th day, including:

In the case that the activity of the ID node in the first ID relationship network is less than the node activity threshold, clearing the ID node out of the first ID relationship network;

In the case that the activity of the ID node relationship in the first ID relationship network is less than the relationship activity threshold, clearing the ID node relationship out of the first ID relationship network;

According to the cleaned first ID relationship network, the second ID relationship network corresponding to the T-th day is acquired.

Optionally, obtaining the second ID relationship network corresponding to the T-th day according to the cleaned up first ID relationship network, including:

In the case that the clearing of the ID node or the relationship between the ID nodes does not result in the splitting of the relationship subnet in the first ID relationship network, the unified identity of the relationship subnet in the second ID relationship network is the Describe the unified identity of the relationship subnet in the first ID relationship network;

In the case where the cleanup of the ID node or the relationship of the ID node causes the relationship subnet in the first ID relationship network to be split into multiple relationship subnets, the multiple relationships in the second ID relationship network The unified identity of a relational subnet in the subnet is the unified identity of the relational subnet in the first ID relational network, and the unified identity of other relational subnets in the plurality of relational subnets is newly generated. unified identity.

Optionally, the first ID relationship network is cleaned up according to the activity of the ID nodes in the first ID relationship network and the activity of the ID node relationship, and after obtaining the second ID relationship network corresponding to the T-th day. ,Also includes:

Obtain the relationship between inactive ID nodes and inactive ID nodes within a preset time from the source ID data record;

updating the activity of the inactive ID node and the activity of the inactive ID node relationship;

The second ID relationship network is cleaned up according to the updated activity degree of the inactive ID node and the updated activity degree of the inactive ID node relationship, and the third ID relationship network corresponding to the T-th day is obtained.

Optionally, the ID node that appears on the T day, the ID node relationship that appears on the T day, and the ID relationship network corresponding to the T-1 day are identified and connected, and the first corresponding to the T day is obtained. Before the ID network, it also includes:

According to the ID node that appears on the T day and the identity mapping dictionary corresponding to the T-1 day, obtain the unified identity that exists on the T-1 day;

Obtain the ID relationship network corresponding to the T-1th day according to the unified identity identifier that exists on the T-1th day.

Optionally, the ID node that appears on the T day, the ID node relationship that appears on the T day, and the ID relationship network corresponding to the T-1 day are identified and connected, and the first corresponding to the T day is obtained. After the ID relationship network, it also includes:

In the case that the relationship subnet in the first ID relationship network has a unified identity, the unified identity of the relationship subnet is the existing unified identity;

In the case that the relationship subnet in the first ID relationship network has more than one unified identity, the unified identity of the relationship subnet is the unified identity with the earliest creation time and the most times of merging or splitting;

In the case where the relationship subnet in the first ID relationship network does not have a unified identity, the unified identity of the relationship subnet is a newly generated unified identity.

The present invention also provides a device for implementing ID Mapping based on a graph database, including:

The first acquisition module is used to obtain the ID node relationship that occurs on the T day and the ID node relationship that occurs on the T day from the source ID data record;

The second acquisition module is used to identify and connect the ID nodes that appear on the T-th day, the ID node relationships that appear on the T-th day, and the ID relationship network corresponding to the T-1 day, and obtain the ID node corresponding to the T-th day. an ID network;

The third acquisition module is used to clean up the first ID relationship network according to the activity of the ID nodes in the first ID relationship network and the activity degree of the ID node relationship, and obtain the second ID relationship network corresponding to the T day .

The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, when the processor executes the program, the processor implements the above-mentioned based on any one of the above A method for implementing ID Mapping in a graph database.

The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements the method for implementing ID Mapping based on a graph database according to any one of the above.

The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the method for implementing ID Mapping based on a graph database according to any one of the above.

The method and device for implementing ID Mapping based on a graph database provided by the present invention realizes the clearing of expired IDs by clearing the ID nodes whose activity is lower than the threshold, and solves the problem of ID expiration. By disconnecting the ID nodes whose activity is lower than the threshold It realizes the cleaning of weak associations between ID nodes, and solves the problems of ID reuse and complex ID relationships, thereby improving the reliability, accuracy and stability of the user ID relationship network.

Description of drawings

In order to illustrate the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings required in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are the For some embodiments of the invention, for those of ordinary skill in the art, other drawings can also be obtained according to these drawings without any creative effort.

1 is one of the schematic flow charts of the method for implementing ID Mapping based on a graph database provided by the present invention;

2 is the second schematic flowchart of the method for implementing ID Mapping based on a graph database provided by the present invention;

3 is a schematic structural diagram of a device for implementing ID Mapping based on a graph database provided by the present invention;

FIG. 4 is a schematic structural diagram of an electronic device provided by the present invention.

