CN110597984A - Method and device for determining abnormal behavior user information, storage medium and terminal - Google Patents
Method and device for determining abnormal behavior user information, storage medium and terminal Download PDFInfo
- Publication number
- CN110597984A CN110597984A CN201910740571.1A CN201910740571A CN110597984A CN 110597984 A CN110597984 A CN 110597984A CN 201910740571 A CN201910740571 A CN 201910740571A CN 110597984 A CN110597984 A CN 110597984A
- Authority
- CN
- China
- Prior art keywords
- behavior
- text data
- classification
- abnormal behavior
- marking
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/335—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
Abstract
The invention discloses a method and a device for determining user information with abnormal behaviors, a storage medium and a terminal, relates to the technical field of data processing, and mainly aims to solve the problems that the conditions that the same user performs the information loss behaviors among different enterprises cannot be avoided only by judging the information loss times of each enterprise, so that the information loss behaviors are screened under single conditions, and a large amount of human resources are consumed. The method comprises the following steps: acquiring all behavior text data in the user service transaction process; classifying and marking the behavior text data according to a classification system, wherein the classification system is used for dividing category hierarchies of different business behaviors in the behavior text data; and performing abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model to determine the user information of the abnormal behavior.
Description
Technical Field
The present invention relates to the field of data processing technologies, and in particular, to a method and an apparatus for determining abnormal behavior user information, a storage medium, and a terminal.
Background
With the rapid development of electronic commerce, users using the internet for financial activities have become a hot business activity. The credit loss behavior of the user can be generated when the bank, the e-commerce and the internet financial enterprise carry out transaction, so that in order to avoid economic loss caused by the credit loss behavior, each enterprise determines the credit loss blacklist of the user according to the transaction data owned by the enterprise, and the enterprise can process the user.
At present, users in the existing credit loss blacklist are determined by manually screening through judging whether the credit loss times of the users in each enterprise exceed specific times, but the condition that the same user performs credit loss behaviors among different enterprises cannot be avoided through judging the credit loss times in each enterprise, so that the credit loss behavior screening condition is single, a large amount of human resources are consumed, and the determination efficiency of users with abnormal behaviors is reduced.
Disclosure of Invention
In view of the above, the present invention provides a method and an apparatus for determining user information with abnormal behavior, a storage medium, a terminal method and an apparatus, and mainly aims to solve the problems that the conditions of the same user performing a trust loss behavior between different enterprises cannot be avoided only by determining the trust loss times in each enterprise, so that the trust loss behavior screening conditions are single, and a large amount of human resources are consumed.
According to an aspect of the present invention, there is provided a method for determining abnormal behavior user information, including:
acquiring all behavior text data in the user service transaction process;
classifying and marking the behavior text data according to a classification system, wherein the classification system is used for dividing category hierarchies of different business behaviors in the behavior text data;
and performing abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model to determine the user information of the abnormal behavior.
Further, the performing abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model to determine the abnormal behavior user information includes:
performing numerical processing on the behavior text data obtained after the classification marking;
and performing prediction operation on the digitized behavior text data by using a preset abnormal behavior language processing model and a preset natural language processing algorithm to obtain abnormal behavior user information.
Further, the performing numerical processing on the behavior text data obtained after the classification marking includes:
and performing serialization sequencing on the behavior text data after the classification marking, and generating a behavior data sequence according to the sequenced behavior text data.
Further, the step of performing prediction operation on the digitized behavior text data by using a preset abnormal behavior language processing model and a preset natural language processing algorithm to obtain the abnormal behavior user information includes:
and processing the behavior data sequence as the input of a preset abnormal behavior language processing model to obtain a sequence set of different labels, performing prediction operation on the sequence set of different labels according to a preset natural language processing algorithm to determine abnormal behavior user information, wherein a characteristic function in the preset natural language processing algorithm is a preset abnormal behavior matching rule.
Further, the classifying and marking the behavior text data according to a classification system includes:
and classifying the behavior text data according to different category hierarchies by using a natural language processing mode, and labeling tag labels on the classified search behavior text data.
Further, the classifying and marking the behavior text data according to a classification system includes:
and determining a category hierarchy and a marking set for classifying and marking the behavior text data according to a classification system in a mode of comparing classification accuracy with marking accuracy.
Further, the determining of the category hierarchy and the label set for classifying and labeling the behavior text data according to the classification system by comparing the classification accuracy with the label accuracy includes:
classifying the behavior text data according to different category hierarchies by using a natural language processing mode, and marking the classified behavior text data by using tag labels;
and respectively calculating the classification accuracy and the marking accuracy of each time, and determining the category level and the marking set of the behavior text data for classification marking according to the product of the classification accuracy and the marking accuracy.
