CN110489307A - Interface exception call monitoring method and device - Google Patents
Interface exception call monitoring method and device Download PDFInfo
- Publication number
- CN110489307A CN110489307A CN201910795654.0A CN201910795654A CN110489307A CN 110489307 A CN110489307 A CN 110489307A CN 201910795654 A CN201910795654 A CN 201910795654A CN 110489307 A CN110489307 A CN 110489307A
- Authority
- CN
- China
- Prior art keywords
- behavior
- interface
- user
- data
- subclassification
- 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
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/3041—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system component is an input/output interface
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Abstract
A kind of interface exception call monitoring method provided by the present application and device, overcome it is existing based on blacklist, it is rule-based it is low with based on machine learning method verification and measurement ratio, false detection rate is high and the problem of can not finding in time to new abnormal behaviour.The feature of clustering algorithm is obtained using user's Portrait brand technology, user is divided into user's subclass that multiple behaviors are comparable, so that the recall rate that exceptional interface calls is higher, false detection rate is lower.Learn the behavior pattern of each user's subclass using neural network model, obtain the envelope of normal behaviour mode, the subsequent normal api interface of client, which is obtained, using trained neural network prediction calls behavior, predictive behavior and the practical api interface of client call behavior to compare, therefore new abnormal behaviour can also be found in time.
Description
Technical field
The present invention relates to interface calling technology fields, more particularly to a kind of interface exception call monitoring method and dress
It sets.
Background technique
With the raising of the level of IT application, more and more enterprises open the service ability of oneself by way of API, and
The ecosphere of oneself is constructed on the basis of servicing open.With the increase of API service interface calling amount, the calling of interface is pacified
Full problem is more and more urgent also more and more important.Based on blacklist and it is the methods of rule-based proposed in succession, but still remain
Recall rate is low and the high problem of false detection rate, and the above method can only detect known safety problem, former does not send out for new
The safety problem given birth to can not then detect.
Summary of the invention
At least one of to solve the above-mentioned problems, the application provides a kind of interface exception call monitoring method, comprising:
Behavioral data is called to generate user's Figure Characteristics according to the history interface of user;
The preset data format for calling behavior prediction model can be inputted by converting user's Figure Characteristics to, and be inputted
To the calling behavior prediction model, obtains at least one prediction interface and call behavioral data;
Behavioral data is called according at least one described prediction interface, monitoring user's further interface calls the abnormal feelings of behavior
Condition;
Wherein, described that behavior prediction model is called to include user's Figure Characteristics and predict that interface calls the mapping of behavioral data
Relationship.
In certain embodiments, further includes:
Behavioral data is called to determine subclassification belonging to user according to the history interface of user;
Transfer the calling behavior prediction model of corresponding subclassification.
In certain embodiments, further includes:
Preset multiple subclassifications;
Corresponding each subclassification establishes multiple calling behavior prediction models.
In certain embodiments, it calls behavioral data to determine subclassification belonging to user according to the history interface of user, wraps
It includes:
Multiple truth labels of user are generated according to the behavioral data that multiple interfaces of user call;
Multiple truth labels are input to preset model label and generate model, generate at least one model label;
According to multiple truth labels and at least one described model label, service label is generated;
According to multiple service labels, the subclassification is determined.
In certain embodiments, further includes:
Call data as training set, training each calling row the history interface of all users in each subclassification of correspondence
For prediction model.
In certain embodiments, each calling behavior prediction model of the training, comprising:
For each subclassification, the history interface for choosing a user in the subclassification calls data sequentially in time
It is divided into first part and second part, using first part's data as input, for second part data as output, training is corresponding
The calling behavior prediction model;
Iterative operation is executed, calls data to divide sequentially in time the history interface of another user in the subclassification
For first part and second part, and then first part's data of another user is made to substitute first part's number of originating subscriber
According to as input, second part data substitute the second part data of originating subscriber as output, the corresponding calling of training
Behavior prediction model, until first part's data of user any in the subclassification are input to the calling behavior prediction mould
Type, when output result includes at least the second part data of the user, the calling behavior prediction model training is completed;
Obtain the calling behavior prediction model of training completion.
In certain embodiments, any one calling behavior prediction model of training, comprising:
The history interface for choosing all users in the subclassification that corresponding one calls behavior prediction model calls data, according to when
Between be sequentially generated and the one-to-one behavior sequence of each user;
For each behavior sequence, maps total interface therein and call behavior, sequence vector is obtained, wherein each interface
Calling behavior correspondence mappings are a behavior vector;
By each sequence vector be split as calling behavior occur the first behavior before the preset time point to
Quantity set, and call behavior that the second behavior vector set after the preset time point occurs;
It, accordingly will be in each behavior sequence using the first behavior vector set in each behavior sequence as input
Each behavior vector is as object vector, the training calling behavior prediction model in the second behavior vector set.
In certain embodiments, behavioral data is called to determine subclassification belonging to user according to the history interface of user, also
Include:
The model label, which is established, based on deep learning generates model;
Using the marked truth labels and the model label as one group of training set, the training model label is raw
At model.
It is in certain embodiments, described to preset multiple subclassifications, comprising:
Business classification is carried out to user according to the history service data of all users and generates multiple subclassifications, wherein often
A corresponding service label of the subclassification.
In certain embodiments, the history service data according to user carry out business classification to user and generate multiple institutes
State subclassification, comprising:
Clustering processing is carried out to the history service data of all users, obtains the corresponding subclassification.