Detailed ways

The dictionary method of ID Mapping is the simplest among the three types of methods. In the dictionary method of ID Mapping, the ID in the source data is represented by the keyword (key), and the generated unified one-ID is represented by the value (value). The process is to determine whether the extracted ID is in the key. If the extracted ID exists in the key, the existing one-ID is used; if the extracted ID does not exist in the key, a new one-ID is created. The ID connection may be generated as the ID increases, so it is also necessary to merge the dictionary, that is, merge the IDs that already have an ID relationship but have not established an inter-ID connection.

The advantage of the dictionary method of ID Mapping is that it is simple in principle, easy to implement, and has high processing efficiency; the disadvantage is that it cannot solve the problem of ID expiration and reuse and the complex relationship between IDs. For example, when there are multiple users on the same device, multiple users will be merged into one user. .

The most common way of organizing IDs is in the form of table records. One record contains multiple IDs that appear at the same time, and also contains ID relationships. Therefore, the table method is the most proposed method among the three types of methods. Its brief process: taking records as a unit, merging records with the same ID to form a unified one-ID, and the key point is to reduce the error rate of merged records.

The advantage of the ID Mapping table method is that the principle and implementation are relatively simple, and the error rate can be reduced when merging; the disadvantage is that the processing speed is slow when the amount of data is large, and it is difficult to independently process a single ID and ID relationship. Multiplexing problems and complex ID relationship problems, such as problems caused by mobile phone number transfer.

The essence of the ID relationship is the network structure. Therefore, the graph method is the most direct among the three methods, and with the maturity of the graph database, the graph method is more widely used. It mainly builds the ID in the record into a graph, and usually sets the edge threshold in the process of constructing the graph to filter out the weak ID relationship; Each subgraph generates a unique one-ID, and one of the subgraphs represents a user ID relationship network; this method sets the user behavior rules (for example, sets the threshold number of a certain type of ID that a user can have within a preset time) and ID priority (for example, setting the ID number with the highest priority) deals with the complex relationship between IDs.

The advantage of the graph method of ID Mapping is that the graph structure represents the ID relationship, which is intuitive and easy to understand, and it is easy to handle a single ID and ID relationship; its disadvantage is that it lacks an effective method to deal with the problem of ID expiration and reuse. , but too simplistic.

In order to effectively deal with the ID expiration and reuse problems and the ID complex relationship problem, and improve the reliability, accuracy and stability of the user ID and ID relationship network, the present invention realizes the realization of Cleaning up expired IDs solves the problem of ID expiration. By disconnecting the ID node relationship whose activity is lower than the threshold, the weak association relationship between ID nodes is cleaned up, and the problem of ID reuse and ID complex relationship is solved, thereby improving the The reliability, accuracy and stability of user ID and ID relationship network are improved.

In order to make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention. , not all examples. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

1 is one of the schematic flow charts of the method for implementing ID Mapping based on a graph database provided by the present invention. As shown in FIG. 1 , the present invention provides a method for implementing ID Mapping based on a graph database. The method includes:

Step 101: Acquire the relationship between the ID nodes that appear on the T-th day and the ID nodes that appear on the T-th day from the source ID data record.

Specifically, FIG. 2 is the second schematic flowchart of the method for implementing ID Mapping based on a graph database provided by the present invention. The source data includes various ID data records, and the relationship between the ID and the ID in each ID data record is written into The graph database, and initializes the properties of the ID node and the properties of the ID node relationship. Table 1 is the attribute table of the ID node, and Table 2 is the attribute table of the ID node relationship.

Table 1. Attribute table of ID node

As can be seen from Table 1, the attributes of the ID node include the node value, node type, the first record date of the node, the date when the node was most recently active, the number of days the node was active, the list of days from the first record of each node occurrence, the node Degree value, node activity, where the node priority and node activity threshold are closely related to the node type, two nodes with the same ID of the node type have the same node priority, and two nodes with the same ID of the node type have the same node Activeness threshold.

Table 2 Attribute table of ID node relationship

As can be seen from Table 2, the attributes of the ID node relationship include the relationship description, the relationship type, the first record date of the relationship node, the date when the relationship was most recently active, the number of days the relationship was active, and the number of days from the first record between each occurrence of the relationship. List, relationship liveness, relationship priority, and relationship liveness threshold, where relationship priority and relationship liveness threshold are closely related to the relationship type, two ID node relationships with the same relationship type have the same relationship priority, and two relationships ID node relationships of the same type have the same relationship liveness threshold.

During initialization, except for the node priority, relationship priority, and node degree value, the attributes of other ID nodes and the attributes of the ID node relationship are all set to 1.

After writing the ID and the relationship between the IDs contained in the source data into the graph database, obtain the relationship between the ID node that appears on the T day and the ID node that appears on the T day from the graph database.

For example, the ID nodes that appear on the T day have the device number dev_001, the mobile phone number 150***, the ID number 430***, the account number ccc, and the ID number 560***. The ID node relationship that appears on the T day includes the relationship between the device number dev_001 and the mobile phone number 150***, and the relationship between the ID number 430*** and the account ccc.