Further, the method further comprises:
and training an abnormal behavior prediction model according to behavior text data and abnormal behavior user information in a preset behavior database, wherein the marked behavior text data and a corresponding relation matched with the abnormal behavior user information are stored in the preset behavior database.
According to another aspect of the present invention, there is provided an apparatus for determining abnormal behavior user information, including:
the acquisition module is used for acquiring all behavior text data in the user service transaction process;
the classification module is used for classifying and marking the behavior text data according to a classification system, and the classification system is used for dividing the category hierarchy of different business behaviors in the behavior text data;
and the determining module is used for performing abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model to determine the user information of the abnormal behavior.
Further, the determining module includes:
the processing unit is used for carrying out numerical processing on the behavior text data obtained after the classification marking;
and the operation unit is used for carrying out prediction operation on the digitized behavior text data by utilizing a preset abnormal behavior language processing model and combining a preset natural language processing algorithm to obtain abnormal behavior user information.
Further, the processing unit is specifically configured to perform serialization and ordering on the behavior text data after the classification is marked, and generate a behavior data sequence according to the ordered behavior text data.
Further, the operation unit is specifically configured to process the behavior data sequence as an input of a preset abnormal behavior language processing model to obtain a sequence set of different tags, perform prediction operation on the sequence set of different tags according to a preset natural language processing algorithm, and determine abnormal behavior user information, where a feature function in the preset natural language processing algorithm is a preset abnormal behavior matching rule.
Further, the classification module is specifically configured to classify the behavior text data according to different category hierarchies by using a natural language processing manner, and perform tag labeling on the classified search behavior text data.
Further, the classification module is specifically configured to determine a category hierarchy and a label set for performing classification labeling on the behavior text data according to a classification system in a classification accuracy and label accuracy comparison manner.
Further, the classification module includes:
the classification unit is used for classifying the behavior text data according to different category hierarchies by using a natural language processing mode and marking the classified behavior text data by using tag labels;
and the determining unit is used for respectively calculating the classification accuracy and the marking accuracy of each time, and determining the category level and the marking set of the behavior text data for classification marking according to the product of the classification accuracy and the marking accuracy.
Further, the apparatus further comprises:
and the training module is used for training an abnormal behavior prediction model according to behavior text data and abnormal behavior user information in a preset behavior database, wherein the marked behavior text data and the corresponding relation matched with the abnormal behavior user information are stored in the preset behavior database.
According to another aspect of the present invention, there is provided a storage medium, wherein at least one executable instruction is stored in the storage medium, and the executable instruction causes a processor to perform an operation corresponding to the determination method of the abnormal behavior user information.
According to still another aspect of the present invention, there is provided a terminal including: the system comprises a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete mutual communication through the communication bus;
the memory is used for storing at least one executable instruction, and the executable instruction enables the processor to execute the operation corresponding to the determination method of the abnormal behavior user information.
By the technical scheme, the technical scheme provided by the embodiment of the invention at least has the following advantages:
compared with the technology that users in the existing credit loss blacklist are manually screened and determined by judging whether the credit loss times of the users in each enterprise exceed specific times, the method and the device for determining the abnormal behavior user information have the advantages that all behavior text data of the users are obtained, classification marking is carried out on the behavior text data according to a classification system, the trained abnormal behavior prediction model is used for carrying out abnormal behavior prediction operation on the classified and marked behavior text data, the abnormal behavior user information is determined, accurate judgment on the abnormal behavior users is achieved, human resource consumption is reduced, screening conditions of the abnormal behavior users are adjusted in a self-adaptive mode, and accordingly determining efficiency and accuracy of the abnormal behavior users are improved.
The foregoing description is only an overview of the technical solutions of the present invention, and the embodiments of the present invention are described below in order to make the technical means of the present invention more clearly understood and to make the above and other objects, features, and advantages of the present invention more clearly understandable.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
fig. 1 is a flowchart illustrating a method for determining abnormal behavior user information according to an embodiment of the present invention;
fig. 2 is a flowchart illustrating another method for determining abnormal behavior user information according to an embodiment of the present invention;
fig. 3 is a block diagram illustrating a device for determining user information with abnormal behavior according to an embodiment of the present invention;
fig. 4 is a block diagram illustrating another apparatus for determining abnormal behavior user information according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a terminal according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
An embodiment of the present invention provides a method for determining user information with abnormal behavior, as shown in fig. 1, the method includes:
101. and acquiring all behavior text data in the user service transaction process.
In the embodiment of the invention, when a user conducts electronic commerce transactions such as finance, commerce and the like, in order to ensure that the user has the ability to pay for the transactions or the ability to repay paid loan amount, the behavior text data of the user is acquired, so that the abnormal behavior user information, namely the user information belonging to the loss of credit, such as a loss of credit blacklist, is determined according to the behavior text data. The behavior text data is specific behavior text data of business transaction performed on the user by different transaction parties, for example, the bank transaction party records specific behavior text data "your No. 1 loan is cleared by a bank card number xxx today" of business transaction performed by the user, or the bank transaction party records specific behavior data "your card number is entered into account" of business transaction performed by the user.
It should be noted that, in order to timely deliver the behavior generated by the user to the user, the transaction party generally sends the behavior text data to the user terminal in the form of short message, email, etc. so as to be convenient for the user to view. In the embodiment of the invention, in order to determine the abnormal behavior user information, a data interaction protocol with a transaction party can be established so as to obtain all behavior text data of the user from different transaction parties.
102. And carrying out classification marking on the behavior text data according to a classification system.
The classification system is used for dividing category hierarchies of different business behaviors in the behavior text data, the classification system can be divided according to businesses such as finance, credit card and loan, different businesses are divided into different category hierarchies, and the different businesses comprise three category hierarchies such as loan behavior/credit card/quota, loan behavior/credit card/deration, loan behavior/credit card/overdue and the like, and specific behaviors corresponding to final leaf categories such as overdue and repayment are not limited specifically.
It should be noted that, in order to determine the abnormal behavior user from the behavior text data, the behavior text data needs to be classified and labeled according to a classification system. In the embodiment of the invention, the named entity recognition NER of the natural language processing method is used for carrying out label recognition on the words belonging to the classification system and the words of other categories, the category of the label can be marked according to the name of a person, time and the category hierarchy of the classification system, for example, the behavior text data is classified and marked according to the behavior category corresponding to the third category hierarchy, and labels such as 'name of a person', 'time', 'loan', 'amount', 'overdue', and the like are obtained. In addition, after the behavior text data is classified and marked by the name entity identification NER, the obtained text data is semi-structured text data, such as bei '10-day repayment 100 yuan' for mr. liu, and { change @ borrowing organization } { 10-day @ time } { repayment @ behavior } {100 yuan @ amount } for mr. liu, which are classified and marked according to the behavior hierarchy in the loan behavior/credit card/repayment category hierarchy.
103. And performing abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model to determine the user information of the abnormal behavior.
In the embodiment of the present invention, the abnormal behavior prediction model is used to obtain abnormal behavior user information, such as a user name and a certificate number, after the abnormal behavior prediction model is operated according to the behavior text data obtained after the classification marking, and the embodiment of the present invention is not specifically limited. The abnormal behavior prediction model may include a trained language learning model Xlnet, and the behavior text data after being classified and labeled is used as the input of the Xlnet model to perform operation to obtain a text set with different set labels, and the preset abnormality is combined with a label corresponding to the user information, for example, the label of the loss of confidence behavior corresponding to the loss of confidence blacklist rule determines the user information of the abnormal behavior, that is, determines the user blacklist.
Compared with the technology that users in the existing credit loss blacklist are manually screened and determined by judging whether the credit loss times of the users in each enterprise exceed specific times, the method and the device for determining the abnormal behavior user information achieve accurate judgment of the abnormal behavior users, reduce human resource consumption, achieve self-adaptive adjustment of screening conditions of the abnormal behavior users and improve determination efficiency and accuracy of the abnormal behavior users by obtaining all behavior text data of the users, classifying and marking the behavior text data according to a classification system, and performing abnormal behavior prediction operation on the classified and marked behavior text data through a trained abnormal behavior prediction model.
An embodiment of the present invention provides another method for determining user information with abnormal behavior, as shown in fig. 2, the method includes:
201. and training an abnormal behavior prediction model according to the behavior text data in the preset behavior database and the abnormal behavior user information.
For the embodiment of the invention, in order to enable the abnormal behavior prediction model to accurately determine the abnormal behavior user information from the behavior text data, the abnormal behavior prediction model needs to be trained in advance. The marked behavior text data are classified and marked behavior text data, the corresponding relation matched with the abnormal behavior user information is an abnormal behavior relation configured according to the behavior text data, specifically, the behavior text data and the abnormal behavior user information in the preset behavior database are stored in a determined user abnormal behavior blacklist, and therefore an abnormal behavior prediction model is trained according to the behavior text data and the determined abnormal behavior user information. In addition, the prediction behavior prediction model is an Xlnet model, the training process is the same as that of the existing Xlnet model, and the embodiment of the invention is not particularly limited.
202. And acquiring all behavior text data in the user service transaction process.
This step is the same as step 101 shown in fig. 1, and is not described herein again.
203a, classifying the behavior text data according to different category hierarchies by using a natural language processing mode, and labeling tag labels on the classified search behavior text data.
For the embodiment of the invention, in order to improve the marking of the behavior text data and make the behavior text data marked according to different category levels so as to realize the prediction operation of the abnormal behavior prediction model of the behavior text data, the behavior text data is classified according to classification systems of different category levels by using a natural language processing mode, and tag labeling is carried out on the classified behavior text data so as to obtain the category levels and a mark set. The words obtained by processing the behavior text data in the natural language processing mode are classified according to different category hierarchies, for example, words "enter" are found from "your card number" enter ", the words are classified into income categories according to the category hierarchies, and then the words in the behavior text data are marked in a tag marking mode. During tag labeling, semi-structured data of the behavior text data is obtained in a named entity identification NER processing mode, such as { mr. liu @ name } your { organization of changes } {10 day @ time } { repayment @ behavior } { 100-element @ amount }, where tags of the semi-structured data of all the behavior text data of a user can be classified according to different category levels as "1 fendm # (user id) behavior tags: loan, repayment, app change loan, debt ".
For the embodiment of the present invention, 203b, which is parallel to 203a, determines the category hierarchy and the label set for classifying and labeling the behavior text data according to the classification system by comparing the classification accuracy with the label accuracy.
For the embodiment of the invention, in order to improve the marking of the behavior text data and make the behavior text data marked according to different category levels, the abnormal behavior prediction model prediction operation of the behavior text data is realized, and the category level and the marking set are determined by comparing the classification accuracy with the marking accuracy. The classification precision is the precision of the category hierarchy representing the user behavior, if the classification category hierarchy is more, the precision of the user behavior is higher, the marking precision is the precision of the text word label representing the user behavior, and if the marking label is more, the precision of the user behavior is lower. In the embodiment of the invention, the optimal classification and marking modes are selected for classification and marking in a self-adaptive classification and marking accuracy comparison mode so as to obtain the optimal category level and marking set.
For further illustration and limitation, step 203b may be: classifying the behavior text data according to different category hierarchies by using a natural language processing mode, and marking the classified behavior text data by using tag labels; and respectively calculating the classification accuracy and the marking accuracy of each time, and determining the category level and the marking set of the behavior text data for classification marking according to the product of the classification accuracy and the marking accuracy.
In order to determine the optimal category hierarchy and label, classifying the behavior text data according to all category hierarchies in a classification system by using a natural language processing mode, labeling the classified behavior text data by using tag labels to obtain a classification result and a label set of each time, and calculating the label accuracy corresponding to the classification accuracy of each classification by using the classification accuracy and the label accuracy, wherein the classification accuracy is the accuracy of embodying the user behavior by the category hierarchy, if the classification category hierarchies are more, the accuracy of embodying the user behavior is higher, the label accuracy is the accuracy of embodying the user behavior by using the text word label labels, if the label labels are more, the accuracy of embodying the user behavior is lower, and the calculation method of the accuracy is not specifically limited in the embodiment of the invention. In addition, the maximum value of the classification accuracy and the marking accuracy which are calculated in each classification and marking process is multiplied, and the corresponding category level and the marked label set are determined.
204. And carrying out numerical processing on the behavior text data obtained after the classification and marking.
For the embodiment of the present invention, since the behavior text data needs to be subjected to prediction operation, the behavior text data obtained after classification labeling, that is, the label set carrying the category hierarchy, needs to be subjected to numerical processing, that is, the text data is converted into numerical data.
For further illustration and limitation of the embodiment of the present invention, step 204 may be: and performing serialization sequencing on the behavior text data after the classification marking, and generating a behavior data sequence according to the sequenced behavior text data.
In the embodiment of the invention, different category hierarchies are randomly sequenced to obtain a group of sequence values, the sequence numbers can be used as id of tags, then the sequence values are sequenced according to time, such as loan, repayment, app change loan and arrearage, and each tag is subjected to id identification to obtain a sequence of 1, 23 and 87, so that text data are converted into a data sequence.
205. And performing prediction operation on the digitized behavior text data by using a preset abnormal behavior language processing model and a preset natural language processing algorithm to obtain abnormal behavior user information.
The preset abnormal behavior language processing model is an Xlnet model, the preset natural language processing algorithm is a conditional random field crf algorithm, and the digitalized behavior text data is subjected to prediction operation through the combination of the Xlnet model crf algorithm to obtain abnormal behavior user information. The crf algorithm is a rule for further defining the abnormal behavior user information, for example, a rule for defining the abnormal behavior corresponding to a specific behavior tag is used, that is, the crf algorithm is used for determining the abnormal behavior user information from a text tag set output by Xlnet model operation according to the rule for defining the abnormal behavior user information, for example, a credit blacklist rule.
For further illustration and limitation, step 205 may be: and processing the behavior data sequence as the input of a preset abnormal behavior language processing model to obtain a sequence set of different labels, performing prediction operation on the sequence set of different labels according to a preset natural language processing algorithm to determine abnormal behavior user information, wherein a characteristic function in the preset natural language processing algorithm is a preset abnormal behavior matching rule.
The crf algorithm can utilize a limited credit investigation blacklist rule to perform prediction screening on a sequence set with different labels obtained after the operation of the preset abnormal behavior language processing model, and determine the abnormal behavior user information, wherein the preset natural language processing algorithm crf performs natural language processing operation by setting a characteristic function as a preset abnormal behavior matching rule, for example, the preset credit investigation blacklist rule, so as to obtain the abnormal behavior user information, namely the user information marked as a blacklist.
The invention provides another method for determining abnormal behavior user information, which comprises the steps of obtaining all behavior text data of a user, classifying and marking the behavior text data according to a classification system, and performing abnormal behavior prediction operation on the classified and marked behavior text data by using a trained abnormal behavior prediction model to determine the abnormal behavior user information, thereby realizing accurate judgment on the abnormal behavior user, reducing the consumption of human resources, and realizing self-adaptive adjustment of screening conditions on the abnormal behavior user, thereby improving the determination efficiency and accuracy of the abnormal behavior user.
Further, as an implementation of the method shown in fig. 1, an embodiment of the present invention provides an apparatus for determining user information with abnormal behavior, as shown in fig. 3, where the apparatus includes: an acquisition module 31, a classification module 32, and a determination module 33.
The acquiring module 31 is used for acquiring all behavior text data in the user service transaction process;
the classification module 32 is configured to perform classification labeling on the behavior text data according to a classification system, where the classification system is configured to divide category hierarchies of different service behaviors in the behavior text data;
the determining module 33 is configured to perform abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model, and determine the user information of the abnormal behavior.
The invention provides a device for determining abnormal behavior user information, which is compared with the technology that users in the existing credit loss blacklist are manually screened and determined by judging whether the credit loss times of the users in each enterprise exceed specific times, the embodiment of the invention obtains all behavior text data of the users, classifies and marks the behavior text data according to a classification system, and performs abnormal behavior prediction operation on the classified and marked behavior text data by utilizing a trained abnormal behavior prediction model to determine the abnormal behavior user information, thereby realizing the accurate judgment of the abnormal behavior users, reducing the consumption of human resources, realizing the self-adaptive adjustment of screening conditions of the abnormal behavior users, and further improving the determination efficiency and accuracy of the abnormal behavior users.
Further, as an implementation of the method shown in fig. 2, an embodiment of the present invention provides another apparatus for determining user information with abnormal behavior, as shown in fig. 4, where the apparatus includes: an acquisition module 41, a classification module 42, a determination module 43, and a training module 44.
An obtaining module 41, configured to obtain all behavior text data in a user service transaction process;
the classification module 42 is configured to perform classification labeling on the behavior text data according to a classification system, where the classification system is configured to divide category hierarchies of different service behaviors in the behavior text data;
and the determining module 43 is configured to perform abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model, and determine the user information of the abnormal behavior.
Further, the determining module 43 includes:
the processing unit 4301 is configured to perform a numeralization process on the behavior text data obtained after the classification marking;
and the operation unit 4302 is configured to perform prediction operation on the digitized behavior text data by using a preset abnormal behavior language processing model in combination with a preset natural language processing algorithm, so as to obtain abnormal behavior user information.
Further, the processing unit 4301 is specifically configured to perform serialization sorting on the behavior text data after the classification marking, and generate a behavior data sequence according to the sorted behavior text data.
Further, the operation unit 4302 is specifically configured to process the behavior data sequence as an input of a preset abnormal behavior language processing model to obtain a sequence set of different tags, perform prediction operation on the sequence set of different tags according to a preset natural language processing algorithm, and determine abnormal behavior user information, where a feature function in the preset natural language processing algorithm is a preset abnormal behavior matching rule.
Further, the classification module 42 is specifically configured to classify the behavior text data according to different category hierarchies by using a natural language processing manner, and perform tag labeling on the classified search behavior text data.
Further, the classification module 42 is specifically configured to determine a category hierarchy and a label set for classifying and labeling the behavior text data according to a classification system by comparing the classification accuracy with the labeling accuracy.
Further, the classification module 42 includes:
a classification unit 4201, configured to classify the behavior text data according to different category hierarchies by using a natural language processing manner, and mark the classified behavior text data with tag labels;
the determining unit 4202 is configured to calculate each classification accuracy and each marking accuracy, and determine a category level and a marking set of the behavior text data for classification marking according to a product of the classification accuracy and the marking accuracy.
Further, the apparatus further comprises:
the training module 44 is configured to train an abnormal behavior prediction model according to behavior text data and abnormal behavior user information in a preset behavior database, where the marked behavior text data and a corresponding relationship matching the abnormal behavior user information are stored in the preset behavior database.
The invention provides another device for determining abnormal behavior user information, which determines the abnormal behavior user information by acquiring all behavior text data of users, classifying and marking the behavior text data according to a classification system and performing abnormal behavior prediction operation on the classified and marked behavior text data by using a trained abnormal behavior prediction model, thereby realizing accurate judgment on the abnormal behavior users, reducing the consumption of human resources and realizing self-adaptive adjustment of screening conditions on the abnormal behavior users, and further improving the determination efficiency and accuracy of the abnormal behavior users.
According to an embodiment of the present invention, a storage medium is provided, where the storage medium stores at least one executable instruction, and the computer executable instruction may execute the method for determining the abnormal behavior user information in any method embodiment described above.
Fig. 5 is a schematic structural diagram of a terminal according to an embodiment of the present invention, and the specific embodiment of the present invention does not limit the specific implementation of the terminal.
As shown in fig. 5, the terminal may include: a processor (processor)502, a Communications Interface 504, a memory 506, and a communication bus 508.
Wherein: the processor 502, communication interface 504, and memory 506 communicate with one another via a communication bus 508.
A communication interface 504 for communicating with network elements of other devices, such as clients or other servers.
The processor 502 is configured to execute the program 510, and may specifically perform relevant steps in the above-described method for determining abnormal behavior user information.
In particular, program 510 may include program code that includes computer operating instructions.
The processor 502 may be a central processing unit CPU, or an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits configured to implement an embodiment of the present invention. The terminal comprises one or more processors, which can be the same type of processor, such as one or more CPUs; or may be different types of processors such as one or more CPUs and one or more ASICs.
And a memory 506 for storing a program 510. The memory 506 may comprise high-speed RAM memory, and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.
The program 510 may specifically be used to cause the processor 502 to perform the following operations:
acquiring all behavior text data in the user service transaction process;
classifying and marking the behavior text data according to a classification system, wherein the classification system is used for dividing category hierarchies of different business behaviors in the behavior text data;
and performing abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model to determine the user information of the abnormal behavior.
It will be apparent to those skilled in the art that the modules or steps of the present invention described above may be implemented by a general purpose computing device, they may be centralized on a single computing device or distributed across a network of multiple computing devices, and alternatively, they may be implemented by program code executable by a computing device, such that they may be stored in a storage device and executed by a computing device, and in some cases, the steps shown or described may be performed in an order different than that described herein, or they may be separately fabricated into individual integrated circuit modules, or multiple ones of them may be fabricated into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A method for determining abnormal behavior user information is characterized by comprising the following steps:
acquiring all behavior text data in the user service transaction process;
classifying and marking the behavior text data according to a classification system, wherein the classification system is used for dividing category hierarchies of different business behaviors in the behavior text data;
and performing abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model to determine the user information of the abnormal behavior.
2. The method according to claim 1, wherein the performing abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model to determine the abnormal behavior user information comprises:
performing numerical processing on the behavior text data obtained after the classification marking;
and performing prediction operation on the digitized behavior text data by using a preset abnormal behavior language processing model and a preset natural language processing algorithm to obtain abnormal behavior user information.
3. The method according to claim 2, wherein the digitizing the behavior text data obtained after the classification labeling comprises:
and performing serialization sequencing on the behavior text data after the classification marking, and generating a behavior data sequence according to the sequenced behavior text data.
4. The method according to claim 3, wherein the step of performing a prediction operation on the digitized behavior text data by using a preset abnormal behavior language processing model in combination with a preset natural language processing algorithm to obtain the abnormal behavior user information comprises:
and processing the behavior data sequence as the input of a preset abnormal behavior language processing model to obtain a sequence set of different labels, performing prediction operation on the sequence set of different labels according to a preset natural language processing algorithm to determine abnormal behavior user information, wherein a characteristic function in the preset natural language processing algorithm is a preset abnormal behavior matching rule.
5. The method of claim 1, wherein the categorically labeling the behavioral text data according to a taxonomy comprises:
and classifying the behavior text data according to different category hierarchies by using a natural language processing mode, and labeling tag labels on the classified search behavior text data.
6. The method of claim 1, wherein the categorically labeling the behavioral text data according to a taxonomy comprises:
and determining a category hierarchy and a marking set for classifying and marking the behavior text data according to a classification system in a mode of comparing classification accuracy with marking accuracy.
7. The method of claim 6, wherein determining a category hierarchy and a tag set for classifying and tagging the behavior text data according to a classification hierarchy by comparing classification accuracy with tag accuracy comprises:
classifying the behavior text data according to different category hierarchies by using a natural language processing mode, and marking the classified behavior text data by using tag labels;
and respectively calculating the classification accuracy and the marking accuracy of each time, and determining the category level and the marking set of the behavior text data for classification marking according to the product of the classification accuracy and the marking accuracy.
8. An apparatus for determining abnormal behavior user information, comprising:
the acquisition module is used for acquiring all behavior text data in the user service transaction process;
the classification module is used for classifying and marking the behavior text data according to a classification system, and the classification system is used for dividing the category hierarchy of different business behaviors in the behavior text data;
and the determining module is used for performing abnormal behavior prediction operation on the behavior text data obtained after the classification and marking by using the trained abnormal behavior prediction model to determine the user information of the abnormal behavior.
9. A storage medium having at least one executable instruction stored therein, the executable instruction causing a processor to perform operations corresponding to the method for determining abnormal behavior user information according to any one of claims 1 to 7.
10. A terminal, comprising: the system comprises a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete mutual communication through the communication bus;
the memory is used for storing at least one executable instruction, and the executable instruction causes the processor to execute the operation corresponding to the determination method of the abnormal behavior user information as claimed in any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910740571.1A CN110597984B (en) | 2019-08-12 | 2019-08-12 | Method and device for determining abnormal behavior user information, storage medium and terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910740571.1A CN110597984B (en) | 2019-08-12 | 2019-08-12 | Method and device for determining abnormal behavior user information, storage medium and terminal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110597984A true CN110597984A (en) | 2019-12-20 |
| CN110597984B CN110597984B (en) | 2022-05-20 |
Family
ID=68854203
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910740571.1A Active CN110597984B (en) | 2019-08-12 | 2019-08-12 | Method and device for determining abnormal behavior user information, storage medium and terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110597984B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240510A (en) * | 2021-05-28 | 2021-08-10 | 平安科技（深圳）有限公司 | Abnormal user prediction method, device, equipment and storage medium |
| CN115203292A (en) * | 2022-09-15 | 2022-10-18 | 昆仑智汇数据科技(北京)有限公司 | Data processing method, device and equipment of industrial equipment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170124615A1 (en) * | 2015-11-04 | 2017-05-04 | Wal-Mart Stores, Inc. | Systems, method, and non-transitory computer-readable storage media for evaluating, storing, and managing labels for classification model evaluation and training |
| CN107239533A (en) * | 2017-05-31 | 2017-10-10 | 北京知道创宇信息技术有限公司 | Generate abnormal patterns, determine method and computing device of the user with the presence or absence of malicious act |
| CN107273454A (en) * | 2017-05-31 | 2017-10-20 | 北京京东尚科信息技术有限公司 | User data sorting technique, device, server and computer-readable recording medium |
| CN108985553A (en) * | 2018-06-05 | 2018-12-11 | 中国平安人寿保险股份有限公司 | A kind of recognition methods and equipment of abnormal user |
| CN109410036A (en) * | 2018-10-09 | 2019-03-01 | 北京芯盾时代科技有限公司 | A kind of fraud detection model training method and device and fraud detection method and device |
| CN109544190A (en) * | 2018-11-28 | 2019-03-29 | 北京芯盾时代科技有限公司 | A kind of fraud identification model training method, fraud recognition methods and device |
| CN109561069A (en) * | 2018-10-25 | 2019-04-02 | 阿里巴巴集团控股有限公司 | A kind of generation method and device, a kind of recognition methods and device of identification model |
| CN109670852A (en) * | 2018-09-26 | 2019-04-23 | 平安普惠企业管理有限公司 | User classification method, device, terminal and storage medium |
| CN110083750A (en) * | 2019-03-15 | 2019-08-02 | 平安科技（深圳）有限公司 | Blacklist screening method, device, computer equipment and storage medium |
-
2019
- 2019-08-12 CN CN201910740571.1A patent/CN110597984B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170124615A1 (en) * | 2015-11-04 | 2017-05-04 | Wal-Mart Stores, Inc. | Systems, method, and non-transitory computer-readable storage media for evaluating, storing, and managing labels for classification model evaluation and training |
| CN107239533A (en) * | 2017-05-31 | 2017-10-10 | 北京知道创宇信息技术有限公司 | Generate abnormal patterns, determine method and computing device of the user with the presence or absence of malicious act |
| CN107273454A (en) * | 2017-05-31 | 2017-10-20 | 北京京东尚科信息技术有限公司 | User data sorting technique, device, server and computer-readable recording medium |
| CN108985553A (en) * | 2018-06-05 | 2018-12-11 | 中国平安人寿保险股份有限公司 | A kind of recognition methods and equipment of abnormal user |
| CN109670852A (en) * | 2018-09-26 | 2019-04-23 | 平安普惠企业管理有限公司 | User classification method, device, terminal and storage medium |
| CN109410036A (en) * | 2018-10-09 | 2019-03-01 | 北京芯盾时代科技有限公司 | A kind of fraud detection model training method and device and fraud detection method and device |
| CN109561069A (en) * | 2018-10-25 | 2019-04-02 | 阿里巴巴集团控股有限公司 | A kind of generation method and device, a kind of recognition methods and device of identification model |
| CN109544190A (en) * | 2018-11-28 | 2019-03-29 | 北京芯盾时代科技有限公司 | A kind of fraud identification model training method, fraud recognition methods and device |
| CN110083750A (en) * | 2019-03-15 | 2019-08-02 | 平安科技（深圳）有限公司 | Blacklist screening method, device, computer equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 何颖婧等: "基于标记集合划分的多标记分类算法", 《昆明理工大学学报（自然科学版）》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240510A (en) * | 2021-05-28 | 2021-08-10 | 平安科技（深圳）有限公司 | Abnormal user prediction method, device, equipment and storage medium |
| CN113240510B (en) * | 2021-05-28 | 2023-08-01 | 平安科技（深圳）有限公司 | Abnormal user prediction method, device, equipment and storage medium |
| CN115203292A (en) * | 2022-09-15 | 2022-10-18 | 昆仑智汇数据科技(北京)有限公司 | Data processing method, device and equipment of industrial equipment |
| CN115203292B (en) * | 2022-09-15 | 2022-11-25 | 昆仑智汇数据科技(北京)有限公司 | Data processing method, device and equipment for industrial equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110597984B (en) | 2022-05-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110428322A (en) | A kind of adaptation method and device of business datum | |
| CN112862298B (en) | Credit evaluation method for user portrait | |
| US11531987B2 (en) | User profiling based on transaction data associated with a user | |
| US10726501B1 (en) | Method to use transaction, account, and company similarity clusters derived from the historic transaction data to match new transactions to accounts | |
| CN107862425B (en) | Wind control data acquisition method, device and system and readable storage medium | |
| CN110597984B (en) | Method and device for determining abnormal behavior user information, storage medium and terminal | |
| CN113688923A (en) | Intelligent order abnormity detection method and device, electronic equipment and storage medium | |
| CN112214508A (en) | Data processing method and device | |
| CN112184005A (en) | Operation task classification method, device, equipment and storage medium | |
| US10922633B2 (en) | Utilizing econometric and machine learning models to maximize total returns for an entity | |
| CN115423578A (en) | Bidding method and system based on micro-service containerization cloud platform | |
| CN110852809A (en) | Data processing method, device, equipment and medium | |
| CN112950347B (en) | Resource data processing optimization method and device, storage medium and terminal | |
| Owda et al. | Financial discussion boards irregularities detection system (fdbs-ids) using information extraction | |
| US20140288979A1 (en) | System and method for selecting an insurance carrier | |
| CN111353728A (en) | Risk analysis method and system | |
| CN116071150A (en) | Data processing method, bank product popularization, wind control system, server and medium | |
| CN115018529A (en) | Financial advertisement generation method, device, equipment and storage medium | |
| CN114723568A (en) | Recommendation method, device, server, medium and product for financial product | |
| CN111695077A (en) | Asset information pushing method, terminal equipment and readable storage medium | |
| CN117272201B (en) | Financial behavior anomaly detection method and system based on 4W1H language model | |
| CN108153725A (en) | A kind of office work management method that operation is carried out by node | |
| CN111178740B (en) | Data processing method and device | |
| Kang | Examining the Impact of ICT Access and ICT Use on Trade in Services: Focused on Selected Service Categories. | |
| CN114862543A (en) | Authorized acquirable information acquisition method and related product |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |