CN113392100A - System intelligent verification method, device and system based on particle swarm optimization neural network - Google Patents
System intelligent verification method, device and system based on particle swarm optimization neural network Download PDFInfo
- Publication number
- CN113392100A CN113392100A CN202110939821.1A CN202110939821A CN113392100A CN 113392100 A CN113392100 A CN 113392100A CN 202110939821 A CN202110939821 A CN 202110939821A CN 113392100 A CN113392100 A CN 113392100A
- Authority
- CN
- China
- Prior art keywords
- neural network
- particle
- model
- value
- network model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000002245 particle Substances 0.000 title claims abstract description 92
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 39
- 238000012795 verification Methods 0.000 title claims abstract description 27
- 238000005457 optimization Methods 0.000 title claims abstract description 24
- 230000003044 adaptive effect Effects 0.000 claims description 59
- 238000003062 neural network model Methods 0.000 claims description 51
- 238000012545 processing Methods 0.000 claims description 28
- 238000012549 training Methods 0.000 claims description 24
- 238000004590 computer program Methods 0.000 claims description 20
- 230000006870 function Effects 0.000 claims description 18
- 238000003860 storage Methods 0.000 claims description 13
- 238000010845 search algorithm Methods 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 abstract description 7
- 238000005516 engineering process Methods 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 12
- 230000008569 process Effects 0.000 description 7
- 230000009471 action Effects 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000013475 authorization Methods 0.000 description 2
- 238000007418 data mining Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000012550 audit Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 230000003340 mental effect Effects 0.000 description 1
- 239000000178 monomer Substances 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000002829 reductive effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a system intelligent checking method, a device and a system for optimizing a neural network based on a particle swarm optimization, wherein an optimal neural network topological structure is found in a self-adaptive mode based on the particle swarm optimization, the logical relation between system input and system output in historical data is learned through a neural network algorithm, then the system output is simulated through a neural network and compared with a real output value, and an intelligent testing technology of on-line data checking, unmanned participation intelligent testing and full link case checking is realized by setting whether the output of a confidence interval decision system is correct or not. The method is used for scenes such as full-link case intelligent verification, online database data intelligent verification, unmanned participation intelligent test and the like.
Description
Technical Field
The invention relates to the technical field of artificial intelligence and databases, in particular to a method, a device and a system for intelligently checking a system based on a particle swarm optimization neural network.
Background
With the coming of big data era, the upgrading of the interconnection and digital transformation of everything, the data mining and application trend to be intelligent, the application scenes based on the data are more and more, and the traditional method based on the data mining, cleaning and statistical analysis can not meet the intelligent requirements on the scenes.
Disclosure of Invention
Aiming at the defects, the technical problem to be solved by the invention is how to intelligently mine and apply data by means of an artificial intelligence technology, and the data is used for the decision of a new business scene.
Aiming at the defects, the invention aims to provide a system intelligent verification method, a system, electronic equipment, a computer storage medium and a program product for optimizing a neural network based on a particle swarm algorithm, which are applied to a server side and used for searching a neural network model topological structure most suitable for a service scene through a search algorithm; calculating parameters of a neural network model under a network model topological structure according to historical data and an error back propagation algorithm, wherein the parameters comprise weight values and bias values of all nodes in the neural network model, and training the model; using the trained model for checking in a new service scene; comparing the predicted value of the algorithm model with the actual service output value, judging whether the predicted value is in the confidence interval, and if so, checking the accuracy; otherwise, the verification fails.
Preferably, in the neural network model, the network topology structure is configured to include a plurality of network layers, the number of nodes in each layer is defined, and the neural network model topology structure algorithm is searched through a particle swarm algorithm to find the neural network model topology structure most suitable for the service scenario.
Preferably, the algorithm comprises the steps of:
s1, initializing particle groups, and endowing each particle with a random initial position and speed;
s2, calculating the adaptive value of each particle according to the fitness function;
s3, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical optimal position, and if the adaptive value of the current position is higher, updating the historical optimal position by using the current position;
s4, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the global optimal position, and if the adaptive value of the current position is higher, updating the global optimal position by using the current position;
s5, calculating and updating the speed and the position of each particle;
and S6, if the ending condition is not met, returning to S2, and if the ending condition is met, ending the algorithm, wherein the global optimal position is the global optimal solution.
Preferably, the training neural network model performs data processing through feature engineering after acquiring data, adjusts learning rate dynamics after selecting the model, and performs model training.
Preferably, the data processing includes data normalization processing for continuous values, discrete values, and enumerated values.
Preferably, the data processing includes data variance scaling processing for continuous values, discrete values, and enumerated values.
Preferably, after the neural network model is generated, the generated model is used for predicting an output value of a new service flow according to input, and when the output of the network model is different from the actual output value of the service, a confidence interval is set to make a decision on the same value of the model predicted value and the actual service value.
The invention provides a system intelligent checking system for optimizing a neural network based on a particle swarm optimization algorithm, which comprises at least one database and at least one server, wherein the parameter data of the RPC request of the database are logged in the RPC request, the parameter values of the RPC request are acquired through a log service, the server is acquired through DRC, offline data are acquired through directly linking the database, and a neural network model topological structure most suitable for a service scene is searched through a search algorithm; calculating parameters of a neural network model under a network model topological structure according to historical data and an error back propagation algorithm, wherein the parameters comprise weight values and bias values of all nodes in the neural network model, and training the model; using the trained model for checking in a new service scene; comparing the predicted value of the algorithm model with the actual service output value, judging whether the predicted value is in the confidence interval, and if so, checking the accuracy; otherwise, the verification fails.
Preferably, the algorithm comprises the following steps:
s1, initializing particle groups, and endowing each particle with a random initial position and speed;
s2, calculating the adaptive value of each particle according to the fitness function;
s3, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical optimal position, and if the adaptive value of the current position is higher, updating the historical optimal position by using the current position;
s4, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the global optimal position, and if the adaptive value of the current position is higher, updating the global optimal position by using the current position;
s5, calculating and updating the speed and the position of each particle;
and S6, if the ending condition is not met, returning to S2, and if the ending condition is met, ending the algorithm, wherein the global optimal position is the global optimal solution. .
Preferably, after the neural network model is trained to obtain data, data processing is carried out through feature engineering, learning rate dynamics is adjusted after the model is selected, model training is carried out, the data processing comprises data normalization or variance scaling processing on a continuous value, a discrete value and an enumerated value, after the neural network model is generated, a generated model is used for predicting an output value of new business flow according to input, and when the output of the network model is different from the actual output value of the business, a confidence interval is set to make a decision on a model predicted value and the actual same value of the business.
The present invention provides a computer readable storage medium having stored thereon a computer program/instructions, characterized in that the computer program/instructions, when executed by a processor, implement the steps of the above-mentioned method.
The present invention provides a computer program product comprising computer programs/instructions, characterized in that the computer programs/instructions, when executed by a processor, implement the steps of the above-mentioned method.
The present invention provides an electronic device, including:
a processor; and
a memory arranged to store computer-executable instructions that, when executed, cause the processor to:
searching a neural network model topological structure most suitable for a service scene through a search algorithm; calculating parameters of a neural network model under a network model topological structure according to historical data and an error back propagation algorithm, wherein the parameters comprise weight values and bias values of all nodes in the neural network model, and training the model; using the trained model for checking in a new service scene; comparing the predicted value of the algorithm model with the actual service output value, judging whether the predicted value is in the confidence interval, and if so, checking the accuracy; otherwise, the verification fails;
the algorithm comprises the following steps:
s1, initializing particle groups, and endowing each particle with a random initial position and speed;
s2, calculating the adaptive value of each particle according to the fitness function;
s3, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical optimal position, and if the adaptive value of the current position is higher, updating the historical optimal position by using the current position;
s4, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the global optimal position, and if the adaptive value of the current position is higher, updating the global optimal position by using the current position;
s5, calculating and updating the speed and the position of each particle;
and S6, if the ending condition is not met, returning to S2, and if the ending condition is met, ending the algorithm, wherein the global optimal position is the global optimal solution.
Compared with the prior art, the scheme of the invention has the advantages of realizing flow automation, and having higher algorithm recall rate and accuracy rate, overcomes the NP-hard problem generated by directly searching the neural network topological structure, and has good adaptability in the current business practice.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments of the present invention will be briefly described below, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic structural diagram of an embodiment of a system intelligent verification method for optimizing a neural network based on a particle swarm optimization algorithm according to the present invention;
FIG. 2 is a schematic structural diagram of another embodiment of the intelligent verification method for a system based on particle swarm optimization neural network;
FIG. 3 is a schematic structural diagram of another embodiment of the intelligent verification method for a system based on particle swarm optimization neural network;
FIG. 4 is a schematic flow chart illustrating an embodiment of the intelligent verification method for a system based on particle swarm optimization neural network;
FIG. 5 is a schematic flow chart of another embodiment of the intelligent verification method for a system based on particle swarm optimization neural network according to the present invention;
FIG. 6 shows a schematic algorithm flow diagram in the system intelligent verification method for optimizing a neural network based on a particle swarm optimization.
Detailed Description
Features and exemplary embodiments of various aspects of the present invention will be described in detail below, and in order to make objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not to be construed as limiting the invention. It will be apparent to one skilled in the art that the present invention may be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the present invention by illustrating examples of the present invention.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
As shown in fig. 1 to 4, an embodiment of the present specification provides a system intelligent verification method for optimizing a neural network based on a particle swarm optimization, including:
step one, searching a neural network model topological structure which is most suitable for a service scene through a particle swarm search algorithm;
and step two, calculating parameters (weight values and offset values of all nodes in the network) of the neural network model under the network model topological structure in the first step according to the error back propagation algorithm through historical data, and finishing the training of the model.
And step three, using the trained model for checking in a new service scene. Comparing the predicted value of the algorithm model with the actual service output value, judging whether the predicted value is in the confidence interval, and if so, checking the accuracy; otherwise, the verification fails.
In some embodiments, step one is specifically:
in the neural network model, the network topology may be defined by the number of layers of the network (set as n), and the number of nodes in each layer (set as s _1 for the first layer, s _2 for the second layer, and s _ n for the nth layer). Therefore, for a neural network with n layers, the algorithm for searching the neural network model topological structure through the particle swarm optimization comprises the following steps:
1. initialization
Initializing a particle group (the particle group has m particles): each particle is assigned a random initial position and velocity. For example:
the particle position (8, 4, 16, 3) represents a neural network structure in which the number of nodes of the first layer is 8, the number of nodes of the second layer is 4, the number of nodes of the third layer is 16, and the number of nodes of the fourth layer is 3.
The particle position (5, 3, 32, 5) represents a neural network structure in which the number of nodes of the first layer is 5, the number of nodes of the second layer is 3, the number of nodes of the third layer is 32, and the number of nodes of the fourth layer is 5.
2. Calculating an adaptation value
Calculating the adaptive value of each particle according to the fitness function
The fitness value of the particle is represented by a loss function of the trained neural network:
for example: the neural network loss value for particle position (5, 3, 32, 5) was 0.006, which is superior to the (8, 4, 16, 3) particle.
3. Calculating the optimum adaptive value of each individual
For each particle, the fitness value for its current location is compared to the fitness value corresponding to the historical best location (pbest), and if the fitness value for the current location is higher, the historical best location is updated with the current location.
4. Calculating the optimum adaptive value of each individual
For each particle, the adapted value of its current position is compared with the adapted value corresponding to its global best position (gbest), and if the adapted value of the current position is higher, the global best position is updated with the current position.
5. Updating particle position and velocity
The velocity and position of each particle is updated according to the following formula
The current iteration speed = the last iteration speed + weight 1 (monomer history best position-current position) + weight 2 (population history best position-current position)
The position after the iteration = the current iteration position + the iteration speed.
The corresponding mathematical formula:
wherein,
in order to have a weight of 1, the weight,
in order to have the weight of 2,
in order to be the best position of the single body history,
for the best location of the group history,
for the position after the iteration of this time,
for the last iteration position (current position),
for the speed of the current iteration to be,
the last iteration speed.
6. Judging whether the algorithm is finished
And if the ending condition is not met, returning to the step 2, and if the ending condition is met, ending the algorithm, wherein the global optimal position (gbest) is the global optimal solution.
An embodiment of the present specification provides a system intelligent calibration method for optimizing a neural network based on a particle swarm algorithm, which is applied to a server side, and searches for a neural network model topological structure most suitable for a service scene through a search algorithm; calculating parameters of a neural network model under a network model topological structure according to historical data and an error back propagation algorithm, wherein the parameters comprise weight values and bias values of all nodes in the neural network model, and training the model; using the trained model for checking in a new service scene; comparing the predicted value of the algorithm model with the actual service output value, judging whether the predicted value is in the confidence interval, and if so, checking the accuracy; otherwise, the verification fails.
In some embodiments, in the neural network model, the network topology structure is configured to include a plurality of network layers, the number of nodes in each layer is defined, and the neural network model topology structure algorithm is searched by the particle swarm algorithm to find the neural network model topology structure most suitable for the service scenario.
In some embodiments, as shown in FIG. 6, the algorithm includes the following steps:
s1, initializing particle groups, and endowing each particle with a random initial position and speed;
s2, calculating the adaptive value of each particle according to the fitness function;
s3, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical optimal position, and if the adaptive value of the current position is higher, updating the historical optimal position by using the current position;
s4, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the global optimal position, and if the adaptive value of the current position is higher, updating the global optimal position by using the current position;
s5, calculating and updating the speed and the position of each particle;
and S6, if the ending condition is not met, returning to S2, and if the ending condition is met, ending the algorithm, wherein the global optimal position is the global optimal solution.
In some embodiments, the training neural network model performs data processing through feature engineering after acquiring data, adjusts learning rate dynamics after selecting a model, and performs model training.
In some embodiments, the data processing includes data normalization processing of continuous, discrete, enumerated values.
In some embodiments, the data processing includes data variance scaling of continuous, discrete, enumerated values.
In some embodiments, after the neural network model is generated, the generated model is used for predicting an output value of a new service flow according to input, and when the output of the network model is different from an actual service output value, a confidence interval is set to make a decision on the model prediction value and the actual service same value.
The invention provides a system intelligent checking system embodiment based on a particle swarm optimization neural network, which comprises at least one database and at least one server, wherein the parameter input and output data of an RPC request of the database are logged in the RPC request, then the parameter input and output values of the RPC request are obtained through a log service, the server is obtained through DRC, offline data are obtained through directly linking the database, and a neural network model topological structure most suitable for a service scene is searched through a search algorithm; calculating parameters of a neural network model under a network model topological structure according to historical data and an error back propagation algorithm, wherein the parameters comprise weight values and bias values of all nodes in the neural network model, and training the model; using the trained model for checking in a new service scene; comparing the predicted value of the algorithm model with the actual service output value, judging whether the predicted value is in the confidence interval, and if so, checking the accuracy; otherwise, the verification fails.
As shown in fig. 6, the algorithm in some embodiments includes the following steps:
s1, initializing particle groups, and endowing each particle with a random initial position and speed;
s2, calculating the adaptive value of each particle according to the fitness function;
s3, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical optimal position, and if the adaptive value of the current position is higher, updating the historical optimal position by using the current position;
s4, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the global optimal position, and if the adaptive value of the current position is higher, updating the global optimal position by using the current position;
s5, calculating and updating the speed and the position of each particle;
and S6, if the ending condition is not met, returning to S2, and if the ending condition is met, ending the algorithm, wherein the global optimal position is the global optimal solution. .
In some embodiments, training a neural network model to obtain data, performing data processing through feature engineering, selecting a model, adjusting learning rate dynamics, and performing model training, wherein the data processing includes performing data normalization or variance scaling on a continuous value, a discrete value, and an enumerated value, after the neural network model is generated, predicting an output value of a new service flow according to an input by using the generated model, and when an output of the network model is different from an actual output value of the service, deciding a model predicted value and the actual same value of the service by setting a confidence interval.
The present invention provides a computer-readable storage medium embodiment having stored thereon a computer program/instructions which, when executed by a processor, implement the operations of the above-described method.
The present invention provides a computer program product embodiment comprising computer programs/instructions which, when executed by a processor, implement the operations of the above-described methods.
The present invention provides an electronic device embodiment, comprising:
a processor; and
a memory arranged to store computer executable instructions that, when executed, cause the processor to perform the operations of the above-described method.
As shown in fig. 3, the present invention provides an embodiment of training a neural network model, wherein the data source is an ingress and egress parameter data of a RPC (Remote Procedure Call Protocol) request, and the ingress and egress parameter data of the RPC request is obtained by logging the RPC request and then obtaining the ingress and egress parameter of the RPC request through a Log Service (SLS). The DB data is retrieved by DRC (database bingo log based retrieval delta data) and the offline DB data is retrieved by direct linking to the database.
In some embodiments, the data includes an internet medical platform or a large medical database, such as a user's data of certain mental, psychological or other ailments. After the user authorization, the user forms a data source through data precipitation of interactive authorization through a PC or a mobile phone terminal.
As another example, in some financial scenarios, the data may be financial data, accounting data, tax data, audit data, and so forth.
In still other scenarios, the data is unstructured, such as image data, electrocardiogram data, etc., and the data needs to be structured, and some of the data are discrete and non-continuous.
Therefore, the invention is very necessary for data processing, and provides the data processing based on feature engineering, wherein the processed data comprises the following data:
1. continuous value of amount type
The value type field value is numerical value continuous, no special treatment is needed, and only normalization operation is needed before model training.
2. Discrete values of string type
Discrete values of string type, such as ip _ id, identification card, are discrete and not enumerable. A hash function is required to be designed to convert the character string into a specific numerical value. The size of the range of the hash function is determined by evaluating the possible number of strings. For example, the field of the identity card, assuming that the user is approximately 1 hundred million in magnitude, the hash function can be set to range from 0 to 1 hundred million.
3. State type enumeration value
For the state types, the values are discrete and can be enumerated, and only One-Hot coding (One-Hot) can be directly carried out. For example, the four states of initialization, in progress, success, failure can be coded as 0001, 0010, 0100, 1000.
4. Data normalization
The first method is a feature scaling method, where x is a feature value, and min (x) and max (x) are the minimum and maximum values of the feature in the data set, respectively, so the formula of the feature scaling method is as follows:
or a variance scaling method is adopted, and after variance scaling, the distribution of the data set is changed into the distribution with the mean value of 0 and the variance value of 1.
In the present embodiment, a variance scaling method is selected. The predicted values from the neural network are determined from the confidence interval and the actual values. The variance scaling method is beneficial to better distinguish confidence interval ranges between different values.
In some embodiments, as shown in fig. 5, after the model selection and the conversion of the business data into mathematical vectors, the training of the model is started after the historical data is divided into two groups, namely a training set and a test set. In different business scenarios, selecting different network models will increase the algorithm's progress. The present embodiment selects a feed-forward neural network to fit the business input to output relationship.
In some embodiments, the learning rate is dynamically adjusted, and after normalization of the input data, the initial learning rate is set to 0.1. Then, according to the number of iterations, setting n times of each iteration, the learning rate is reduced by 10 times (learning rate 0.1).
In some embodiments, the model is trained, and finally the model training is performed by an error back propagation algorithm.
In some embodiments, after the neural network model is generated by making a decision on the verification result by using the confidence interval, the generated model can be used to predict the output value of the new service flow according to the input. The output of the network model will be different from the actual output value of the service, for example, 35 is the actual service output, and 34.995 or 35.2 is the predicted output of the network model. At this time, the confidence interval is set to make a decision on the model predicted value and the actual same value of the service. The confidence interval can be set when the error between the two is less than 3%, the two are considered to be consistent, namely the service is checked correctly; otherwise, the service is considered to be in error.
The invention adaptively finds the optimal neural network topological structure based on a particle swarm search algorithm, learns the logical relation between system input and system output in historical data through a neural network algorithm, then simulates the system output through a neural network and compares the system output with a real output value, and realizes the intelligent test technology of on-line data check, intelligent test without human participation in a gray level environment and full link case check by setting whether the system output is correct or not in a confidence interval decision making system. Compared with the current academical method for solving the Test Oracle problem through the neural network, the method has the advantages of realizing flow automation, and being high in algorithm recall rate and accuracy rate, overcomes the NP-hard problem caused by direct search of the neural network topological structure, and has good adaptability in current business practice.
For convenience of description, the above devices are described as being divided into various units by function, and are described separately. Of course, the functionality of the units may be implemented in one or more software and/or hardware when implementing the present application.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (13)
1. A system intelligent calibration method based on particle swarm optimization neural network is applied to a server side, and a neural network model topological structure most suitable for a service scene is searched through a search algorithm; calculating parameters of a neural network model under a network model topological structure according to historical data and an error back propagation algorithm, wherein the parameters comprise weight values and bias values of all nodes in the neural network model, and training the model; using the trained model for checking in a new service scene; comparing the predicted value of the algorithm model with the actual service output value, judging whether the predicted value is in the confidence interval, and if so, checking the accuracy; otherwise, the verification fails.
2. The method for system intelligent calibration based on particle swarm optimization neural network of claim 1, wherein in the neural network model, the network topology is configured to include a plurality of network layers, the number of nodes in each layer is defined, and the neural network model topology is searched by the particle swarm optimization to find the neural network model topology most suitable for the service scenario.
3. The intelligent verification method for the system based on the particle swarm optimization neural network according to claim 1 or 2, wherein the algorithm comprises the following steps:
s1, initializing particle groups, and endowing each particle with a random initial position and speed;
s2, calculating the adaptive value of each particle according to the fitness function;
s3, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical optimal position, and if the adaptive value of the current position is higher, updating the historical optimal position by using the current position;
s4, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the global optimal position, and if the adaptive value of the current position is higher, updating the global optimal position by using the current position;
s5, calculating and updating the speed and the position of each particle;
and S6, if the ending condition is not met, returning to S2, and if the ending condition is met, ending the algorithm, wherein the global optimal position is the global optimal solution.
4. The intelligent system verification method based on particle swarm optimization neural network of claim 1 or 2, wherein the training of the neural network model is performed after data acquisition and data processing through feature engineering, and after the model is selected, the learning rate dynamics is adjusted, and then the model training is performed.
5. The particle swarm optimization neural network-based system intelligent verification method of claim 4, wherein the data processing comprises data normalization processing on continuous values, discrete values and enumerated values.
6. The particle swarm optimization neural network-based system intelligent verification method of claim 4, wherein the data processing comprises data variance scaling processing on continuous values, discrete values and enumerated values.
7. The system intelligent verification method based on particle swarm optimization neural network of claim 1 or 2, wherein after the neural network model is generated, the generated model is used for predicting the output value of the new service flow according to the input, and when the output of the network model is different from the actual output value of the service, the confidence interval is set to make a decision on the same value of the model predicted value and the actual service value.
8. A system intelligent calibration system based on a particle swarm optimization neural network comprises at least one database and at least one server, wherein the parameter input and output data of an RPC request of the database are logged in the RPC request, then the parameter input and output values of the RPC request are obtained through a log service, the server is obtained through DRC, offline data are obtained through a direct link database, and a neural network model topological structure most suitable for a service scene is searched through a search algorithm; calculating parameters of a neural network model under a network model topological structure according to historical data and an error back propagation algorithm, wherein the parameters comprise weight values and bias values of all nodes in the neural network model, and training the model; using the trained model for checking in a new service scene; comparing the predicted value of the algorithm model with the actual service output value, judging whether the predicted value is in the confidence interval, and if so, checking the accuracy; otherwise, the verification fails.
9. The system of claim 8, the algorithm comprising the steps of:
s1, initializing particle groups, and endowing each particle with a random initial position and speed;
s2, calculating the adaptive value of each particle according to the fitness function;
s3, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical optimal position, and if the adaptive value of the current position is higher, updating the historical optimal position by using the current position;
s4, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the global optimal position, and if the adaptive value of the current position is higher, updating the global optimal position by using the current position;
s5, calculating and updating the speed and the position of each particle;
and S6, if the ending condition is not met, returning to S2, and if the ending condition is met, ending the algorithm, wherein the global optimal position is the global optimal solution.
10. The system according to claim 8 or 9, wherein a neural network model is trained to perform data processing after acquiring data, a learning rate dynamics is adjusted after selecting a model, and model training is performed, wherein the data processing includes performing data normalization or variance scaling on a continuous value, a discrete value, and an enumerated value, after the neural network model is generated, a generated model is used to predict a new traffic flow according to an input to obtain an output value, and when an output of the network model is different from an actual traffic output value, a confidence interval is set to make a decision on a predicted value of the model and an actual traffic same value.
11. A computer-readable storage medium, on which a computer program/instructions are stored, characterized in that the computer program/instructions, when executed by a processor, implement the steps of the method according to one of claims 1 to 7.
12. A computer program product comprising computer programs/instructions, characterized in that the computer programs/instructions, when executed by a processor, implement the steps of the method according to one of claims 1 to 7.
13. An electronic device, comprising:
a processor; and
a memory arranged to store computer-executable instructions that, when executed, cause the processor to:
searching a neural network model topological structure most suitable for a service scene through a search algorithm; calculating parameters of a neural network model under a network model topological structure according to historical data and an error back propagation algorithm, wherein the parameters comprise weight values and bias values of all nodes in the neural network model, and training the model; using the trained model for checking in a new service scene; comparing the predicted value of the algorithm model with the actual service output value, judging whether the predicted value is in the confidence interval, and if so, checking the accuracy; otherwise, the verification fails;
the algorithm comprises the following steps:
s1, initializing particle groups, and endowing each particle with a random initial position and speed;
s2, calculating the adaptive value of each particle according to the fitness function;
s3, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical optimal position, and if the adaptive value of the current position is higher, updating the historical optimal position by using the current position;
s4, for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the global optimal position, and if the adaptive value of the current position is higher, updating the global optimal position by using the current position;
s5, calculating and updating the speed and the position of each particle;
and S6, if the ending condition is not met, returning to S2, and if the ending condition is met, ending the algorithm, wherein the global optimal position is the global optimal solution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110939821.1A CN113392100A (en) | 2021-08-17 | 2021-08-17 | System intelligent verification method, device and system based on particle swarm optimization neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110939821.1A CN113392100A (en) | 2021-08-17 | 2021-08-17 | System intelligent verification method, device and system based on particle swarm optimization neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113392100A true CN113392100A (en) | 2021-09-14 |
Family
ID=77622874
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110939821.1A Pending CN113392100A (en) | 2021-08-17 | 2021-08-17 | System intelligent verification method, device and system based on particle swarm optimization neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113392100A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114374636A (en) * | 2021-12-21 | 2022-04-19 | 航天科工网络信息发展有限公司 | Intelligent routing method, device and network equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108334059A (en) * | 2018-02-26 | 2018-07-27 | 上海理工大学 | Based on particle cluster algorithm Optimized BP Neural Network Symbolic fault diagnosis method |
| CN109146121A (en) * | 2018-06-25 | 2019-01-04 | 华北电力大学 | The power predicating method stopped in the case of limited production based on PSO-BP model |
| CN111210082A (en) * | 2020-01-13 | 2020-05-29 | 东南大学 | Optimized BP neural network algorithm-based precipitation prediction method |
| CN111882041A (en) * | 2020-07-31 | 2020-11-03 | 国网重庆市电力公司电力科学研究院 | Power grid attack detection method and device based on improved RNN (neural network) |
-
2021
- 2021-08-17 CN CN202110939821.1A patent/CN113392100A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108334059A (en) * | 2018-02-26 | 2018-07-27 | 上海理工大学 | Based on particle cluster algorithm Optimized BP Neural Network Symbolic fault diagnosis method |
| CN109146121A (en) * | 2018-06-25 | 2019-01-04 | 华北电力大学 | The power predicating method stopped in the case of limited production based on PSO-BP model |
| CN111210082A (en) * | 2020-01-13 | 2020-05-29 | 东南大学 | Optimized BP neural network algorithm-based precipitation prediction method |
| CN111882041A (en) * | 2020-07-31 | 2020-11-03 | 国网重庆市电力公司电力科学研究院 | Power grid attack detection method and device based on improved RNN (neural network) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114374636A (en) * | 2021-12-21 | 2022-04-19 | 航天科工网络信息发展有限公司 | Intelligent routing method, device and network equipment |
| CN114374636B (en) * | 2021-12-21 | 2024-04-02 | 航天科工网络信息发展有限公司 | Intelligent routing method, device and network equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109408731B (en) | Multi-target recommendation method, multi-target recommendation model generation method and device | |
| CN115130536A (en) | Training method of feature extraction model, data processing method, device and equipment | |
| KR102472800B1 (en) | Artificial intelligence-based enterprise rating method, device and system | |
| CN113379536A (en) | Default probability prediction method for optimizing recurrent neural network based on gravity search algorithm | |
| CN113342474A (en) | Method, device and storage medium for forecasting customer flow and training model | |
| CN114511063A (en) | Power data prediction model construction method, device, apparatus, medium, and program | |
| CN111461353A (en) | Model training method and system | |
| CN117668622B (en) | Training method of equipment fault diagnosis model, fault diagnosis method and device | |
| CN113392100A (en) | System intelligent verification method, device and system based on particle swarm optimization neural network | |
| CN114580546A (en) | Industrial pump fault prediction method and system based on federal learning framework | |
| US20220171985A1 (en) | Item recommendation with application to automated artificial intelligence | |
| CN112905987B (en) | Account identification method, device, server and storage medium | |
| CN114092162B (en) | Recommendation quality determination method, and training method and device of recommendation quality determination model | |
| CN114548382B (en) | Migration training method, device, equipment, storage medium and program product | |
| CN110889316A (en) | Target object identification method and device and storage medium | |
| CN116958608A (en) | Method, device, equipment, medium and program product for updating object recognition model | |
| CN117560162A (en) | Network security protection method, device, storage medium and apparatus | |
| CN115102868A (en) | Web service QoS prediction method based on SOM clustering and depth self-encoder | |
| CN113723593A (en) | Load shedding prediction method and system based on neural network | |
| CN114332480A (en) | Graph data classification model training method, device, equipment and storage medium | |
| CN113553270B (en) | Target object determining method and device | |
| CN111737319A (en) | User cluster prediction method and device, computer equipment and storage medium | |
| CN113537269B (en) | Image processing method, device and equipment | |
| Su et al. | Neural network based fusion of global and local information in predicting time series | |
| CN113988185B (en) | Data processing method and related 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 | ||
| CB02 | Change of applicant information |
Address after: Rooms 1601 and 1602, 16th floor, No. 52, North Fourth Ring West Road, Haidian District, Beijing 100080 Applicant after: Good Feeling Health Industry Group Co.,Ltd. Address before: Rooms 1601 and 1602, 16th floor, No. 52, North Fourth Ring West Road, Haidian District, Beijing 100080 Applicant before: BEIJING HAOXINQING MOBILE MEDICAL TECHNOLOGY CO.,LTD. |
|
| CB02 | Change of applicant information |