Step 102, identify and connect the ID nodes that appear on the T day, the ID node relationships that appear on the T day, and the ID relationship network corresponding to the T-1 day, and obtain the first ID relationship network corresponding to the T day. .

Specifically, before the identification connection is performed to obtain the first ID relationship network corresponding to the T-th day, the ID relationship network corresponding to the T-1th day needs to be obtained first.

Optionally, identify and connect the ID node that occurs on the T day, the ID node relationship that occurs on the T day, and the ID relationship network corresponding to the T-1 day, and before acquiring the first ID relationship network corresponding to the T day, also include:

Obtain the unified identity that exists on the T-1 day according to the ID node that appears on the T day and the identity mapping dictionary corresponding to the T-1 day;

Obtain the ID relationship network corresponding to the T-1 day according to the unified identity that exists on the T-1 day.

Specifically, the unified identity is one-ID, and the unified identity is the Chinese translation of one-ID. The identification mapping dictionary contains one-ID, the ID node corresponding to each one-ID, and the ID node relationship.

Compare the ID node that appears on the T-th day with the ID node included in the ID mapping dictionary corresponding to the T-1 day, and find the ID node that appears on the T-th day and also contains the ID corresponding to the T-1 day. For the ID node in the mapping dictionary, according to the corresponding relationship in the ID mapping dictionary corresponding to the found ID node on the T-1 day, the one-ID corresponding to the found ID node is obtained, that is, the one that exists on the T-1 day. -ID. After obtaining the one-ID existing on the T-1 day, the ID node and the ID node relationship corresponding to the one-ID existing on the T-1 day are searched from the graph database, so as to obtain the ID relationship corresponding to the T-1 day. network.

For example, the node mobile phone number 150*** that appeared on the T day also exists in the identity mapping dictionary corresponding to the T-1 day. In the identity mapping dictionary corresponding to the T-1 day, the mobile phone number 150*** corresponds to one-ID is one-ID01, and then according to one-ID01, find the ID node and ID node relationship that has a corresponding relationship with it in the graph database, and find the ID number 320***, account aaa, mobile phone associated with one-ID01 Number 150***, as well as the relationship between ID number 320*** and mobile phone number 150***, and the relationship between ID number 320*** and account aaa.

By first using the ID node that appears on the T day and the identity mapping dictionary corresponding to the T-1 day, the one-ID existing on the T-1 day is obtained, and then the one-ID existing on the T-1 day is searched according to the graph database. The relationship between the ID node and the ID node corresponding to the one-ID is obtained, and the ID relationship network corresponding to the T-1th day is obtained. In order to use the ID relationship network corresponding to the T-1th day for subsequent identification and connection, the corresponding ID relationship network of the T-th day is obtained. An ID relationship network laid the foundation.

After obtaining the first ID relationship network corresponding to the T day, the maximum connected subgraph algorithm is used to identify the ID nodes that appear on the T day, the ID node relationships that appear on the T day, and the ID relationship network corresponding to the T-1 day. Connectivity, that is, connecting the ID node and the ID node relationship appearing on the T day to the ID relation network corresponding to the T-1 day, thereby obtaining the first ID relation network corresponding to the T day.

After the first ID relationship network is acquired, the one-ID of the relationship subnet in the first ID relationship network needs to be determined.

Optionally, identify and connect the ID node that occurs on the T day, the ID node relationship that occurs on the T day, and the ID relationship network corresponding to the T-1 day, and after obtaining the first ID relationship network corresponding to the T day, also include:

In the case that the relationship subnet in the first ID relationship network has a unified identity, the unified identity of the relationship subnet is the existing unified identity;

In the case that more than one unified identity identifier exists in the relational subnet in the first ID relational network, the unified identity of the relational subnet is the one with the earliest creation time and the most times of merging or splitting;

In the case where the relationship subnet in the first ID relationship network does not have a unified identity, the unified identity of the relationship subnet is a newly generated unified identity.

Specifically, Table 3 is an attribute table of one-ID, and the attributes of one-ID include node value, node generation time, merge or split time, and merge or split times. The relationship attribute between the ID node and the one-ID has the relationship establishment time, and the relationship establishment time refers to the time when the ID node and the one-ID establish the relationship.

Table 3 attribute table of one-ID

Numbering Attributes describe 1 node value Unique identification value for the relational subnet 2 Node generation time Time to create one-ID node 3 Merge or split time Time field updated when merging or splitting relational subnets 4 Number of merges or splits Number of merges and splits

In the case that there is one one-ID in the relational subnet in the first ID relational network, the one-ID of the relational subnet is the existing one-ID.

When there are multiple one-IDs in the relational subnet in the first ID relational network, the one-ID with the earliest creation time and the most times of merging or splitting is selected from the multiple one-IDs as the one-ID of the relational subnet. ID.

If the creation time and the number of times of merging or splitting are the same for multiple one-IDs, one of the multiple one-IDs is randomly selected as the one-ID of the relational subnet.

In the case where there is no one-ID in the relationship subnet in the first ID relationship network, a new one-ID is generated according to the relationship between the ID node and the ID node in the relationship subnet, and the newly generated one-ID is used as the relationship The one-ID of the subnet.

The determination of the one-ID of the relationship subnet in the first ID relationship network is clarified, the one-ID has opened up the association between the IDs, and the determination of the one-ID is conducive to obtaining all ID nodes and all IDs related to it according to the one-ID. Node relationship.

Step 103: Clean up the first ID relationship network according to the activity of the ID nodes in the first ID relationship network and the activity of the ID node relationship, and obtain the second ID relationship network corresponding to the T-th day.

Specifically, the ID nodes whose node activity is lower than the node activity threshold in the first ID relation network are removed from the first ID relation network; the ID nodes whose relation activity is lower than the relation activity threshold in the first ID relation network are removed from the first ID relation network. The node relationship is cleared from the first ID relationship network. Obtain the second ID relationship network corresponding to the T-th day according to the cleaned first ID relationship network.

Optionally, before cleaning the first ID relationship network according to the activity of the ID nodes and the activity of the ID node relationship in the first ID relationship network, the method further includes:

Update the attributes of the ID node and the attributes of the ID node relationship respectively;

Perform feature extraction on the updated attribute of the ID node and the updated attribute of the ID node relationship, and obtain the feature value of the ID node and the feature value of the ID node relationship;

Obtain the activity of the ID node according to the eigenvalue of the ID node and the weight corresponding to the eigenvalue of the ID node;

The activity of the ID node relationship is obtained according to the eigenvalues of the ID node relationship and the weights corresponding to the eigenvalues of the ID node relationship.

Specifically, there are two types of relationships between ID nodes and ID nodes in the first ID relationship network. One is the relationship between the ID node and the ID node that does not appear on the T-1th day and appears on the T-th day; the other is the relationship between the ID node and the ID node on the T-1th day. The ID node and ID node relationship that appear on day T and do not appear on day T.

For the ID nodes and ID node relationships that do not appear on the T-1th day and appear on the T day, the attributes of all ID nodes in Table 1 are updated, and the attributes of all ID node relationships in Table 2 are updated.

For ID nodes and ID node relationships that appear on day T-1 and do not appear on day T, only the node liveness attribute of the ID node is updated, only the relationship liveness attribute of the ID node relationship is updated, and other attributes are not updated.

Features for evaluating the node liveness of the ID node and evaluating the relationship liveness of the ID node relationship are extracted from the updated attribute of the ID node and the updated attribute of the ID node relationship. Table 4 is a feature extraction processing table of ID node and ID node relationship, and s in Table 4 represents the feature value before normalized feature processing.

Table 4 Feature extraction processing table of ID node and ID node relationship

As can be seen from Table 4, the features used to evaluate the node activity of ID nodes include node priority, the ratio of active days to the number of days from the first record to the current processing date, the average active interval days, and the distance from the latest active date to the current processing date. The number of days of the date, the standard deviation of the active interval days, the regularity of the interval days, and the node degree value.

The features used to evaluate the relationship activity of the ID node relationship are the ratio of the number of active days to the number of days from the first record to the current processing date, the average number of days between active days, the number of days from the most recent active date to the current processing date, and the two nodes of the relationship are most recently active. The number of days apart between dates, the standard deviation of the active interval days, the regularity of the interval days, and the relationship priority.

The extracted features are normalized to obtain the feature value of the ID node and the feature value of the ID node relationship.

The activity of the ID node is obtained according to the eigenvalue of the ID node and the weight corresponding to the eigenvalue of the ID node.

The expression for the liveness of the ID node is as follows:

In the formula, X represents the activity of the ID node, x represents the eigenvalue after normalization processing, the subscript of x corresponds to the number in Table 4, α represents the weight corresponding to the eigenvalue of each node, α ₁ is x ₂ Corresponding weight, α ₂ is the weight corresponding to x ₃ , α ₃ is the weight corresponding to x ₄ , α ₄ is the weight corresponding to x ₅ , α ₅ is the weight corresponding to x ₇ , α ₆ is the weight corresponding to x ₈ , weight Scored by experts through the analytic hierarchy process.

The activity of the ID node relationship is obtained according to the eigenvalues of the ID node relationship and the weights corresponding to the eigenvalues of the ID node relationship.

The expression for the liveness of the ID node relationship is as follows:

In the formula, Y represents the activity of the ID node relationship, x represents the eigenvalue after normalization processing, the subscript of x corresponds to the number in Table 4, β represents the weight corresponding to each relationship eigenvalue, β ₁ is x ₂ corresponds to the weight, β ₂ is the weight corresponding to x ₃ , β ₃ is the weight corresponding to the combination of x ₄ and x ₅ , β ₄ is the weight corresponding to x ₆ , β ₅ is the weight corresponding to x ₇ , and the weights pass the AHP Scored by experts.

By updating the attributes of the ID node and the relationship between the ID nodes, extracting features according to the updated attributes, normalizing the features, and finally obtaining the activity according to the eigenvalue and the weight corresponding to the eigenvalue, the calculation of the activity is clarified The method further uses the activity to clean up the ID node and the ID node relationship.

Optionally, clean up the first ID relationship network according to the activity of the ID nodes in the first ID relationship network and the activity degree of the ID node relationship, and obtain the second ID relationship network corresponding to the T-th day, including:

In the case that the activity of the ID node in the first ID relationship network is less than the node activity threshold, clear the ID node out of the first ID relationship network;

When the activity of the ID node relationship in the first ID relationship network is less than the relationship activity threshold, clear the ID node relationship out of the first ID relationship network;

Obtain the second ID relationship network corresponding to the T-th day according to the cleaned first ID relationship network.

Specifically, after the activity of the ID node is obtained, the activity of the ID node is compared with the node activity threshold, and the ID nodes whose activity is less than the node activity threshold are cleaned up.

After obtaining the activity of the ID node relationship, compare the activity of the ID node relationship with the relationship activity threshold, and clean up the ID node relationship whose activity is less than the relationship activity threshold.

Obtain the second ID relationship network corresponding to the T-th day according to the cleaned first ID relationship network.

By cleaning up ID nodes whose activity is lower than the threshold, the expired ID can be cleaned up. By disconnecting the ID node relationship whose activity is lower than the threshold, the weak relationship between ID nodes can be cleaned up, and the relationship between user IDs and IDs can be improved. reliability, accuracy and stability of the network.

Optionally, according to the cleaned up first ID relationship network, obtain the second ID relationship network corresponding to the T-th day, including:

In the case that the clearing of the ID node or ID node relationship does not cause the split of the relationship subnet in the first ID relationship network, the unified identity of the relationship subnet in the second ID relationship network is the relationship child in the first ID relationship network. The unified identity of the network;

In the case where the cleanup of the ID node or ID node relationship causes the relationship subnet in the first ID relationship network to be split into multiple relationship subnets, a relationship in the multiple relationship subnets in the second ID relationship network The unified identity of the subnet is the unified identity of the relationship subnet in the first ID relationship network, and the unified identity of other relationship subnets in the multiple relationship subnets is the newly generated unified identity.

Specifically, when cleaning up the relationship between ID nodes and ID nodes, it is necessary to consider whether a new relational subnet will be generated, that is, whether the relational subnet will be split when cleaning is performed, and whether after splitting, after splitting The one-ID of the relational subnet is determined.

In the case where the cleanup of the ID node or ID node relationship does not cause the split of the relationship subnet in the first ID relationship network, the one-ID of the relationship subnet remains unchanged, and the one-ID of the relationship subnet in the second ID relationship network -ID is the one-ID of the relationship subnet in the first ID relationship network.

For example, in the first ID relationship network, the one-ID of the relationship subnet 1 is one-ID1, some ID nodes and some ID node relationships are cleaned up from the relationship subnet 1, and the relationship subnet 1 after cleaning is not split. , the relationship subnet 1 after cleaning is the relationship subnet 2 in the second ID relationship network, and the one-ID of the relationship subnet 2 is still one-ID1.

In the case that the cleaning of the ID node or ID node relationship causes the relationship subnet in the first ID relationship network to be split into multiple relationship subnets, select a relationship that conforms to the preset rule from the multiple relationship subnets according to a preset rule The subnet inherits the one-IDs of the relational subnets in the first ID relational network, and the one-IDs of other relational subnets are new one-IDs generated according to their own ID node and the ID node relationship.

The preset rule may be to select a relationship subnet from multiple relationship subnets in the order of "the most active ID node, the most active days of the ID node, the largest number of ID nodes, and the most recent active date of the ID node".

Select a relationship randomly from multiple relationship subnets under the condition that "the most active ID node, the most active days of the ID node, the largest number of ID nodes, and the most recent active date of the ID node" corresponding to multiple relationship subnets are all the same. subnet.

For example, in the first ID relationship network, the one-ID of the relationship subnet A is one-IDA, some ID nodes and some ID node relationships are cleaned up from the relationship subnet A, and the cleaned relationship subnet A splits. The cleaned-up relational subnet A is the relational subnet B, the relational subnet C, and the relational subnet D in the second ID relational network.

Compare the ID nodes corresponding to relation subnet B, relation subnet C, and relation subnet D with the highest activity. If the ID node activity corresponding to relation subnet B is the largest among the three relation subnets, then the relation subnet The one-ID of B is the one-IDA of the relational subnet A, and the relational subnet C and the relational subnet D generate new one-IDs based on their own ID nodes and the relationship between the ID nodes.

If the activity of ID nodes corresponding to relation subnet B, relation subnet C and relation subnet D are the same at maximum, then compare the active days of ID nodes corresponding to relation subnet B, relation subnet C and relation subnet D respectively. At most, the number of active days of the ID node corresponding to the relationship subnet B is at most 15 days, the number of active days of the ID node corresponding to the relationship subnet C is at most 25 days, and the number of active days of the ID node corresponding to the relationship subnet D is at most 12 days. The one-ID of the network C is the one-IDA of the relational subnet A, and the relational subnet B and the relational subnet D generate new one-IDs according to their own ID nodes and the relationship between the ID nodes.

If relationship subnet B, relationship subnet C, and relationship subnet D have the same "ID node activity is the largest, ID node has the most active days, the number of ID nodes is the largest, and the ID node's latest active date" are the same, the relationship Subnet B, relational subnet C and relational subnet D randomly select a relational subnet to inherit the one-IDA of relational subnet A, and the other two relational subnets generate a new one based on their own ID node and ID node relationship -ID.

Update the acquired second ID relational network to the graph database, and update the relational subnet with the updated attributes and activity to the graph database for the relational subnet that already exists one-ID.

According to whether the cleaning of the ID node and the relationship between the ID nodes leads to the split of the ID relationship subnet, the one-ID of the ID relationship subnet is re-determined, which improves the reliability, accuracy and stability of the ID relationship network.

Optionally, clean up the first ID relationship network according to the activity of the ID nodes in the first ID relationship network and the activity degree of the ID node relationship, and after acquiring the second ID relationship network corresponding to the T day, it also includes:

Obtain the relationship between inactive ID nodes and inactive ID nodes within a preset time from the source ID data record;

Update the activity of inactive ID nodes and the activity of inactive ID node relationships;

The second ID relationship network is cleaned up according to the updated activity degree of the inactive ID node and the updated activity degree of the inactive ID node relationship, and the third ID relationship network corresponding to the T-th day is obtained.

Specifically, only the ID nodes of the T-th day, the ID node relationships of the T-th day, and the ID relationship network of the T-1th day have been screened, while the inactive ID nodes and inactive IDs within the preset time have been screened. Node relationships are not considered.

From the graph database constructed by the source ID data records, obtain the inactive ID nodes and the inactive ID node relationships within the preset time. The preset time can be reasonably set according to the type of the ID node and the type of the relationship between the ID nodes.

The activity of the inactive ID node is updated by using the attribute of the inactive ID node. The calculation method of the active degree of the ID node is the same as the previous one, and will not be repeated here.

The activeness of the inactive ID node relationship is updated by using the attribute of the inactive ID node relationship. The calculation method of the activeness of the ID node relationship is the same as the previous one, and will not be repeated here.

Compare the updated liveness of the inactive ID nodes with the node liveness threshold, and clean up the inactive ID nodes whose liveness is less than the node liveness threshold.

Compare the updated liveness of inactive ID node relationships with the relationship liveness threshold, and clean up inactive ID node relationships whose liveness is less than the relationship liveness threshold.

According to the second ID relationship network in which the relationship between the inactive ID nodes and the inactive ID nodes is cleaned up, the third ID relationship network corresponding to the T-th day is obtained.

After cleaning the relationship between the inactive ID nodes and the inactive ID nodes, it is still necessary to consider whether a new relationship subnet will be generated, and it is still necessary to consider the determination of the one-ID of the new relationship subnet.

In the case where the cleanup of the inactive ID node or the inactive ID node relationship does not result in the splitting of the relationship subnet in the second ID relationship network, the one-ID of the relationship subnet remains unchanged, and the third ID relationship network in the The one-ID of the relationship subnet is the one-ID of the relationship subnet in the second ID relationship network.

In the case that the cleanup of inactive ID nodes or inactive ID node relationships causes the relationship subnet in the second ID relationship network to be split into multiple relationship subnets, select from the multiple relationship subnets according to preset rules. One relational subnet of the preset rule inherits the one-ID of the relational subnet in the second ID relational network, and the one-IDs of other relational subnets are new one-IDs generated according to their own ID node and the ID node relationship. The third ID relation network is updated into the graph database, and the ID relation subnets whose ID nodes are all deleted are deleted from the graph database.

Generate an identity mapping dictionary corresponding to the T-th day according to the updated graph database.

The reliability, accuracy and stability of the ID relationship network are improved by updating the activity level of ID nodes and ID node relationships that are not active within a preset time, and then cleaning up the ID relationship network according to the updated activity level.

The method for implementing ID Mapping based on a graph database provided by the present invention realizes the clearing of expired IDs by clearing the ID nodes whose activity is lower than the threshold, and solves the problem of ID expiration. By disconnecting the ID nodes whose activity is lower than the threshold, Realize the cleaning of weak relationship between ID nodes, solve the problem of ID reuse and ID complex relationship, thereby improving the reliability, accuracy and stability of the user ID relationship network

The apparatus for implementing ID Mapping based on a graph database provided by the present invention is described below. The apparatus for implementing ID Mapping based on a graph database described below and the method for implementing ID Mapping based on a graph database described above may refer to each other correspondingly.

FIG. 3 is a schematic structural diagram of an apparatus for implementing ID Mapping based on a graph database provided by the present invention. As shown in FIG. 3 , the present invention also provides an apparatus for implementing ID Mapping based on a graph database, including: a first acquisition module 301, a second The acquisition module 302 and the third acquisition module 303, wherein:

The first obtaining module 301 is used to obtain the ID node relationship that occurs on the T-th day and the ID node relationship that occurs on the T-th day from the source ID data record;

The second obtaining module 302 is configured to identify and connect the ID nodes that appear on the T-th day, the ID node relationships that appear on the T-th day, and the ID relation network corresponding to the T-1 day, to obtain the ID node corresponding to the T-th day. an ID network;

The third obtaining module 303 is configured to clean up the first ID relationship network according to the activity of the ID nodes in the first ID relationship network and the activity of the ID node relationship, and obtain the second ID relationship network corresponding to the T day .

Specifically, the apparatus for implementing ID Mapping based on a graph database provided in the embodiment of the present application can implement all the method steps implemented by the above method embodiments, and can achieve the same technical effect. The same parts and beneficial effects of the embodiments are described in detail.

FIG. 4 is a schematic structural diagram of an electronic device provided by the present invention. As shown in FIG. 4 , the electronic device may include: a processor (processor) 410, a communication interface (Communications Interface) 420, a memory (memory) 430 and a communication bus 440, The processor 410 , the communication interface 420 , and the memory 430 communicate with each other through the communication bus 440 . The processor 410 may invoke logic instructions in the memory 430 to execute a method for implementing ID Mapping based on a graph database, the method comprising: obtaining an ID node that occurs on the T-th day and an ID that occurs on the T-th day from a source ID data record Node relationship; identify and connect the ID node that occurs on the T day, the ID node relationship that occurs on the T day, and the ID relationship network corresponding to the T-1 day, and obtain the first ID relationship network corresponding to the T day. ; Clean up the first ID relationship network according to the activity degree of the ID node in the first ID relationship network and the activity degree of the ID node relationship, and obtain the second ID relationship network corresponding to the Tth day.

In addition, the above-mentioned logic instructions in the memory 430 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product in essence, or the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .

In another aspect, the present invention also provides a computer program product, the computer program product includes a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer can Executing the method for implementing ID Mapping based on a graph database provided by the above methods, the method includes: obtaining the ID node appearing on the T-th day and the ID node relationship appearing on the T-th day from the source ID data record; The ID node that appears on the T day, the ID node relationship that appears on the T day, and the ID relationship network corresponding to the T-1 day are identified and connected, and the first ID relationship network corresponding to the T day is obtained; according to the first ID The activity of the ID node and the activity of the relationship between the ID nodes in the relationship network are used to clean up the first ID relationship network, and obtain the second ID relationship network corresponding to the T-th day.

In another aspect, the present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored, the computer program is implemented by a processor to execute the method for implementing IDMapping based on a graph database provided by the above methods, The method includes: obtaining the relationship between the ID node appearing on the T-th day and the ID node appearing on the T-th day from the source ID data record; The relationship and the ID relationship network corresponding to the T-1th day are identified and connected, and the first ID relationship network corresponding to the Tth day is obtained; The first ID relationship network is cleaned up, and the second ID relationship network corresponding to the T-th day is obtained.

The device embodiments described above are only illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in One place, or it can be distributed over multiple network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this embodiment. Those of ordinary skill in the art can understand and implement it without creative effort.

From the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and certainly can also be implemented by hardware. Based on this understanding, the above-mentioned technical solutions can be embodied in the form of software products in essence or the parts that make contributions to the prior art, and the computer software products can be stored in computer-readable storage media, such as ROM/RAM, magnetic A disc, an optical disc, etc., includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the methods described in various embodiments or some parts of the embodiments.

The terms "first", "second", etc. in the embodiments of the present application are used to distinguish similar objects, and are not used to describe a specific sequence or sequence. It is to be understood that the terms so used are interchangeable under appropriate circumstances so that the embodiments of the present application can be practiced in sequences other than those illustrated or described herein, and that "first", "second" distinguishes Usually it is a class, and the number of objects is not limited. For example, the first object may be one or multiple.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that it can still be The technical solutions described in the foregoing embodiments are modified, or some technical features thereof are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (10)

1. a method for realizing ID Mapping based on graph database, is characterized in that, comprises: From the source ID data record, obtain the relationship between the ID node that occurs on the T-th day and the ID node that occurs on the T-th day; The ID node that occurs on the T day, the ID node relationship that occurs on the T day, and the ID relationship network corresponding to the T-1 day are identified and connected, and the first ID relationship network corresponding to the T day is obtained; The first ID relationship network is cleaned up according to the activity degree of the ID nodes in the first ID relationship network and the activity degree of the ID node relationship, and the second ID relationship network corresponding to the T-th day is obtained. 2 . The method for implementing ID Mapping based on a graph database according to claim 1 , wherein, according to the activity of the ID nodes in the first ID relationship network and the activity of the ID node relationship, the first Before the ID network is cleaned up, it also includes: respectively update the attribute of the ID node and the attribute of the ID node relationship; Perform feature extraction on the updated attribute of the ID node and the updated attribute of the ID node relationship, respectively, to obtain the feature value of the ID node and the feature value of the ID node relationship; Obtain the activity of the ID node according to the feature value of the ID node and the weight corresponding to the feature value of the ID node; The activity of the ID node relationship is acquired according to the feature value of the ID node relationship and the weight corresponding to the feature value of the ID node relationship. 3. The method for implementing ID Mapping based on a graph database according to claim 1, wherein the first ID relationship is determined according to the activity of the ID node and the activity of the ID node relationship in the first ID relationship network. The network is cleaned up, and the second ID relationship network corresponding to the T day is obtained, including: In the case that the activity of the ID node in the first ID relationship network is less than the node activity threshold, clearing the ID node out of the first ID relationship network; In the case that the activity of the ID node relationship in the first ID relationship network is less than the relationship activity threshold, clearing the ID node relationship out of the first ID relationship network; According to the cleaned first ID relationship network, the second ID relationship network corresponding to the T-th day is acquired. 4. the method for realizing ID Mapping based on graph database according to claim 3, is characterized in that, described according to the described first ID relation network after cleaning, obtain the second ID relation network corresponding to T day, comprising: In the case that the clearing of the ID node or the relationship between the ID nodes does not result in the splitting of the relationship subnet in the first ID relationship network, the unified identity of the relationship subnet in the second ID relationship network is the Describe the unified identity of the relationship subnet in the first ID relationship network; In the case where the cleanup of the ID node or the relationship of the ID node causes the relationship subnet in the first ID relationship network to be split into multiple relationship subnets, the multiple relationships in the second ID relationship network The unified identity of a relational subnet in the subnet is the unified identity of the relational subnet in the first ID relational network, and the unified identity of other relational subnets in the plurality of relational subnets is newly generated. unified identity. 5. The method for implementing ID Mapping based on a graph database according to claim 1, wherein the first ID node activity and the activity degree of the ID node relationship in the first ID relationship network The ID relationship network is cleaned up, and after obtaining the second ID relationship network corresponding to the T day, it also includes: Obtain the relationship between inactive ID nodes and inactive ID nodes within a preset time from the source ID data record; updating the activity of the inactive ID node and the activity of the inactive ID node relationship; The second ID relationship network is cleaned up according to the updated activity degree of the inactive ID node and the updated activity degree of the inactive ID node relationship, and the third ID relationship network corresponding to the T-th day is obtained. 6 . The method for implementing ID Mapping based on a graph database according to claim 1 , wherein the ID nodes that appear on the T-th day, the ID node relationships that appear on the T-th day, and the T-1 The ID relationship network corresponding to the day is identified and connected, and before obtaining the first ID relationship network corresponding to the T day, it also includes: According to the ID node that appears on the T day and the identity mapping dictionary corresponding to the T-1 day, obtain the unified identity that exists on the T-1 day; Obtain the ID relationship network corresponding to the T-1th day according to the unified identity identifier that exists on the T-1th day. 7 . The method for implementing ID Mapping based on a graph database according to claim 1 , wherein the ID nodes that appear on the T-th day, the ID node relationships that appear on the T-th day, and the T-1 The ID relationship network corresponding to the day is identified and connected, and after obtaining the first ID relationship network corresponding to the T day, it also includes: In the case that the relationship subnet in the first ID relationship network has a unified identity, the unified identity of the relationship subnet is the existing unified identity; In the case that the relationship subnet in the first ID relationship network has more than one unified identity, the unified identity of the relationship subnet is the unified identity with the earliest creation time and the most times of merging or splitting; In the case where the relationship subnet in the first ID relationship network does not have a unified identity, the unified identity of the relationship subnet is a newly generated unified identity. 8. A device for realizing ID Mapping based on a graph database, characterized in that, comprising: The first acquisition module is used to obtain the ID node relationship that occurs on the T day and the ID node relationship that occurs on the T day from the source ID data record; The second acquisition module is used to identify and connect the ID nodes that appear on the T-th day, the ID node relationships that appear on the T-th day, and the ID relationship network corresponding to the T-1 day, and obtain the ID node corresponding to the T-th day. an ID network; The third acquisition module is used to clean up the first ID relationship network according to the activity of the ID nodes in the first ID relationship network and the activity degree of the ID node relationship, and obtain the second ID relationship network corresponding to the T day . 9. An electronic device comprising a memory, a processor and a computer program stored on the memory and running on the processor, wherein the processor implements the computer program as claimed in the claims The method for implementing ID Mapping based on a graph database according to any one of 1 to 7. 10. A non-transitory computer-readable storage medium on which a computer program is stored, characterized in that, when the computer program is executed by a processor, the graph-based database according to any one of claims 1 to 7 is implemented A method to implement IDMapping.

Images