In certain embodiments, it includes predictive behavior type and each predictive behavior that the prediction interface, which calls behavioral data,
The called probability value of type;Behavioral data is called according at least one described prediction interface, monitoring user's further interface calls
The abnormal conditions of behavior, comprising:
Each predictive behavior type is ranked up from big to small according to probability value, forms predictive behavior sequence;
The interface for obtaining the setting quantity of user's subsequent execution calls behavior;
Judge that the interface obtained calls whether behavior is in front of the setting position of the predictive behavior sequence, if
It is judged as YES, it is determined that user's further interface calls behavior normal.
The application also provides a kind of interface exception call monitoring device, comprising:
User's Figure Characteristics generation module calls behavioral data to generate user's Figure Characteristics according to the history interface of user;
Input module, the preset data lattice for calling behavior prediction model can be inputted by converting user's Figure Characteristics to
Formula, and it is input to the calling behavior prediction model, it obtains at least one prediction interface and calls behavioral data;
Abnormal conditions monitoring modular calls behavioral data according at least one described prediction interface, monitors continued access after user
Mouth calls the abnormal conditions of behavior;
Wherein, described that behavior prediction model is called to include user's Figure Characteristics and predict that interface calls the mapping of behavioral data
Relationship.
In certain embodiments, further includes:
Subclassification determining module calls behavioral data to determine subclassification belonging to user according to the history interface of user;
It calls behavior prediction model to transfer module, transfers the calling behavior prediction model of corresponding subclassification.
In certain embodiments, further includes:
Subclassification pre-sets module, presets multiple subclassifications;
Behavior prediction model building module is called, corresponding each subclassification establishes multiple calling behavior prediction models.
In certain embodiments, it calls behavioral data to determine subclassification belonging to user according to the history interface of user, wraps
It includes:
Truth labels generation unit generates multiple true marks of user according to the behavioral data that multiple interfaces of user call
Label;
Multiple truth labels are input to preset model label and generate model, generated by model label generation unit
At least one model label;
Service label generation unit generates business according to multiple truth labels and at least one described model label
Label;
Subclassification determination unit determines the subclassification according to multiple service labels.
In certain embodiments, further includes:
The history interface of all users in each subclassification of correspondence is called data as training set, training by training module
Each calling behavior prediction model.
In certain embodiments, the training module includes:
Training unit, for each subclassification, the history interface for choosing in the subclassification user calls data to press
It is divided into first part and second part according to time sequencing, using first part's data as input, second part data are used as output,
The corresponding calling behavior prediction model of training;
Repetitive exercise unit executes iterative operation, the history interface of another user in the subclassification is called data
It is divided into first part and second part sequentially in time, and then so that first part's data of another user is substituted starting and use
For first part's data at family as input, second part data substitute the second part data of originating subscriber as output, training
The corresponding calling behavior prediction model, until first part's data of user any in the subclassification are input to the tune
With behavior prediction model, when exporting second part data of the result including at least the user, the calling behavior prediction model instruction
Practice and completes;
Output unit obtains the calling behavior prediction model of training completion.
In certain embodiments, the training module includes:
Behavior sequence generation unit, the history for choosing all users in the subclassification of corresponding calling behavior prediction model connect
Mouth calls data, generates and the one-to-one behavior sequence of each user sequentially in time;
Map unit maps total interface therein and calls behavior, obtain sequence vector for each behavior sequence,
In each interface to call behavior correspondence mappings be a behavior vector;
Each sequence vector is split as calling behavior and occurred in the preset time point by sequence vector split cells
The first behavior vector set before, and call behavior that the second behavior vector set after the preset time point occurs;
Training unit, using the first behavior vector set in each behavior sequence as input, accordingly by each row
It is each behavior vector in the second behavior vector set in sequence as object vector, the training calling behavior prediction mould
Type.
In certain embodiments, subclassification determining module, further includes:
Model label generates model foundation unit, establishes the model label based on deep learning and generates model;
Model label generates model training unit, using the marked truth labels and the model label as one group
Training set, the training model label generate model.
In certain embodiments, the subclassification pre-set module according to the history service data of all users to user into
Industry business classification generates multiple subclassifications, wherein the corresponding service label of each subclassification.
In certain embodiments, the subclassification pre-sets module and carries out at cluster to the history service data of all users
Reason, obtains the corresponding subclassification.
In certain embodiments, it includes predictive behavior type and each predictive behavior that the prediction interface, which calls behavioral data,
The called probability value of type, the abnormal conditions monitoring modular include:
Predictive behavior sequence forms unit, is ranked up from big to small to each predictive behavior type according to probability value, shape
At predictive behavior sequence;
Further interface calls behavior acquiring unit, and the interface for obtaining the setting quantity of user's subsequent execution calls behavior;
Behavior normal judgment unit judges that the interface obtained calls whether behavior is in the predictive behavior sequence
Setting position before, if being judged as YES, it is determined that user's further interface call behavior it is normal.
The application also provides a kind of computer equipment, including memory, processor and storage on a memory and can located
The step of computer program run on reason device, the processor realizes method as described above when executing described program.
The application also provides a kind of computer readable storage medium, is stored thereon with computer program, the computer program
The step of method as described above is realized when being executed by processor.
The invention has the following beneficial effects:
A kind of interface exception call monitoring method provided by the invention and device, overcome it is existing based on blacklist, based on rule
It is then low with based on machine learning method verification and measurement ratio, false detection rate is high and the problem of can not finding in time to new abnormal behaviour.It uses
User's Portrait brand technology obtains the feature of clustering algorithm, user is divided into user's subclass that multiple behaviors are comparable, is made
The recall rate for obtaining exceptional interface calling is higher, and false detection rate is lower.Learn the behavior of each user's subclass using neural network model
Mode obtains the envelope of normal behaviour mode, obtains the subsequent normal API of client using trained neural network prediction and connects
Mouthful behavior is called, predictive behavior and the practical api interface of client call behavior to compare, therefore can also be with for new abnormal behaviour
Discovery in time.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 shows a kind of interface exception call monitoring method flow diagram in the application.
Fig. 2 shows the generating process schematic diagrames of behavior vector in the application.
Fig. 3 shows the training schematic diagram that behavior prediction model is called in the application.
Fig. 4 shows a kind of interface exception call monitoring device structural schematic diagram in the application.
Fig. 5 shows the structural schematic diagram for being suitable for the computer equipment for being used to realize the embodiment of the present application.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Processing Interface is called in safety problem at present, based on blacklist and it is the methods of rule-based there is recall rate it is low and
The problems such as false detection rate is high, and can only detect known safety problem, the safety problem never occurred can not be solved.
In view of this, the application obtains the feature of clustering algorithm using user's Portrait brand technology, user is divided into multiple rows
For in user's subclass for being comparable, so that the recall rate that exceptional interface calls is higher, false detection rate is lower.Use neural network
The behavior pattern of each user's subclass of model learning, obtains the envelope of normal behaviour mode, uses trained neural network
Prediction obtains the subsequent normal api interface of client and calls behavior, and predictive behavior and the practical api interface of client call behavior to do ratio
Compared with, therefore new abnormal behaviour can also be found in time.
Fig. 1 shows one of the embodiment of the present application interface exception call monitoring method, comprising:
S1: behavioral data is called to generate user's Figure Characteristics according to the history interface of user;
S2: the preset data format for calling behavior prediction model can be inputted by converting user's Figure Characteristics to, and
It is input to the calling behavior prediction model, at least one prediction interface is obtained and calls behavioral data;
S3: behavioral data is called according at least one described prediction interface, monitoring user's further interface calls the different of behavior
Reason condition;
Wherein, described that behavior prediction model is called to include user's Figure Characteristics and predict that interface calls the mapping of behavioral data
Relationship.
In the application, calling behavior prediction model is the neural network model that training is completed, the calling behavior that training is completed
It include user's Figure Characteristics and the mapping relations for predicting interface calling behavioral data in prediction model.
User's Figure Characteristics refer to analyzing user interface calling behavior, and what is obtained can describe the feature of user
Data.
In some embodiments, the attribute feature of each user is all different, and based on different subscriber datas, can will be used
Family group is divided into multiple classification, corresponding, the corresponding calling behavior prediction model of each classification.
Certainly, each calling behavior prediction model is consistent when initially setting up, by using corresponding user
Data are trained, and call behavior prediction model due to the difference of training data, inevitable difference, and only needle two-by-two after training
There is accurate prediction effect to corresponding user's classification.
Obviously, in the embodiment, the above method further include:
S01: behavioral data is called to determine subclassification belonging to user according to the history interface of user;
S02: the calling behavior prediction model of corresponding subclassification is transferred.
Corresponding, the above method further include:
Preset multiple subclassifications;
Corresponding each subclassification establishes multiple calling behavior prediction models.
In some embodiments, user's classification is carried out based on clustering algorithm, for example, by using k-means clustering algorithm pair
All users do clustering processing.
In the specific implementation, it needs to generate corresponding service label to each user, then according to service label to user
Classify.
Specifically, in some embodiments, user's classification can be carried out based on service label, step S01 is specifically included:
S011: multiple truth labels of user are generated according to the behavioral data that multiple interfaces of user call;
S012: multiple truth labels are input to preset model label and generate model, generate at least one model
Label;
S013: according to multiple truth labels and at least one described model label, service label is generated;
S014: according to multiple service labels, the subclassification is determined.
Model label is generated model and is also possible to be established based on deep learning, model label generate model can it is online or
It is offline to establish, for establishing online, i.e. above-mentioned steps S01 further include:
S0101: the model label is established based on deep learning and generates model;
S0102: using the marked truth labels and the model label as one group of training set, the training model
Label generates model.
Service label is label label corresponding with each subclassification, and subclassification is the history service number according to all users
According to generation, such as according to passing historical data, subclassification is divided into robustness, radical type, assets value and is greater than 100w, available assets value
Greater than 30w etc., exhaustion is not done herein.
The foundation of subclassification needs to carry out all historical datas clustering processing, can be by historical data before clustering processing
It is converted into numeric type feature, clustering processing is then done to all users using k-means clustering algorithm.
In the application, behavior prediction model is called to can be based on neural network.Specifically, in some embodiments
In, using LSTM neural network as example.
User is numbered in the calling behavior of each of limited calling action space, user calls in space in total
There is N kind behavior behavior, then assigns a behavior number to every kind of calling behavior.Such as there are five types of behaviors in action space, respectively
It is inquiry combustion gas expense, pays combustion gas expense, inquires water rate, pays water rate and pay tuition fee, then successively to number every kind of behavior be 0,
1,2,3 and 4.
To user's subclass that the history interface by user calls behavioral data to determine, each behavior obtained above is compiled
It number is mapped in the behavior vector that a length is M, the initial value of the vector can generate at random, can also be according to certain point
Cloth generates, as shown in Figure 2.
From the behavior sequence for obtaining user in each user's subclass in database, by each behavior row in this behavior sequence
To obtain the sequence vector of behavior by number and mapping.Using current behavior vector each in sequence vector as object vector,
The neural network mould that the behavior vector at all moment before current time is made of as list entries, training LSTM neural network
Type, as shown in Figure 3.
In the application, model training can be carried out online or be carried out offline, i.e., the step of model training may include
In the method and step that the application executes, it can also be detached from the method and step executed in the application, called by way of pre-stored
The model that training is completed.By taking on-line training model as an example, in the embodiment, the step of training pattern, includes:
S021: the history interface for choosing all users in the subclassification that corresponding one calls behavior prediction model calls data,
It generates and the one-to-one behavior sequence of each user sequentially in time;
S022: being directed to each behavior sequence, maps total interface therein and calls behavior, sequence vector is obtained, wherein often
It is a behavior vector that a interface, which calls behavior correspondence mappings,;
S023: each sequence vector is split as the first row of the calling behavior generation before the preset time point
The second behavior vector set after the preset time point occurs for vector set, and calling behavior;
S024: using the first behavior vector set in each behavior sequence as input, accordingly by each behavior sequence
Each behavior vector is as object vector, the training calling behavior prediction model in the second behavior vector set in column.
In some embodiments, can input the preset data format for calling behavior model is numeric type format, Yong Huhua
Picture feature is character type data, can convert character type data to numeric type data, and then be input to preset calling behavior
Prediction model.
It includes called general of predictive behavior type and each predictive behavior type that the prediction interface, which calls behavioral data,
Rate value.
Such as each behavior type corresponds to some binary character, probability value is indicated by hexadecimal character.
In the embodiment, step S3 is specifically included:
S31: being ranked up each predictive behavior type according to probability value from big to small, forms predictive behavior sequence;
S32: the interface for obtaining the setting quantity of user's subsequent execution calls behavior;
S33: judge obtain the interface call behavior whether be in the predictive behavior sequence setting position it
Before, if being judged as YES, it is determined that user's further interface calls behavior normal.
In the embodiment, by carrying out packing judgement to multiple calling behaviors, the reliability and accuracy of monitoring are increased.
Certainly, in other embodiments, behavior can also be called to carry out independent judgment for the interface of each subsequent generation.
Such as judge that user's further interface calls whether behavior belongs at least the one of the calling behavior prediction model output
A prediction interface calls one in behavioral data, if belonging to, determine that user interface calls behavior normal, otherwise user interface tune
Use abnormal behavior.
In the embodiment, model training is again based on above-mentioned similar mode, and model training specifically includes:
For each subclassification, the history interface for choosing a user in the subclassification calls data sequentially in time
It is divided into first part and second part, using first part's data as input, for second part data as output, training is corresponding
The calling behavior prediction model;
Iterative operation is executed, calls data to divide sequentially in time the history interface of another user in the subclassification
For first part and second part, and then first part's data of another user is made to substitute first part's number of originating subscriber
According to as input, second part data substitute the second part data of originating subscriber as output, the corresponding calling of training
Behavior prediction model, until first part's data of user any in the subclassification are input to the calling behavior prediction mould
Type, when output result includes at least the second part data of the user, the calling behavior prediction model training is completed;
Obtain the calling behavior prediction model of training completion.
Since preset calling interface behavioral data itself has been converted into numeric type formatted data, or according to numeric type
Formatted data storage, therefore does not need still further to be mapped as sequence vector etc., by recording modes such as binary system, hexadecimals,
The mapping to each behavior (the corresponding numerical value of each behavior type) can be realized.
It is appreciated that interface exception call monitoring method provided by the present application, it both can be to each interface of subsequent transmission
Whether calling behavior individually judges abnormal, can also by calling behavior to carry out comprehensive descision multiple interfaces of subsequent generation,
On the one hand it improves and judges speed, on the other hand can provide judgment accuracy with comprehensive descision.
In addition, the application overcomes existing based on blacklist, rule-based, erroneous detection low with based on machine learning method verification and measurement ratio
Rate is high and the problem of can not finding in time to new abnormal behaviour.The feature of clustering algorithm is obtained using user's Portrait brand technology, it will
User is divided into user's subclass that multiple behaviors are comparable, so that the recall rate that exceptional interface calls is higher, false detection rate
It is lower.The behavior pattern for learning each user's subclass using neural network model obtains the envelope of normal behaviour mode, uses
Trained neural network prediction obtains the subsequent normal api interface of client and calls behavior, and predictive behavior and the practical API of client connect
Mouth calling behavior compares, therefore new abnormal behaviour can also be found in time.
Based on identical inventive concept, Fig. 4 shows a kind of interface exception call monitoring device in the embodiment of the present application, tool
Body includes: user's Figure Characteristics generation module 1, calls behavioral data to generate user's Figure Characteristics according to the history interface of user;
Input module 2, the preset data format for calling behavior prediction model can be inputted by converting user's Figure Characteristics to, and defeated
Enter to the calling behavior prediction model, obtains at least one prediction interface and call behavioral data;Abnormal conditions monitoring modular 3,
Behavioral data is called according at least one described prediction interface, monitoring user's further interface calls the abnormal conditions of behavior;Wherein,
It is described that behavior prediction model is called to include user's Figure Characteristics and predict that interface calls the mapping relations of behavioral data.
Based on identical inventive concept, in some embodiments, interface exception call monitoring device further include: subclassification determines
Module calls behavioral data to determine subclassification belonging to user according to the history interface of user;
It calls behavior prediction model to transfer module, transfers the calling behavior prediction model of corresponding subclassification.
Based on identical inventive concept, in some embodiments, interface exception call monitoring device further include: subclassification is default
Module is set, multiple subclassifications are preset;Behavior prediction model building module is called, corresponding each subclassification is established multiple
Call behavior prediction model.
Based on identical inventive concept, in some embodiments, calls behavioral data to determine according to the history interface of user and use
Subclassification belonging to family, comprising: truth labels generation unit generates user according to the behavioral data that multiple interfaces of user call
Multiple truth labels;Multiple truth labels are input to preset model label and generate mould by model label generation unit
Type generates at least one model label;Service label generation unit, according to multiple truth labels and at least one described mould
Type label generates service label;Subclassification determination unit determines the subclassification according to multiple service labels.
Based on identical inventive concept, in some embodiments, interface exception call monitoring device further include: training module,
Call data as training set, training each calling behavior prediction mould the history interface of all users in each subclassification of correspondence
Type.
Based on identical inventive concept, in some embodiments, the training module includes: training unit, for every height
Classification, the history interface for choosing in the subclassification user call data to be divided into first part and second sequentially in time
Part, using first part's data as input, second part data are as output, the corresponding calling behavior prediction mould of training
Type;Repetitive exercise unit executes iterative operation, by the history interface of another user in the subclassification call data according to when
Between be sequentially divided into first part and second part, and then first part's data of another user is made to substitute the of originating subscriber
A part of data are as input, and second part data substitute the second part data of originating subscriber as output, and training is corresponding
The calling behavior prediction model, until first part's data of user any in the subclassification are input to the calling behavior
Prediction model, when output result includes at least the second part data of the user, the calling behavior prediction model training is completed;
Output unit obtains the calling behavior prediction model of training completion.
Based on identical inventive concept, in some embodiments, the training module includes: behavior sequence generation unit, choosing
It takes the history interface of all users in corresponding one subclassification for calling behavior prediction model to call data, generates sequentially in time
With the one-to-one behavior sequence of each user;Map unit maps total interface therein and calls for each behavior sequence
Behavior obtains sequence vector, wherein it is a behavior vector that each interface, which calls behavior correspondence mappings,;Sequence vector splits single
Each sequence vector is split as first behavior vector set of the calling behavior generation before the preset time point by member,
And call behavior that the second behavior vector set after the preset time point occurs;Training unit, by each behavior sequence
In the first behavior vector set as input, accordingly will be every in the second behavior vector set in each behavior sequence
A behavior vector is as object vector, the training calling behavior prediction model.
Based on identical inventive concept, in some embodiments, subclassification determining module, further includes: model label generates mould
Type establishes unit, establishes the model label based on deep learning and generates model;Model label generates model training unit, will
The truth labels and the model label of label generate model as one group of training set, the training model label.
Based on identical inventive concept, in some embodiments, the subclassification pre-sets module going through according to all users
History business datum carries out business classification to user and generates multiple subclassifications, wherein the corresponding business of each subclassification
Label.
Based on identical inventive concept, in some embodiments, the subclassification pre-sets module to the history of all users
Business datum carries out clustering processing, obtains the corresponding subclassification.
Based on identical inventive concept, in some embodiments, it includes predictive behavior that the prediction interface, which calls behavioral data,
The probability value that type and each predictive behavior type are called, the abnormal conditions monitoring modular includes: predictive behavior sequence shape
At unit, each predictive behavior type is ranked up from big to small according to probability value, forms predictive behavior sequence;Further interface
Calling behavior acquiring unit, the interface for obtaining the setting quantity of user's subsequent execution call behavior;Behavior normal judgment unit, sentences
The disconnected interface obtained calls whether behavior is in front of the setting position of the predictive behavior sequence, if being judged as YES,
Then determine that user's further interface calls behavior normal.
It is appreciated that a kind of interface exception call monitoring device provided by the present application, overcomes existing based on blacklist, be based on
Rule is low with based on machine learning method verification and measurement ratio, false detection rate is high and the problem of can not finding in time to new abnormal behaviour.Make
The feature of clustering algorithm is obtained with user's Portrait brand technology, user is divided into user's subclass that multiple behaviors are comparable,
So that the recall rate that exceptional interface calls is higher, false detection rate is lower.Learn the row of each user's subclass using neural network model
For mode, the envelope of normal behaviour mode is obtained, obtains the subsequent normal API of client using trained neural network prediction
Interface calls behavior, and predictive behavior and the practical api interface of client call behavior to compare, therefore can also for new abnormal behaviour
To find in time.
System, device, module or the unit that above-described embodiment illustrates can specifically realize by computer chip or entity,
Or it is realized by the product with certain function.It is a kind of typically to realize that equipment is computer equipment, specifically, computer is set
It is standby for example can for personal computer, laptop computer, cellular phone, camera phone, smart phone, personal digital assistant,
Media player, navigation equipment, electronic mail equipment, game console, tablet computer, wearable device or these equipment
In any equipment combination.
Computer equipment specifically includes memory, processor and storage on a memory simultaneously in a typical example
The computer program that can be run on a processor is realized when the processor executes described program and is held as described above by client
Capable method, alternatively, the processor realizes the method executed as described above by server when executing described program.
Below with reference to Fig. 5, it illustrates the structural schematic diagrams for the computer equipment for being suitable for being used to realize the embodiment of the present application.
As shown in figure 5, computer equipment 600 includes central processing unit (CPU) 601, it can be read-only according to being stored in
Program in memory (ROM) 602 is loaded into random access storage device (RAM) from storage section 608) program in 603
And execute various work appropriate and processing.In RAM603, also it is stored with system 600 and operates required various program sum numbers
According to.CPU601, ROM602 and RAM603 are connected with each other by bus 604.Input/output (I/O) interface 605 is also connected to
Bus 604.
I/O interface 605 is connected to lower component: the importation 606 including keyboard, mouse etc.;It is penetrated including such as cathode
The output par, c 607 of spool (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.;Storage section 608 including hard disk etc.;
And including such as LAN card, the communications portion 609 of the network interface card of modem etc..Communications portion 609 via such as because
The network of spy's net executes communication process.Driver 610 is also connected to I/O interface 606 as needed.Detachable media 611, such as
Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 610, in order to read from thereon
Computer program be mounted as needed such as storage section 608.
Particularly, according to an embodiment of the invention, may be implemented as computer above with reference to the process of flow chart description
Software program.For example, the embodiment of the present invention includes a kind of computer program product comprising be tangibly embodied in machine readable
Computer program on medium, the computer program include the program code for method shown in execution flow chart.At this
In the embodiment of sample, which can be downloaded and installed from network by communications portion 609, and/or from removable
Medium 611 is unloaded to be mounted.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method
Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves
State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable
Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM),
Digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape magnetic disk storage or other magnetic storage devices
Or any other non-transmission medium, can be used for storage can be accessed by a computing device information.As defined in this article, it calculates
Machine readable medium does not include temporary computer readable media (transitory media), such as the data-signal and carrier wave of modulation.
For convenience of description, it is divided into various units when description apparatus above with function to describe respectively.Certainly, implementing this
The function of each unit can be realized in the same or multiple software and or hardware when application.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
It should also be noted that, the terms "include", "comprise" or its any other variant are intended to nonexcludability
It include so that the process, method, commodity or the equipment that include a series of elements not only include those elements, but also to wrap
Include other elements that are not explicitly listed, or further include for this process, method, commodity or equipment intrinsic want
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including described want
There is also other identical elements in the process, method of element, commodity or equipment.
It will be understood by those skilled in the art that embodiments herein can provide as method, system or computer program product.
Therefore, complete hardware embodiment, complete software embodiment or embodiment combining software and hardware aspects can be used in the application
Form.It is deposited moreover, the application can be used to can be used in the computer that one or more wherein includes computer usable program code
The shape for the computer program product implemented on storage media (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
Formula.
The application can describe in the general context of computer-executable instructions executed by a computer, such as program
Module.Generally, program module includes routines performing specific tasks or implementing specific abstract data types, programs, objects, group
Part, data structure etc..The application can also be practiced in a distributed computing environment, in these distributed computing environments, by
Task is executed by the connected remote processing devices of communication network.In a distributed computing environment, program module can be with
In the local and remote computer storage media including storage equipment.
All the embodiments in this specification are described in a progressive manner, same and similar portion between each embodiment
Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.Especially for system reality
For applying example, since it is substantially similar to the method embodiment, so being described relatively simple, related place is referring to embodiment of the method
Part explanation.
The above description is only an example of the present application, is not intended to limit this application.For those skilled in the art
For, various changes and changes are possible in this application.All any modifications made within the spirit and principles of the present application are equal
Replacement, improvement etc., should be included within the scope of the claims of this application.
Claims (24)
1. a kind of interface exception call monitoring method characterized by comprising
Behavioral data is called to generate user's Figure Characteristics according to the history interface of user;
The preset data format for calling behavior prediction model can be inputted by converting user's Figure Characteristics to, and be input to institute
It states and calls behavior prediction model, obtain at least one prediction interface and call behavioral data;
Behavioral data is called according at least one described prediction interface, monitoring user's further interface calls the abnormal conditions of behavior;
Wherein, described that behavior prediction model is called to include user's Figure Characteristics and predict that interface calls the mapping of behavioral data to close
System.
2. interface exception call monitoring method according to claim 1, which is characterized in that further include:
Behavioral data is called to determine subclassification belonging to user according to the history interface of user;
Transfer the calling behavior prediction model of corresponding subclassification.
3. interface exception call monitoring method according to claim 2, which is characterized in that further include:
Preset multiple subclassifications;
Corresponding each subclassification establishes multiple calling behavior prediction models.
4. interface exception call monitoring method according to claim 2, which is characterized in that according to the history interface tune of user
With behavioral data determine user belonging to subclassification, comprising:
Multiple truth labels of user are generated according to the behavioral data that multiple interfaces of user call;
Multiple truth labels are input to preset model label and generate model, generate at least one model label;
According to multiple truth labels and at least one described model label, service label is generated;
According to multiple service labels, the subclassification is determined.
5. interface exception call monitoring method according to claim 3, which is characterized in that further include:
Call data as training set the history interface of all users in each subclassification of correspondence, each calling behavior of training is pre-
Survey model.
6. interface exception call monitoring method according to claim 5, which is characterized in that each calling behavior of training
Prediction model, comprising:
For each subclassification, the history interface for choosing in the subclassification user calls data to be divided into sequentially in time
First part and second part, using first part's data as input, for second part data as output, training is corresponding described
Call behavior prediction model;
Iterative operation is executed, calls data to be divided into the sequentially in time the history interface of another user in the subclassification
A part and second part, and then the first part's data for making first part's data of another user substitute originating subscriber are made
For input, second part data substitute the second part data of originating subscriber as output, the corresponding calling behavior of training
Prediction model, until first part's data of user any in the subclassification are input to the calling behavior prediction model, it is defeated
When result includes at least the second part data of the user out, the calling behavior prediction model training is completed;
Obtain the calling behavior prediction model of training completion.
7. interface exception call monitoring method according to claim 5, which is characterized in that any one calling behavior of training
Prediction model, comprising:
The history interface for choosing all users in the subclassification that corresponding one calls behavior prediction model calls data, suitable according to the time
Sequence generates and the one-to-one behavior sequence of each user;
It for each behavior sequence, maps total interface therein and calls behavior, obtain sequence vector, wherein each interface calls
Behavior correspondence mappings are a behavior vector;
Each sequence vector is split as first behavior vector set of the calling behavior generation before the preset time point,
And call behavior that the second behavior vector set after the preset time point occurs;
It, accordingly will be described in each behavior sequence using the first behavior vector set in each behavior sequence as input
Each behavior vector is as object vector, the training calling behavior prediction model in second behavior vector set.
8. interface exception call monitoring method according to claim 4, which is characterized in that according to the history interface tune of user
With behavioral data determine user belonging to subclassification, further includes:
The model label, which is established, based on deep learning generates model;
Using the marked truth labels and the model label as one group of training set, the training model label generates mould
Type.
9. interface exception call monitoring method according to claim 3, which is characterized in that it is described preset it is multiple described
Subclassification, comprising:
Business classification is carried out to user according to the history service data of all users and generates multiple subclassifications, wherein each institute
State the corresponding service label of subclassification.
10. interface exception call monitoring method according to claim 9, which is characterized in that the history according to user
Business datum carries out business classification to user and generates multiple subclassifications, comprising:
Clustering processing is carried out to the history service data of all users, obtains the corresponding subclassification.
11. interface exception call monitoring method according to claim 1, which is characterized in that the prediction interface calls row
It include the probability value that predictive behavior type and each predictive behavior type are called for data;It is connect according at least one described prediction
Mouth calls behavioral data, and monitoring user's further interface calls the abnormal conditions of behavior, comprising:
Each predictive behavior type is ranked up from big to small according to probability value, forms predictive behavior sequence;
The interface for obtaining the setting quantity of user's subsequent execution calls behavior;
Judge that the interface obtained calls whether behavior is in front of the setting position of the predictive behavior sequence, if judgement
It is yes, it is determined that user's further interface calls behavior normal.
12. a kind of interface exception call monitoring device characterized by comprising
User's Figure Characteristics generation module calls behavioral data to generate user's Figure Characteristics according to the history interface of user;
Input module, the preset data format for calling behavior prediction model can be inputted by converting user's Figure Characteristics to,
And it is input to the calling behavior prediction model, it obtains at least one prediction interface and calls behavioral data;
Abnormal conditions monitoring modular calls behavioral data according at least one described prediction interface, monitors user's further interface tune
With the abnormal conditions of behavior;
Wherein, described that behavior prediction model is called to include user's Figure Characteristics and predict that interface calls the mapping of behavioral data to close
System.
13. interface exception call monitoring device according to claim 12, which is characterized in that further include:
Subclassification determining module calls behavioral data to determine subclassification belonging to user according to the history interface of user;
It calls behavior prediction model to transfer module, transfers the calling behavior prediction model of corresponding subclassification.
14. interface exception call monitoring device according to claim 13, which is characterized in that further include:
Subclassification pre-sets module, presets multiple subclassifications;
Behavior prediction model building module is called, corresponding each subclassification establishes multiple calling behavior prediction models.
15. interface exception call monitoring device according to claim 13, which is characterized in that according to the history interface of user
Behavioral data is called to determine subclassification belonging to user, comprising:
Truth labels generation unit generates multiple truth labels of user according to the behavioral data that multiple interfaces of user call;
Multiple truth labels are input to preset model label and generate model, generated at least by model label generation unit
One model label;
Service label generation unit generates service label according to multiple truth labels and at least one described model label;
Subclassification determination unit determines the subclassification according to multiple service labels.
16. interface exception call monitoring device according to claim 14, which is characterized in that further include:
The history interface of all users in each subclassification of correspondence is called data as training set by training module, and training is each
Call behavior prediction model.
17. interface exception call monitoring device according to claim 16, which is characterized in that the training module includes:
Training unit, for each subclassification, choose a user in the subclassification history interface call data according to when
Between be sequentially divided into first part and second part, using first part's data as input, second part data as output, training
The corresponding calling behavior prediction model;
Repetitive exercise unit executes iterative operation, by the history interface of another user in the subclassification call data according to
Time sequencing is divided into first part and second part, and then first part's data of another user is made to substitute originating subscriber
First part's data are as input, and second part data substitute the second part data of originating subscriber as output, and training corresponds to
The calling behavior prediction model, until first part's data of user any in the subclassification are input to calling row
For prediction model, when exporting second part data of the result including at least the user, the calling behavior prediction model training is complete
At;
Output unit obtains the calling behavior prediction model of training completion.
18. interface exception call monitoring device according to claim 16, which is characterized in that the training module includes:
Behavior sequence generation unit chooses the history interface tune of all users in the subclassification that corresponding one calls behavior prediction model
With data, generate and the one-to-one behavior sequence of each user sequentially in time;
Map unit maps total interface therein and calls behavior, sequence vector is obtained, wherein often for each behavior sequence
It is a behavior vector that a interface, which calls behavior correspondence mappings,;
Each sequence vector is split as calling behavior and occurred before the preset time point by sequence vector split cells
The first behavior vector set, and call behavior that the second behavior vector set after the preset time point occurs;
Training unit, using the first behavior vector set in each behavior sequence as input, accordingly by each behavior sequence
Each behavior vector is as object vector, the training calling behavior prediction model in the second behavior vector set in column.
19. interface exception call monitoring device according to claim 15, which is characterized in that subclassification determining module, also
Include:
Model label generates model foundation unit, establishes the model label based on deep learning and generates model;
Model label generates model training unit, using the marked truth labels and the model label as one group of training
Collection, the training model label generate model.
20. interface exception call monitoring device according to claim 14, which is characterized in that the subclassification pre-sets mould
Root tuber carries out business classification to user according to the history service data of all users and generates multiple subclassifications, wherein each described
Subclassification corresponds to a service label.
21. interface exception call monitoring device according to claim 20, which is characterized in that the subclassification pre-sets mould
Block carries out clustering processing to the history service data of all users, obtains the corresponding subclassification.
22. interface exception call monitoring device according to claim 12, which is characterized in that the prediction interface calls row
It include the probability value that predictive behavior type and each predictive behavior type are called, the abnormal conditions monitoring modular packet for data
It includes:
Predictive behavior sequence forms unit, is ranked up from big to small to each predictive behavior type according to probability value, is formed pre-
Survey behavior sequence;
Further interface calls behavior acquiring unit, and the interface for obtaining the setting quantity of user's subsequent execution calls behavior;
Behavior normal judgment unit judges that the interface obtained calls whether behavior is in setting for the predictive behavior sequence
Before positioning is set, if being judged as YES, it is determined that user's further interface calls behavior normal.
23. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor
Calculation machine program, which is characterized in that the processor realizes the described in any item sides of claim 1 to 11 when executing described program
Method.
24. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program quilt
Claim 1 to 11 described in any item methods are realized when processor executes.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910795654.0A CN110489307B (en) | 2019-08-27 | 2019-08-27 | Interface abnormal call monitoring method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910795654.0A CN110489307B (en) | 2019-08-27 | 2019-08-27 | Interface abnormal call monitoring method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110489307A true CN110489307A (en) | 2019-11-22 |
CN110489307B CN110489307B (en) | 2023-04-07 |
Family
ID=68554522
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910795654.0A Active CN110489307B (en) | 2019-08-27 | 2019-08-27 | Interface abnormal call monitoring method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110489307B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111901347A (en) * | 2020-07-29 | 2020-11-06 | 南方电网科学研究院有限责任公司 | Dynamic identity authentication method and device under zero trust |
CN112527601A (en) * | 2020-12-17 | 2021-03-19 | 航天信息股份有限公司 | Monitoring early warning method and device |
CN112989332A (en) * | 2021-04-08 | 2021-06-18 | 北京安天网络安全技术有限公司 | Abnormal user behavior detection method and device |
CN113537616A (en) * | 2021-07-28 | 2021-10-22 | 北京达佳互联信息技术有限公司 | Account prediction model training method and device, electronic equipment and storage medium |
CN115292081A (en) * | 2022-08-10 | 2022-11-04 | 朴道征信有限公司 | Information sending method, information sending device, electronic equipment, medium and computer program product |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150222504A1 (en) * | 2014-02-03 | 2015-08-06 | Apigee Corporation | System and method for monitoring and reporting data in api processing systems |
CN108846520A (en) * | 2018-06-22 | 2018-11-20 | 北京京东金融科技控股有限公司 | Overdue loan prediction technique, device and computer readable storage medium |
CN110166422A (en) * | 2019-04-01 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Domain name Activity recognition method, apparatus, readable storage medium storing program for executing and computer equipment |
-
2019
- 2019-08-27 CN CN201910795654.0A patent/CN110489307B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150222504A1 (en) * | 2014-02-03 | 2015-08-06 | Apigee Corporation | System and method for monitoring and reporting data in api processing systems |
CN108846520A (en) * | 2018-06-22 | 2018-11-20 | 北京京东金融科技控股有限公司 | Overdue loan prediction technique, device and computer readable storage medium |
CN110166422A (en) * | 2019-04-01 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Domain name Activity recognition method, apparatus, readable storage medium storing program for executing and computer equipment |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111901347A (en) * | 2020-07-29 | 2020-11-06 | 南方电网科学研究院有限责任公司 | Dynamic identity authentication method and device under zero trust |
CN112527601A (en) * | 2020-12-17 | 2021-03-19 | 航天信息股份有限公司 | Monitoring early warning method and device |
CN112989332A (en) * | 2021-04-08 | 2021-06-18 | 北京安天网络安全技术有限公司 | Abnormal user behavior detection method and device |
CN113537616A (en) * | 2021-07-28 | 2021-10-22 | 北京达佳互联信息技术有限公司 | Account prediction model training method and device, electronic equipment and storage medium |
CN115292081A (en) * | 2022-08-10 | 2022-11-04 | 朴道征信有限公司 | Information sending method, information sending device, electronic equipment, medium and computer program product |
CN115292081B (en) * | 2022-08-10 | 2023-10-20 | 朴道征信有限公司 | Information sending method, device, electronic equipment and medium |
Also Published As
Publication number | Publication date |
---|---|
CN110489307B (en) | 2023-04-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110489307A (en) | Interface exception call monitoring method and device | |
CN106570778B (en) | A kind of method that data integration based on big data is calculated with line loss analyzing | |
CN106844781B (en) | Data processing method and device | |
CN106919957B (en) | Method and device for processing data | |
CN108470022B (en) | Intelligent work order quality inspection method based on operation and maintenance management | |
CN105446988B (en) | The method and apparatus for predicting classification | |
CN109242135A (en) | A kind of model method for running, device and service server | |
CN108573355A (en) | The method, apparatus and service server of operation are replaced after model modification | |
CN109684052A (en) | Transaction analysis method, apparatus, equipment and storage medium | |
CN108255706A (en) | Edit methods, device, terminal device and the storage medium of automatic test script | |
CN107622326A (en) | User's classification, available resources Forecasting Methodology, device and equipment | |
CN109154938A (en) | Using discrete non-trace location data by the entity classification in digitized map | |
CN105867930B (en) | A kind of display methods of function calling relationship, device and terminal | |
CN103617146B (en) | A kind of machine learning method and device based on hardware resource consumption | |
CN109784352A (en) | A kind of method and apparatus for assessing disaggregated model | |
CN114169389A (en) | Class-expanded target detection model training method and storage device | |
CN109213831A (en) | Event detecting method and device calculate equipment and storage medium | |
CN115545103A (en) | Abnormal data identification method, label identification method and abnormal data identification device | |
CN111898247A (en) | Landslide displacement prediction method, equipment and storage medium | |
CN110335061A (en) | Trade mode portrait method for building up, device, medium and electronic equipment | |
CN110188322A (en) | A kind of wave-shape amplitude uncertainty determines method and system | |
CN110232131A (en) | Intention material searching method and device based on intention label | |
CN106294115B (en) | A kind of test method and device of application system migration | |
CN107256254A (en) | A kind of Industrial Cycle index acquisition methods, storage device and terminal | |
CN116229211A (en) | Sample generation method, model training method, object detection method and device |
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 |