CN113988870A - Group partner identification method and device - Google Patents
Group partner identification method and device Download PDFInfo
- Publication number
- CN113988870A CN113988870A CN202111258230.4A CN202111258230A CN113988870A CN 113988870 A CN113988870 A CN 113988870A CN 202111258230 A CN202111258230 A CN 202111258230A CN 113988870 A CN113988870 A CN 113988870A
- Authority
- CN
- China
- Prior art keywords
- virtual
- virtual relationship
- business objects
- relationship
- attribute
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 230000011218 segmentation Effects 0.000 claims description 94
- 238000010586 diagram Methods 0.000 claims description 40
- 238000005065 mining Methods 0.000 claims description 20
- 238000012216 screening Methods 0.000 claims description 16
- 238000003860 storage Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 8
- 238000013459 approach Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000004576 sand Substances 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000010923 batch production Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000008014 freezing Effects 0.000 description 1
- 238000007710 freezing Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000011144 upstream manufacturing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/38—Payment protocols; Details thereof
- G06Q20/40—Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
- G06Q20/401—Transaction verification
- G06Q20/4016—Transaction verification involving fraud or risk level assessment in transaction processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Security & Cryptography (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the specification provides a group partner identification method and a group partner identification device, wherein the method comprises the steps of obtaining a plurality of business objects and a plurality of associated non-entity attributes of the business objects, wherein the non-entity attributes respectively correspond to a plurality of non-entity associated objects of the business objects; determining a virtual relationship among a plurality of business objects according to a plurality of non-entity attributes, wherein the representation of the virtual relationship takes a plurality of non-entity associations as media for association; generating an object virtual relationship graph according to the business object and the virtual relationship, wherein the object virtual relationship graph comprises a plurality of nodes and edges among the nodes, the nodes correspond to the business object, and the edges correspond to the virtual relationship; and excavating an object group consisting of a plurality of business objects based on the object virtual relationship graph.
Description
Technical Field
One or more embodiments of the present disclosure relate to the field of digital security and risk prevention and control, and more particularly, to a group partner identification method and apparatus.
Background
In recent years, the business scale and the user scale of a plurality of internet enterprises are continuously enlarged, and correspondingly, illegal industries which obtain illegal benefits from the businesses by illegal means also have the trend of industrial chaining of upstream and downstream division, and further the illegal means are continuously developed and rapidly iterated, so that the difficulty of the internet business enterprises in resisting the means is continuously improved. For example, the illegal industry purchases personal identity information in an illegal way, and the personal identity information is used for false registration of the vacant companies in batches, account opening of an enterprise payment platform and the like. Further, the personal identity information, the shell company and the payment account number are utilized to carry out false business activities aiming at obtaining illegal benefits and having abnormal purposes, so that economic losses of internet enterprises providing services are caused. With this new form of illicit approach, it is difficult to efficiently identify risk object groups consisting of, for example, a large number of false accounts registered by illicit industries using existing technical methods.
Therefore, a new method of group identification is needed.
Disclosure of Invention
The embodiments in the present specification aim to provide a more effective identification method for a group formed by risky business objects in business activities, and solve the deficiencies in the prior art.
According to a first aspect, there is provided a group partner identification method comprising:
acquiring a plurality of business objects and a plurality of associated non-entity attributes thereof, wherein the non-entity attributes respectively correspond to a plurality of non-entity associated objects of the business objects;
determining a virtual relationship among the plurality of business objects according to the plurality of non-entity attributes, wherein the virtual relationship represents that the plurality of non-entity associations are associated by taking the plurality of non-entity associations as media;
generating an object virtual relationship graph according to the business object and the virtual relationship, wherein the object virtual relationship graph comprises a plurality of nodes and edges among the nodes, the nodes correspond to the business object, and the edges correspond to the virtual relationship;
and mining an object group consisting of a plurality of business objects based on the object virtual relationship graph.
In one possible embodiment, the number of non-entity attributes includes: and one or more of the name, the registration address, the email address and the associated website of the business object.
In one possible embodiment, the number of non-entity attributes includes a first attribute and a second attribute;
determining a virtual relationship between the plurality of business objects according to the plurality of non-entity attributes, including:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
the generating of the object virtual relationship graph comprises:
generating a first relation graph according to the business object and the first virtual relation;
generating a second relation graph according to the business object and the second virtual relation;
and generating the object virtual relationship diagram based on the first relationship diagram and the second relationship diagram.
In one possible embodiment, the number of non-entity attributes includes a first attribute and a second attribute;
determining a virtual relationship between the plurality of business objects according to the plurality of non-entity attributes, including:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
determining a synthetic virtual relationship based at least on the first virtual relationship and the second virtual relationship;
the generating of the object virtual relationship graph comprises:
and generating the object virtual relationship graph according to the service object and the comprehensive virtual relationship.
In a possible implementation manner, the plurality of non-entity attributes include a first attribute, and the first attribute is an associated website of the business object;
determining a virtual relationship between the plurality of objects according to the non-entity attributes, including:
and determining the virtual relationship according to one or more of the name of the associated website and the webpage picture of the associated website.
In one possible embodiment, the number of non-entity attributes includes a first attribute, and determining the virtual relationship between the plurality of objects according to the number of non-entity attributes includes:
performing text word segmentation on the texts with the first attributes of the plurality of service objects to obtain word segmentation results of the texts;
and determining a first virtual relationship corresponding to the first attribute among the plurality of business objects based on the word segmentation result.
In one possible implementation, the word segmentation result comprises a plurality of segmented words obtained by text word segmentation;
determining a first virtual relationship corresponding to the first attribute among the plurality of business objects based on the word segmentation result, including:
encoding each division word of the text;
and determining whether a first virtual relation exists between the plurality of business objects based on codes corresponding to one or more dividing words of the text.
In one possible embodiment, text-tokenizing the text of the non-entity attribute of a plurality of business objects includes:
performing pattern matching on the text;
and if the text conforms to a preset text mode, performing text word segmentation on the text according to the text mode to obtain a word segmentation result.
In one possible embodiment, determining a first virtual relationship corresponding to the first attribute among a plurality of business objects based on the word segmentation result includes:
determining whether a first virtual relationship exists among the plurality of business objects based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same word segmentation/word segmentation combination/mode word combination, wherein the word segmentation is obtained based on text word segmentation, and the mode word is obtained based on text word segmentation according to a text mode.
In one possible embodiment, determining a first virtual relationship corresponding to the first attribute among a plurality of business objects based on the word segmentation result includes:
determining whether a first virtual relationship exists among the plurality of business objects or not based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same hot word combination or not, wherein the hot word combination comprises at least one hot word.
In a possible implementation manner, the hotword is determined according to the frequency of occurrence of each divided word in the word segmentation result in the texts with a plurality of non-entity attributes of the plurality of business objects.
In a possible implementation manner, the hotword is determined according to a bad sample rate of each division word in the division word result, and the bad sample rate is determined according to a ratio of a business object corresponding to the division word and having a pre-labeled high-risk label to all business objects corresponding to the division word.
In one possible implementation, mining an object group composed of a plurality of the business objects based on the object virtual relationship graph comprises:
and acquiring a plurality of node groups comprising a plurality of nodes based on a community mining algorithm according to the object virtual relationship graph, and acquiring object groups corresponding to the node groups.
In one possible implementation, the community mining algorithm includes one of a Louvain algorithm, a tag propagation algorithm, and an Infomap algorithm.
In one possible embodiment, the method further comprises:
based on an object entity relationship graph obtained according to entity attributes of a plurality of service objects, screening the object group and/or the service objects in the object group to obtain a screening result, wherein the entity attributes correspond to entity associations of the service objects;
and taking the screening result as an output result of the group recognition.
In one possible embodiment, the method further comprises:
based on a pre-acquired negative business object library, screening the object group and/or the business objects in the object group to obtain a screening result;
and taking the screening result as an output result of the group recognition.
According to a second aspect, there is provided a group partner identifying apparatus comprising:
an object non-entity attribute obtaining unit configured to obtain a plurality of business objects and a plurality of non-entity attributes associated therewith, the plurality of non-entity attributes corresponding to a plurality of non-entity associations possessed by the business objects, respectively;
the virtual relation determining unit is configured to determine a virtual relation among the plurality of business objects according to the plurality of non-entity attributes, wherein the virtual relation represents that the plurality of non-entity associations are associated by taking the plurality of non-entity associations as media;
a virtual relationship graph generating unit configured to generate an object virtual relationship graph according to the service object and the virtual relationship, where the object virtual relationship graph includes a plurality of nodes and edges between the nodes, where the nodes correspond to the service object and the edges correspond to the virtual relationship;
and the group-partner digging unit is configured to dig out an object group formed by a plurality of business objects based on the object virtual relationship graph.
In one possible embodiment, the number of non-entity attributes includes a first attribute and a second attribute;
a virtual relationship determination unit further configured to:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
a virtual relationship diagram generation unit further configured to:
generating a first relation graph according to the business object and the first virtual relation;
generating a second relation graph according to the business object and the second virtual relation;
and generating the object virtual relationship diagram based on the first relationship diagram and the second relationship diagram.
In one possible embodiment, the number of non-entity attributes includes a first attribute and a second attribute;
a virtual relationship determination unit further configured to:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
determining a synthetic virtual relationship based at least on the first virtual relationship and the second virtual relationship;
a virtual relationship diagram generation unit further configured to:
and generating the object virtual relationship graph according to the service object and the comprehensive virtual relationship.
In one possible embodiment, the number of non-entity attributes includes a first attribute,
a virtual relationship determination unit further configured to:
performing text word segmentation on the texts with the first attributes of the plurality of service objects to obtain word segmentation results of the texts;
and determining a first virtual relationship corresponding to the first attribute among the plurality of business objects based on the word segmentation result.
In a possible implementation, the virtual relationship determining unit is further configured to:
performing pattern matching on the text;
and if the text conforms to a preset text mode, performing text word segmentation on the text according to the text mode to obtain a word segmentation result.
In a possible implementation, the virtual relationship determining unit is further configured to:
determining whether a first virtual relationship exists among the plurality of business objects based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same word segmentation/word segmentation combination/mode word combination, wherein the word segmentation is obtained based on text word segmentation, and the mode word is obtained based on text word segmentation according to a text mode.
In a possible implementation, the virtual relationship determining unit is further configured to:
determining whether a first virtual relationship exists among the plurality of business objects or not based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same hot word combination or not, wherein the hot word combination comprises at least one hot word.
According to a third aspect, there is provided a computer readable storage medium having stored thereon a computer program which, when executed in a computer, causes the computer to perform the method of the first aspect.
According to a fourth aspect, there is provided a computing device comprising a memory having stored therein executable code and a processor that, when executing the executable code, implements the method of the first aspect.
By using one or more of the methods, the devices, the computing equipment and the storage medium in the aspects, the group of the risky business objects in the business activity can be effectively identified, illegal business performed through the group can be conveniently controlled, and economic loss caused by the illegal business is reduced.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 illustrates a schematic diagram of identifying a group based on an entity's media;
FIG. 2 illustrates a schematic diagram of identifying a group based on a funding relationship;
FIG. 3 illustrates a schematic diagram of a group identification method according to an embodiment of the present disclosure;
FIG. 4 illustrates an example diagram of a text schema and its corresponding word encoding in accordance with an embodiment of the present description;
FIG. 5 illustrates an example diagram of hotwords and hotword combinations in accordance with an embodiment of the present description.
FIG. 6 illustrates a flow diagram of a method of group identification according to an embodiment of the present description;
fig. 7 is a block diagram illustrating a group recognition apparatus according to an embodiment of the present disclosure.
Detailed Description
The solution provided by the present specification will be described below with reference to the accompanying drawings.
As described above, as the business scale of many internet enterprises is continuously enlarged, the technical means of illegal industry is rapidly iterated, and the difficulty of enterprises in dealing with illegal means, such as real-time countermeasure, is continuously increased. For example, illegal industries purchase personal identification information, and use the personal information to bulk-falsely register a vacant company, register a private domain name, open an enterprise payment account, and the like for performing illegal business activities on the internet. For an internet platform or a service enterprise, it is very difficult to accurately identify such a group consisting of high-risk account numbers such as batch registration and buying and selling through the existing technical means in a service cycle, especially in an early stage of a service, such as a merchant admission stage, so as to reduce the loss caused by illegal service activities performed by the service enterprise.
In order to more clearly describe the core idea of the group identification method provided by the embodiment of the present specification, the following first introduces the existing main identification method for high-risk group. Currently, the main technical solutions for performing the above group identification in the industry include the following:
the first approach is to identify risk groups through entity-media relationships (or strong-media relationships). The scheme is mainly characterized in that entity medium relations of a plurality of business objects (such as accounts) are established, and then a set of business objects with high association aggregation degree is identified from the business objects to serve as risk groups according to the entity medium relations. The entity medium is a medium relying on entity equipment or entity identification, and the entity equipment can be various equipment possibly used by a user in the process of using the internet business, such as a smart phone, a PC computer, a tablet computer, a wearable intelligent device, a router and the like; the entity identity certificate may be issued by an authority after user identity authentication, and may be a certification medium capable of certifying or identifying the user identity, for example, various certificates (identity cards and passports) issued by the police department, bank cards issued by banks, SIM cards (corresponding to mobile phone numbers) distributed by operators, and the like. Because the method relies on the characteristics of the entity, the entity medium has stronger equipment identity/user identity characterization capability. Common entity-medium relationships are, for example, associations between different business objects established by using entity objects such as certificates, bank cards, mobile phone numbers, and devices Mac as media. For example, the physical-media relationship between accounts is established according to whether multiple accounts use the same device, mobile phone number, bank card, etc. to perform login, signing operation, etc. or whether multiple accounts belong to one company/legal person within a predetermined period of time.
Figure 1 shows a schematic diagram of identifying a group based on an entity's medium. The disadvantage of this solution is that the illicit industry can use various means, especially masquerading its registered batch accounts as separate accounts without strong media association, operated by different natural persons, after it perceives, for example through experience, that the service enterprise recognizes policy rules based on the medium of the entity. Logging in an account at a different terminal, for example, through a low-priced acquisition device; or illegally acquiring the identity information of the idle personnel, carrying out false enterprise registration to obtain the business card of the enterprise and the like, and further respectively registering the account by utilizing the identity information, the business card and the like. Making it difficult to identify these false accounts based on the physical media relationship.
The second approach is to identify the risk groups through the fund relationship. The scheme mainly judges whether a group relationship exists in a plurality of business objects through fund exchange among different business objects. For example, whether risk groups exist in multiple accounts can be identified through whether fund action relations such as transfer, account separation, collection and the like occur among the multiple accounts within a limited time period and for example, the proportion condition of the transaction number and the transaction amount is combined. Figure 2 shows a schematic diagram of identifying a group according to a funding relationship. A disadvantage of this approach is that in many business scenarios, especially in the early stages of many business activities, the risky accounts may not have obvious funds transactions, or their funds transactions are deliberately circumvented by illegal industries, for example, by using transactions prior to account registration. Therefore, the effect of the group identification by the fund relationship is also not obvious in these cases.
In order to solve the above technical problem, an embodiment of the present specification provides a group partner identification method. The method has the core idea that firstly, the non-entity attributes corresponding to the non-entity associated objects of the business objects are utilized to establish the virtual relationship among a plurality of business objects, then the business objects and the virtual relationship among the business objects are utilized to establish a virtual object relationship graph, and finally, the virtual object relationship graph is utilized to dig out a set of risk objects in the business objects through a community digging algorithm, namely risk object gangs. In contrast to the aforementioned entity medium, the non-entity association does not exist in the form of a physical entity, nor falls within the category of the aforementioned entity identification. Non-entity associations include, for example, electronic mailboxes, registered addresses, associated websites, and the like. As previously mentioned, an entity medium is referred to as a "strong medium" because it has a strong identification capability, as opposed to a non-entity association, which is referred to as a "weak medium" because it is not entity-dependent and has a weak identification capability.
The implementation process of the above core idea is described with an example. Fig. 3 is a schematic diagram illustrating a group identification method according to an embodiment of the present disclosure. As shown in fig. 3, first, for example, a business object (e.g., an account) of a current stock and non-entity attributes thereof are obtained, where the obtained account includes, for example, account 1, account 2, account 3, account 4, and account 5, where the obtained non-entity attributes of account 1 include, for example, "associated website top page photo: ajpg "," Email address: mycom03@ vp1.com ", the non-entity attribute of account 2 includes, for example," associated website home photo: n1.jpg "," company name: jiangsu la company ", the non-entity attributes of account 3 include, for example," photo of home page of associated website: ajpg "," company name: xx La general company "; non-entity attributes of account 4 include, for example, "associated website homepage photo: ajpg "," Email address: mycom01@ vp1. com; non-entity attributes of account 5 include, for example, "associated website homepage photo: j1.jpg "," Email address: mycom02@ vp1.com "," company name: xx Lala company ". Note that for convenience of description, in the attribute values of the photos of the top page of the associated website, different photo files are identified by photo file names to distinguish the different photo files, and in different embodiments, the photo files may also be identified by using, for example, codes of the photo files or specific coding processing results.
Then, it may be determined that accounts 1, 3, 4 have a first virtual relationship therebetween based on the first page photos "a 1. jpg" that all have the same associated website. In various embodiments, it may also be determined whether there is a virtual relationship between accounts, for example, based on the proximity of the photos of the top page of the associated website. In different embodiments, the determination of the photo similarity may be based on different algorithms, which is not limited in this specification.
As shown in fig. 3, it may also be determined that there is a second virtual relationship between accounts 2, 3, 5 based on them having similar company names. And determining that the accounts 1, 4 and 5 have a third virtual relationship according to the fact that the accounts have similar mail addresses. Since the commonly acquired account numbers are not identical in, for example, company name or mail address, for identifying their similarity, text pattern matching may be performed on their texts, and then the similarity of the non-entity attribute, for example, the company name or mail address, may be determined according to a combination of one or more of different parts of the texts obtained after the text pattern matching.
Then, a plurality of corresponding virtual relationship sub-graphs can be respectively constructed according to the accounts 1, 3 and 4 and the first virtual relationship between the accounts, the accounts 2, 3 and 5 and the second virtual relationship between the accounts, the accounts 1, 4 and 5 and the third virtual relationship between the accounts. For example, the first, second and third relationship graphs in fig. 3, where nodes correspond to business objects and edges between nodes correspond to virtual relationships. In different embodiments, the constructed virtual relationship subgraph may be determined according to a specific virtual relationship, for example, at least one relationship subgraph may be established according to at least one virtual relationship. And then, combining the first relational graph, the second relational graph and the third relational graph to obtain a comprehensive virtual relational graph. In different embodiments, for example, the first, second, and third virtual relationships of the accounts 1 to 5 may be combined to determine the comprehensive virtual relationship therebetween, and then the comprehensive virtual relationship graph may be established according to the comprehensive virtual relationship between the accounts 1 to 5.
And finally, excavating a node set in the comprehensive virtual relationship graph according to the obtained comprehensive virtual relationship graph through a community mining algorithm such as louvain and the like, and further obtaining a business object group corresponding to the node set, namely a high-risk business object group. In different embodiments, after the business object group is mined, the business objects can be screened by combining the entity medium relationship among the business objects and/or the pre-acquired negative business object library, so that the identification accuracy of the finally acquired risk object group is further improved.
The method has the following advantages: firstly, a non-entity medium, namely a virtual medium, of a business object is fully utilized to discover the association relationship between the objects; secondly, after the names, the mailboxes, the websites and the like of the original non-entities are subjected to word segmentation, hot word screening, recoding, word reconstruction and other processing, the non-entity attribute texts which cannot be used for business object association originally are converted into association word combinations or coding combinations which can be used for business object association, and composition media (relationship edges) of the object relationship graph are expanded from the entity media to the virtual media. And the group mining based on the object virtual relationship diagram constructed according to the virtual media greatly increases the mining coverage rate of illegal groups in batch production compared with the method for group mining based on the entity media such as equipment, bank cards, legal persons and the like. Third, with respect to a funding relationship-based group mining method, since for an account that is newly registered, for example, even if it has no funding transaction yet, its registration information typically includes one or more of information such as name, address, mailbox, etc. Therefore, the method carries out group mining based on the non-entity attributes, and the stability and the accuracy of group identification are higher. In general, the solution of the embodiment of the present specification overcomes the problems of low coverage and low recognition efficiency of group mining by constructing a virtual relationship graph based on a virtual medium.
The details of the process are further set forth below. Fig. 6 shows a flow diagram of a group partner identification method according to an embodiment of the present description. As shown in fig. 6, the method at least comprises the following steps:
step 62, determining a virtual relationship among the plurality of business objects according to a plurality of non-entity attributes, wherein the representation of the virtual relationship takes a plurality of non-entity associations as media for association;
and step 64, excavating an object group consisting of a plurality of business objects based on the object virtual relationship graph.
First, in step 61, a plurality of business objects and associated non-entity attributes are obtained. In various embodiments, the business object may be an account, customer, participating user involved in the network business activity, or other identification corresponding to a business participant. The non-entity attribute may correspond to a non-entity association that the business object has, which may be any possessory object of the business object that does not exist in a physical entity form or modality. For example, the name of the account, the registered address of the company, the email address, the associated website of the account, etc. In various embodiments, the non-entity attribute may be, for example, the non-entity association itself of the business object, e.g., the non-entity association itself, such as an account name, a registration address, an associated website of the account, etc. It may also be, for example, a component of a non-physical associate or an identification of the physical associate. For example, a home page title or a web page picture (e.g., a picture of a particular location) of an account's associated website. The specification does not limit the specific types of the entity associations and the non-entity attributes corresponding to the entity associations.
Then, at step 62, a virtual relationship between the plurality of business objects is determined based on the number of non-entity attributes.
In this step, the virtual relationship represents the association between the business objects that takes the non-entity association as a medium. For example, in one example, where multiple accounts have similar company names, a first virtual relationship between the accounts that characterizes the similarity of the company names may be determined using the similar company names as a medium. In another example, where multiple accounts have similar email addresses, a second virtual relationship between the accounts that characterizes the similarity of the email addresses may be determined using the similar company names as the medium. In different embodiments, different virtual relationships may be determined according to different non-entity attributes, which is not limited in this specification.
In an actual production scenario, usually the acquired non-entity attributes of different business objects are not exactly the same. For example, names, email addresses, etc. of different accounts are not always identical, but names, email addresses, registration addresses of companies, etc. of accounts, for example, which are registered in large quantities by illegal industries, often have specific identical parts, composition patterns, or laws. Therefore, after specific data processing is performed on the non-entity attributes, whether the non-entity attributes of different business objects have specific same parts, same composition modes or rules can be determined based on the data processing result, and whether corresponding virtual relationships exist among the non-entity attributes or the non-entity attributes can be further determined.
FIG. 4 illustrates an example diagram of a text schema and its corresponding word encodings in accordance with an embodiment of the present description. As shown in fig. 4, the email address and the company name generally have their specific text patterns or text formats, respectively, for example, the text pattern of the email address is generally "user identification + @ domain name", and the text pattern of the company name is generally "region + company name + company type", so that they may be divided into different parts according to their text patterns, and then the similarity of the non-entity attributes is determined according to the similarity or coincidence degree of the different parts. For comparison, the different parts may also be encoded, for example, different user identifiers and domain names in the email address are encoded separately in fig. 4, and then the similarity of the email addresses may be determined according to whether the encoding of the user identifiers and/or the domain names is the same/similar.
In different embodiments, the text of the non-entity attribute may be segmented based on a predetermined segmentation algorithm or model, for example, the text may be segmented by an n-gram algorithm, and then, for example, the hotword in the segmentation result may be determined by calculating a high-risk word in the segmentation result (for example, the high-risk word may be determined according to a bad sample rate of each word in the segmentation result, the bad sample rate may be determined according to a business object with a pre-labeled high-risk label, which contains the word, in the non-entity attribute thereof and the number of all business objects containing the word in the non-entity attribute thereof) and/or a high-frequency word (for example, the hotword may be determined by a TF-IDF [ term frequency-inverse document frequency ] algorithm according to the frequency of each word in the segmentation result occurring in the high-frequency text of all non-entity attributes), and then determining whether different business objects have corresponding virtual relationship according to whether the non-entity attributes of the different business objects contain the same hot words or hot word combinations.
Fig. 5 illustrates an exemplary diagram of hotwords and hotword combinations according to an embodiment of the present disclosure, and as shown in fig. 5, various hotwords such as "guangzhou", "guangzhou south sand area", "memory bitter", and combinations thereof, for example, may be extracted from the company name "guangzhou south sand area memory bitter network technology limited" of an account for determination of virtual relationships with other accounts based on the company name. For example, "fujian province", "dragon city", "virtual industry park", and combinations thereof, may also be extracted from the registered address of an account for determination of virtual relationships with other accounts based on the registered address. In one example, the hotword may also be encoded, for example, the hotword 'network technologies company' is encoded as 'composition 01' for determining the virtual relationship.
Thus, in one embodiment, assume that the number of non-entity attributes includes a first attribute characterized in textual form; then, text word segmentation can be performed on the texts with the first attributes of the plurality of business objects to obtain word segmentation results of the texts; and determining a first virtual relationship corresponding to the first attribute among the plurality of business objects based on the word segmentation result.
More specifically, in one embodiment, the segmentation result may include a plurality of segmented words obtained by text segmentation. Thus, each divisional word of the text may be encoded; and determining whether a first virtual relation exists between the plurality of business objects based on codes corresponding to one or more dividing words of the text. In another embodiment, the text may be pattern matched; and if the text conforms to a preset text mode, performing text word segmentation on the text according to the text mode to obtain a word segmentation result.
In one embodiment, it may be determined whether a first virtual relationship exists between the plurality of service objects based on whether the segmentation results corresponding to the first attributes of the plurality of service objects include the same segmentation word/segmentation word combination/pattern word combination, where the segmentation word is obtained based on text segmentation, and the pattern word is obtained based on text segmentation according to a text pattern.
In one embodiment, it may be determined whether a first virtual relationship exists between the plurality of business objects based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects include the same hotword combination, where the hotword combination includes at least one hotword.
In one embodiment, the hotword may be determined according to the frequency of occurrence of each divided word in the word segmentation result in the text of several non-entity attributes of the plurality of business objects. In another embodiment, the hot word is determined according to a bad sample rate of each division word in the division word result, and the bad sample rate is determined according to a ratio of the business object corresponding to the division word and having the pre-labeled high-risk label to all the business objects corresponding to the division word.
In various embodiments, various virtual relationships may be determined based on various non-entity attributes, and thus, in one embodiment, several non-entity attributes may include, for example, a first attribute and a second attribute. A first virtual relationship between the plurality of business objects can be determined according to the first attribute; and determining a second virtual relationship among the plurality of business objects according to the second attribute.
And determining various virtual relationships according to various non-entity attributes respectively, and then combining the various virtual relationships to obtain a comprehensive virtual relationship. Thus, in one embodiment, the number of non-entity attributes may include a first attribute and a second attribute; a first virtual relationship between the plurality of business objects can be determined according to the first attribute; determining a second virtual relationship among the plurality of business objects according to the second attribute; determining a synthetic virtual relationship based at least on the first virtual relationship and the second virtual relationship.
Next, at step 63, an object virtual relationship graph is generated based on the business objects and virtual relationships.
In this step, according to the business object and the virtual relationship, the generated object virtual relationship graph includes a plurality of nodes and edges between the nodes, the nodes correspond to the business object, and the edges correspond to the virtual relationship.
In the embodiment of generating the comprehensive virtual relationship, the object virtual relationship graph may be generated according to the business object and the comprehensive virtual relationship between the business objects.
In the above embodiment of generating multiple virtual relationships, multiple corresponding relationship subgraphs may be generated according to multiple virtual relationships, and then a complete object virtual relationship graph is obtained by combining the multiple relationship subgraphs. In one example, a first relationship graph may be generated based on the business object and the first virtual relationship; generating a second relation graph according to the business object and the second virtual relation; and generating the object virtual relationship diagram based on the first relationship diagram and the second relationship diagram.
Thereafter, in step 64, an object group consisting of a plurality of business objects is mined based on the object virtual relationship graph obtained in step 63.
In this step, a plurality of node groups including a plurality of nodes may be obtained based on a community mining algorithm according to the object virtual relationship graph, and an object group corresponding to the node group is obtained. In one embodiment, mined object groups may be determined to be high risk groups, and governing measures such as freezing accounts, controlling business behavior, etc. may be applied to the business objects that they contain. In different embodiments, the group mining may be performed according to different community mining algorithms, which is not limited in this specification. In one embodiment, the community mining algorithm may include one of a Louvain algorithm, a tag propagation algorithm, and an Infomap algorithm.
In an actual production scenario, high-risk groups mined based on the object virtual relationship graph have difficulty reaching full accuracy. For example, the names of a plurality of accounts registered by illegal industries have a pattern "mycomponyxxxx", wherein "X" is an arbitrary number, and after identifying the pattern and establishing corresponding virtual relations and further constructing an object virtual relation graph, high-risk groups containing the accounts registered by illegal industries are mined. However, an account name "mycompany 0001" with which a regular enterprise is registered is also identified into the high-risk group. In order to improve the identification accuracy of the risk group, in one embodiment, the object group and/or the business objects in the object group may be filtered based on an object entity relationship diagram obtained according to entity attributes of a plurality of business objects, so as to obtain a filtering result, where the entity attributes correspond to entity associations that the business objects have; and taking the screening result as an output result of the group recognition. In another embodiment, based on a pre-obtained negative business object library, the business objects in the object group and/or the object group are screened to obtain a screening result; and taking the screening result as an output result of the group recognition.
According to an embodiment of another aspect, a group partner identifying device is also provided. Fig. 7 is a block diagram illustrating a group recognition apparatus according to an embodiment of the present disclosure. As shown in fig. 7, the apparatus 700 includes:
an object non-entity attribute obtaining unit 71, configured to obtain a plurality of business objects and a plurality of non-entity attributes associated therewith, where the plurality of non-entity attributes respectively correspond to a plurality of non-entity associated objects possessed by the business objects;
a virtual relationship determining unit configured to 72 determine a virtual relationship between the plurality of business objects according to the plurality of non-entity attributes, wherein the virtual relationship represents that the plurality of non-entity associations are associated with each other by taking the plurality of non-entity associations as a medium;
a virtual relationship graph generating unit 73 configured to generate an object virtual relationship graph according to the service object and the virtual relationship, where the object virtual relationship graph includes a plurality of nodes and edges between the nodes, where the nodes correspond to the service object, and the edges correspond to the virtual relationship;
a group mining unit 74 configured to mine an object group consisting of a number of the business objects based on the object virtual relationship graph.
In one embodiment, the number of non-entity attributes may include a first attribute and a second attribute;
a virtual relationship determination unit, which may be further configured to:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
the virtual relationship diagram generation unit may be further configured to:
generating a first relation graph according to the business object and the first virtual relation;
generating a second relation graph according to the business object and the second virtual relation;
and generating the object virtual relationship diagram based on the first relationship diagram and the second relationship diagram.
In one embodiment, the number of non-entity attributes may include a first attribute and a second attribute;
a virtual relationship determination unit, which may be further configured to:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
determining a synthetic virtual relationship based at least on the first virtual relationship and the second virtual relationship;
the virtual relationship diagram generation unit may be further configured to:
and generating the object virtual relationship graph according to the service object and the comprehensive virtual relationship.
In one embodiment, the number of non-entity attributes may include a first attribute,
a virtual relationship determination unit, which may be further configured to:
performing text word segmentation on the texts with the first attributes of the plurality of service objects to obtain word segmentation results of the texts;
and determining a first virtual relationship corresponding to the first attribute among the plurality of business objects based on the word segmentation result.
In one embodiment, the virtual relationship determining unit may be further configured to:
performing pattern matching on the text;
and if the text conforms to a preset text mode, performing text word segmentation on the text according to the text mode to obtain a word segmentation result.
In one embodiment, the virtual relationship determining unit may be further configured to:
determining whether a first virtual relationship exists among the plurality of business objects based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same word segmentation/word segmentation combination/mode word combination, wherein the word segmentation is obtained based on text word segmentation, and the mode word is obtained based on text word segmentation according to a text mode.
In one embodiment, the virtual relationship determining unit may be further configured to:
determining whether a first virtual relationship exists among the plurality of business objects or not based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same hot word combination or not, wherein the hot word combination comprises at least one hot word.
Yet another aspect of the present specification provides a computer readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform any of the methods described above.
Yet another aspect of the present specification provides a computing device comprising a memory having stored therein executable code, and a processor that, when executing the executable code, implements any of the methods described above.
It is to be understood that the terms "first," "second," and the like, herein are used for descriptive purposes only and not for purposes of limitation, to distinguish between similar concepts.
Those skilled in the art will recognize that, in one or more of the examples described above, the functions described in this invention may be implemented in hardware, software, firmware, or any combination thereof. When implemented in software, the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium.
The above-mentioned embodiments, objects, technical solutions and advantages of the present invention are further described in detail, it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made on the basis of the technical solutions of the present invention should be included in the scope of the present invention.
Claims (24)
1. A group partner identification method, comprising:
acquiring a plurality of business objects and a plurality of associated non-entity attributes thereof, wherein the non-entity attributes respectively correspond to a plurality of non-entity associated objects of the business objects;
determining a virtual relationship among the plurality of business objects according to the plurality of non-entity attributes, wherein the virtual relationship represents that the plurality of non-entity associations are associated by taking the plurality of non-entity associations as media;
generating an object virtual relationship graph according to the business object and the virtual relationship, wherein the object virtual relationship graph comprises a plurality of nodes and edges among the nodes, the nodes correspond to the business object, and the edges correspond to the virtual relationship;
and mining an object group consisting of a plurality of business objects based on the object virtual relationship graph.
2. The method of claim 1, wherein the number of non-entity attributes comprises: and one or more of the name, the registration address, the email address and the associated website of the business object.
3. The method of claim 1, wherein the number of non-entity attributes includes a first attribute and a second attribute;
determining a virtual relationship between the plurality of business objects according to the plurality of non-entity attributes, including:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
the generating of the object virtual relationship graph comprises:
generating a first relation graph according to the business object and the first virtual relation;
generating a second relation graph according to the business object and the second virtual relation;
and generating the object virtual relationship diagram based on the first relationship diagram and the second relationship diagram.
4. The method of claim 1, wherein the number of non-entity attributes includes a first attribute and a second attribute;
determining a virtual relationship between the plurality of business objects according to the plurality of non-entity attributes, including:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
determining a synthetic virtual relationship based at least on the first virtual relationship and the second virtual relationship;
the generating of the object virtual relationship graph comprises:
and generating the object virtual relationship graph according to the service object and the comprehensive virtual relationship.
5. The method of claim 1, wherein the number of non-entity attributes includes a first attribute, and determining a virtual relationship between the plurality of objects based on the number of non-entity attributes includes:
performing text word segmentation on the texts with the first attributes of the plurality of service objects to obtain word segmentation results of the texts;
and determining a first virtual relationship corresponding to the first attribute among the plurality of business objects based on the word segmentation result.
6. The method of claim 5, wherein the word segmentation result comprises a plurality of segmented words obtained by text word segmentation;
determining a first virtual relationship corresponding to the first attribute among the plurality of business objects based on the word segmentation result, including:
encoding each division word of the text;
and determining whether a first virtual relation exists between the plurality of business objects based on codes corresponding to one or more dividing words of the text.
7. The method of claim 5, wherein text-tokenizing text of the first attribute of a plurality of business objects comprises:
performing pattern matching on the text;
and if the text conforms to a preset text mode, performing text word segmentation on the text according to the text mode to obtain a word segmentation result.
8. The method of claim 5, wherein determining, based on the word segmentation result, a first virtual relationship between a plurality of business objects corresponding to the first attribute comprises:
determining whether a first virtual relationship exists among the plurality of business objects based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same word segmentation/word segmentation combination/mode word combination, wherein the word segmentation is obtained based on text word segmentation, and the mode word is obtained based on text word segmentation according to a text mode.
9. The method of claim 5, wherein determining, based on the word segmentation result, a first virtual relationship between a plurality of business objects corresponding to the first attribute comprises:
determining whether a first virtual relationship exists among the plurality of business objects or not based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same hot word combination or not, wherein the hot word combination comprises at least one hot word.
10. The method of claim 9, wherein the hotword is determined according to a frequency of occurrence of each segmented word in the segmentation result in text of non-entity attributes of the plurality of business objects.
11. The method according to claim 9, wherein the hotword is determined according to a bad sample rate of each divided word in the word segmentation result, and the bad sample rate is determined according to a ratio of a business object corresponding to the divided word and having a pre-labeled high-risk label to all business objects corresponding to the divided word.
12. The method of claim 1, wherein mining an object group consisting of a number of the business objects based on the object virtual relationship graph comprises:
and acquiring a plurality of node groups comprising a plurality of nodes based on a community mining algorithm according to the object virtual relationship graph, and acquiring object groups corresponding to the node groups.
13. The method of claim 12, wherein the community mining algorithm comprises one of a Louva I n algorithm, a label propagation algorithm, an I nfomap algorithm.
14. The method of claim 1, further comprising:
based on an object entity relationship graph obtained according to entity attributes of a plurality of service objects, screening the object group and/or the service objects in the object group to obtain a screening result, wherein the entity attributes correspond to entity associations of the service objects;
and taking the screening result as an output result of the group recognition.
15. The method of claim 1, further comprising:
based on a pre-acquired negative business object library, screening the object group and/or the business objects in the object group to obtain a screening result;
and taking the screening result as an output result of the group recognition.
16. A group partner identifying device, comprising:
an object non-entity attribute obtaining unit configured to obtain a plurality of business objects and a plurality of non-entity attributes associated therewith, the plurality of non-entity attributes corresponding to a plurality of non-entity associations possessed by the business objects, respectively;
the virtual relation determining unit is configured to determine a virtual relation among the plurality of business objects according to the plurality of non-entity attributes, wherein the virtual relation represents that the plurality of non-entity associations are associated by taking the plurality of non-entity associations as media;
a virtual relationship graph generating unit configured to generate an object virtual relationship graph according to the service object and the virtual relationship, where the object virtual relationship graph includes a plurality of nodes and edges between the nodes, where the nodes correspond to the service object and the edges correspond to the virtual relationship;
and the group-partner digging unit is configured to dig out an object group formed by a plurality of business objects based on the object virtual relationship graph.
17. The apparatus of claim 16, wherein the number of non-entity attributes includes a first attribute and a second attribute;
a virtual relationship determination unit further configured to:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
a virtual relationship diagram generation unit further configured to:
generating a first relation graph according to the business object and the first virtual relation;
generating a second relation graph according to the business object and the second virtual relation;
and generating the object virtual relationship diagram based on the first relationship diagram and the second relationship diagram.
18. The apparatus of claim 16, wherein the number of non-entity attributes includes a first attribute and a second attribute;
a virtual relationship determination unit further configured to:
determining a first virtual relationship among the plurality of business objects according to the first attribute;
determining a second virtual relationship among the plurality of business objects according to the second attribute;
determining a synthetic virtual relationship based at least on the first virtual relationship and the second virtual relationship;
a virtual relationship diagram generation unit further configured to:
and generating the object virtual relationship graph according to the service object and the comprehensive virtual relationship.
19. The apparatus of claim 16, wherein the number of non-entity attributes includes a first attribute,
a virtual relationship determination unit further configured to:
performing text word segmentation on the texts with the first attributes of the plurality of service objects to obtain word segmentation results of the texts;
and determining a first virtual relationship corresponding to the first attribute among the plurality of business objects based on the word segmentation result.
20. The apparatus of claim 19, wherein the virtual relationship determining unit is further configured to:
performing pattern matching on the text;
and if the text conforms to a preset text mode, performing text word segmentation on the text according to the text mode to obtain a word segmentation result.
21. The apparatus of claim 19, wherein the virtual relationship determining unit is further configured to:
determining whether a first virtual relationship exists among the plurality of business objects based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same word segmentation/word segmentation combination/mode word combination, wherein the word segmentation is obtained based on text word segmentation, and the mode word is obtained based on text word segmentation according to a text mode.
22. The apparatus of claim 19, wherein the virtual relationship determining unit is further configured to:
determining whether a first virtual relationship exists among the plurality of business objects or not based on whether the word segmentation results corresponding to the first attributes of the plurality of business objects comprise the same hot word combination or not, wherein the hot word combination comprises at least one hot word.
23. A computer-readable storage medium, having stored thereon a computer program which, when executed in a computer, causes the computer to perform the method of any of claims 1-15.
24. A computing device comprising a memory having executable code stored therein and a processor that, when executing the executable code, implements the method of any of claims 1-15.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111258230.4A CN113988870A (en) | 2021-10-27 | 2021-10-27 | Group partner identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111258230.4A CN113988870A (en) | 2021-10-27 | 2021-10-27 | Group partner identification method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113988870A true CN113988870A (en) | 2022-01-28 |
Family
ID=79742853
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111258230.4A Pending CN113988870A (en) | 2021-10-27 | 2021-10-27 | Group partner identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113988870A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7805457B1 (en) * | 2008-02-14 | 2010-09-28 | Securus Technologies, Inc. | System and method for identifying members of a gang or security threat group |
| CN103257957A (en) * | 2012-02-15 | 2013-08-21 | 深圳市腾讯计算机系统有限公司 | Chinese word segmentation based text similarity identifying method and device |
| US20140136547A1 (en) * | 2012-11-14 | 2014-05-15 | International Business Machines Corporation | Determining Potential Enterprise Partnerships |
| CN106301978A (en) * | 2015-05-26 | 2017-01-04 | 阿里巴巴集团控股有限公司 | The recognition methods of gang member account, device and equipment |
| CN106921504A (en) * | 2015-12-24 | 2017-07-04 | 阿里巴巴集团控股有限公司 | A kind of method and apparatus of the associated path for determining different user |
| US20190340294A1 (en) * | 2018-05-04 | 2019-11-07 | International Business Machines Corporation | Combining semantic relationship information with entities and non-entities for predictive analytics in a cognitive system |
| CN112910888A (en) * | 2021-01-29 | 2021-06-04 | 杭州迪普科技股份有限公司 | Illegal domain name registration group mining method and device |
| CN113536070A (en) * | 2021-08-11 | 2021-10-22 | 汉唐信通(北京)咨询股份有限公司 | Address resolution method, system, computer equipment and storage medium |
-
2021
- 2021-10-27 CN CN202111258230.4A patent/CN113988870A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7805457B1 (en) * | 2008-02-14 | 2010-09-28 | Securus Technologies, Inc. | System and method for identifying members of a gang or security threat group |
| CN103257957A (en) * | 2012-02-15 | 2013-08-21 | 深圳市腾讯计算机系统有限公司 | Chinese word segmentation based text similarity identifying method and device |
| US20140136547A1 (en) * | 2012-11-14 | 2014-05-15 | International Business Machines Corporation | Determining Potential Enterprise Partnerships |
| CN106301978A (en) * | 2015-05-26 | 2017-01-04 | 阿里巴巴集团控股有限公司 | The recognition methods of gang member account, device and equipment |
| CN106921504A (en) * | 2015-12-24 | 2017-07-04 | 阿里巴巴集团控股有限公司 | A kind of method and apparatus of the associated path for determining different user |
| US20190340294A1 (en) * | 2018-05-04 | 2019-11-07 | International Business Machines Corporation | Combining semantic relationship information with entities and non-entities for predictive analytics in a cognitive system |
| CN112910888A (en) * | 2021-01-29 | 2021-06-04 | 杭州迪普科技股份有限公司 | Illegal domain name registration group mining method and device |
| CN113536070A (en) * | 2021-08-11 | 2021-10-22 | 汉唐信通(北京)咨询股份有限公司 | Address resolution method, system, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ahmed et al. | A survey of anomaly detection techniques in financial domain | |
| Dou et al. | Enhancing graph neural network-based fraud detectors against camouflaged fraudsters | |
| US11544501B2 (en) | Systems and methods for training a data classification model | |
| JP5941163B2 (en) | Spam detection system and method using frequency spectrum of character string | |
| KR20180115778A (en) | Integrated block chain-based data transfer control method and system | |
| CN111046237B (en) | User behavior data processing method and device, electronic equipment and readable medium | |
| KR20190025005A (en) | Method and device for controlling data risk | |
| CN111932130B (en) | Service type identification method and device | |
| CN111367965B (en) | Target object determining method, device, electronic equipment and storage medium | |
| CN112435137B (en) | Cheating information detection method and system based on community mining | |
| WO2020048056A1 (en) | Risk decision method and apparatus | |
| CN111340612B (en) | Account risk identification method and device and electronic equipment | |
| US11687836B2 (en) | System and method for automatic labeling of clusters created by machine learning methods | |
| US20130332374A1 (en) | Fraud prevention for real estate transactions | |
| EP3361704A1 (en) | User data sharing method and device | |
| JP2022548501A (en) | Data acquisition method and device for analyzing cryptocurrency transactions | |
| CN110197426A (en) | A kind of method for building up of credit scoring model, device and readable storage medium storing program for executing | |
| CN108492119A (en) | Information anti-fake authentication method, device, equipment and medium based on block chain | |
| CN105988998B (en) | Relational network construction method and device | |
| CN112750038B (en) | Transaction risk determination method, device and server | |
| CN111245815B (en) | Data processing method and device, storage medium and electronic equipment | |
| CN110363534B (en) | Method and device for identifying abnormal transaction | |
| CN113988870A (en) | Group partner identification method and device | |
| CN113806555B (en) | Operation abnormality identification method, system and device for APP and storage medium | |
| US20210357942A1 (en) | Method and apparatus for identifying risky vertices |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |