Summary of the invention
The purpose of this specification one or more embodiment is to provide the statistical method and device of a kind of exception information, to
Abnormal data is intelligently analyzed in realization, to efficiently and accurately monitor the Exception Type of abnormal data.
In order to solve the above technical problems, this specification one or more embodiment is achieved in that
On the one hand, this specification one or more embodiment provides a kind of statistical method of exception information, comprising:
Obtain the network structure for the RBF nerve network built using sample data as input data;
Wherein, the network structure of the RBF neural includes input layer, output layer and hidden layer;The sample data includes sample
The sample abnormal class of abnormal data and the sample abnormal data;Each sample data corresponds to respective radial base letter
Number;
According to the mapping relations between the sample abnormal data and the sample abnormal class, the RBF nerve is determined
The parameter of network;
According to the network structure of the RBF neural, the RBF mind is determined using the input data and the parameter
Output data through network;And each sample abnormal class is sorted out according to the output data.
In one embodiment, the radial basis function RBF mind built using sample data as input data is being obtained
Before network structure through network, the method also includes:
Using the sample data as the input data, the network structure of the RBF neural is built.
In one embodiment, described using the sample data as the input data, build the RBF neural
Network structure, comprising:
According to the first quantity of the sample abnormal data and the second quantity of the sample abnormal class, institute is determined respectively
State the interstitial content of input layer, the output layer and the hidden layer;
The RBF neural is built according to the interstitial content of the input layer, the output layer and the hidden layer
Network structure.
In one embodiment, first quantity according to the sample abnormal data and the sample abnormal class
Second quantity determines the interstitial content of the input layer, the output layer and the hidden layer respectively, comprising:
Determine that the interstitial content of the input layer is equal to the value of first quantity, and, determine the node of the output layer
Number is equal to the value of second quantity.
In one embodiment, the parameter includes Basis Function Center;
Correspondingly, the parameter of the determination RBF neural, comprising:
Select multiple first sample abnormal datas as initial cluster center from each sample abnormal data;
It is according to the first distance between each sample abnormal data and the initial cluster center, each sample is different
Regular data is clustered respectively at least one cluster set;
The average value of each sample abnormal data in each cluster set is calculated, and is determined according to the average value
First cluster centre of each cluster set;
Judge whether the difference between first cluster centre and the initial cluster center is located within preset difference value;
If so, determining that the initial cluster center is the Basis Function Center;If it is not, then using first cluster centre as described in
Initial cluster center continues to calculate the difference, until the difference is located within the preset difference value.
In one embodiment, the parameter includes basic function variance;
Correspondingly, the parameter of the determination RBF neural, comprising:
Determine the maximum distance between each first sample abnormal data;
The basic function variance is calculated according to the interstitial content of the maximum distance and the hidden layer.
In one embodiment, the parameter includes the weight between the hidden layer and the output layer;
Correspondingly, the parameter of the determination RBF neural, comprising:
Calculate the second distance between each sample abnormal data and the Basis Function Center;
Interstitial content, the maximum distance and the second distance based on the hidden layer, utilize least square method meter
Calculate the weight between the hidden layer and the output layer.
In one embodiment, described to determine the defeated of the RBF neural using the input data and the parameter
Data out, comprising:
According to the sample abnormal data, the basic function variance and the second distance, each radial base letter is determined
Several values;
According to the value of the radial basis function and the weight, the output data of the RBF neural is determined.
It is in one embodiment, described that each sample abnormal class is sorted out according to the output data, comprising:
Each output data is mapped as data to be sorted out according to default mapping mode;
Whether it is identical data according to the data to be sorted out, each sample abnormal class is sorted out.
On the other hand, this specification one or more embodiment provides a kind of statistic device of exception information, comprising:
Module is obtained, for obtaining the radial basis function RBF nerve net built using sample data as input data
The network structure of network;Wherein, the network structure of the RBF neural includes input layer, output layer and hidden layer;The sample
Data include the sample abnormal class of sample abnormal data and the sample abnormal data;Each sample data is corresponding respective
Radial basis function;
First determining module, for being closed according to the mapping between the sample abnormal data and the sample abnormal class
System, determines the parameter of the RBF neural;
Determining and classifying module utilizes the input data and institute for the network structure according to the RBF neural
State the output data that parameter determines the RBF neural;And according to the output data to each sample abnormal class into
Row is sorted out.
In one embodiment, described device further include:
Module is built, for obtaining the radial basis function RBF nerve built using sample data as input data
Before the network structure of network, using the sample data as the input data, the network knot of the RBF neural is built
Structure.
In one embodiment, the module of building includes:
Determination unit, for according to the first quantity of the sample abnormal data and the second number of the sample abnormal class
Amount, determines the interstitial content of the input layer, the output layer and the hidden layer respectively;
Unit is built, it is described for being built according to the interstitial content of the input layer, the output layer and the hidden layer
The network structure of RBF neural.
In one embodiment, the determination unit is also used to:
Determine that the interstitial content of the input layer is equal to the value of first quantity, and, determine the node of the output layer
Number is equal to the value of second quantity.
In one embodiment, the parameter includes Basis Function Center;
Correspondingly, first determining module includes:
Selecting unit, for selecting multiple first sample abnormal datas as initial poly- from each sample abnormal data
Class center;
Cluster cell, for according to the first distance between each sample abnormal data and the initial cluster center,
Each sample abnormal data is clustered respectively at least one cluster set;
First computing unit, for calculating the average value of each sample abnormal data in each cluster set, and
The first cluster centre of each cluster set is determined according to the average value;
Judgement and determination unit, for judging that the difference between first cluster centre and the initial cluster center is
It is no to be located within preset difference value;If so, determining that the initial cluster center is the Basis Function Center;If it is not, then will be described
First cluster centre continues to calculate the difference as the initial cluster center, until the difference is located at the preset difference value
Within.
In one embodiment, the parameter includes basic function variance;
Correspondingly, first determining module includes:
First determination unit, for determining the maximum distance between each first sample abnormal data;
Second computing unit, for calculating the basic function according to the interstitial content of the maximum distance and the hidden layer
Variance.
In one embodiment, the parameter includes the weight between the hidden layer and the output layer;
Correspondingly, first determining module includes:
Third computing unit, for calculate between each sample abnormal data and the Basis Function Center second away from
From;
4th computing unit, for interstitial content, the maximum distance and the second distance based on the hidden layer,
The weight between the hidden layer and the output layer is calculated using least square method.
In one embodiment, the determination and classifying module include:
Second determination unit is used for according to the sample abnormal data, the basic function variance and the second distance, really
The value of fixed each radial basis function;
Third determination unit, for according to the radial basis function value and the weight, determine the RBF neural
Output data.
In one embodiment, the determination and classifying module include:
Map unit, for each output data to be mapped as data to be sorted out according to default mapping mode;
Sort out unit, for whether being identical data according to the data to be sorted out, to each sample abnormal class
Sorted out.
In another aspect, this specification one or more embodiment provides a kind of statistics equipment of exception information, comprising:
Processor;And
It is arranged to the memory of storage computer executable instructions, the executable instruction makes the place when executed
Manage device:
Obtain the network structure for the RBF nerve network built using sample data as input data;
Wherein, the network structure of the RBF neural includes input layer, output layer and hidden layer;The sample data includes sample
The sample abnormal class of abnormal data and the sample abnormal data;Each sample data corresponds to respective radial base letter
Number;
According to the mapping relations between the sample abnormal data and the sample abnormal class, the RBF nerve is determined
The parameter of network;
According to the network structure of the RBF neural, the RBF mind is determined using the input data and the parameter
Output data through network;And each sample abnormal class is sorted out according to the output data.
In another aspect, the embodiment of the present application provides a kind of storage medium, for storing computer executable instructions, it is described can
It executes instruction and realizes following below scheme when executed:
Obtain the network structure for the RBF nerve network built using sample data as input data;
Wherein, the network structure of the RBF neural includes input layer, output layer and hidden layer;The sample data includes sample
The sample abnormal class of abnormal data and the sample abnormal data;Each sample data corresponds to respective radial base letter
Number;
According to the mapping relations between the sample abnormal data and the sample abnormal class, the RBF nerve is determined
The parameter of network;
According to the network structure of the RBF neural, the RBF mind is determined using the input data and the parameter
Output data through network;And each sample abnormal class is sorted out according to the output data.
Using the technical solution of this specification one or more embodiment, by obtaining using sample data as input number
According to the network structure for the RBF nerve network built, and according to sample abnormal data and sample abnormal class it
Between mapping relations, determine the parameter of RBF neural, and then according to the network structure of RBF neural, and utilize input number
According to and parameter determine the output data of RBF neural, each sample abnormal class is sorted out according to output data.As it can be seen that
The technical solution data processing and data analysis capabilities powerful using computer system, can pass through the exception to abnormal data
Classification is sorted out to reach abnormal data management effect that is unified and standardizing, reflects to realize to the intelligence of abnormal class
Process is penetrated, the efficiency and accuracy of abnormal data analysis are improved, analyzing abnormal data for developer reduces cost.
Specific embodiment
This specification one or more embodiment provides the statistical method and device of a kind of exception information, to realize intelligence
Change ground analysis abnormal data, to efficiently and accurately monitor the Exception Type of abnormal data.
In order to make those skilled in the art more fully understand the technical solution in this specification one or more embodiment,
Below in conjunction with the attached drawing in this specification one or more embodiment, to the technology in this specification one or more embodiment
Scheme is clearly and completely described, it is clear that and described embodiment is only this specification a part of the embodiment, rather than
Whole embodiments.Based on this specification one or more embodiment, those of ordinary skill in the art are not making creativeness
The model of this specification one or more embodiment protection all should belong in every other embodiment obtained under the premise of labour
It encloses.
Fig. 1 is according to a kind of schematic flow chart of the statistical method of exception information of one embodiment of this specification, such as Fig. 1
It is shown, this method comprises:
S102 obtains the network for the RBF nerve network built using sample data as input data
Structure.
Wherein, the network structure of RBF neural includes input layer, output layer and hidden layer;Sample data includes sample
The sample abnormal class of abnormal data and sample abnormal data;Each sample data corresponds to respective radial basis function.
S104 determines the ginseng of RBF neural according to the mapping relations between sample abnormal data and sample abnormal class
Number.
Wherein, the parameter of RBF neural includes between Basis Function Center, basic function variance and hidden layer and output layer
Weight.
S106 determines the defeated of RBF neural using input data and parameter according to the network structure of RBF neural
Data out;And each sample abnormal class is sorted out according to output data.
Using the technical solution of this specification one or more embodiment, by obtaining using sample data as input number
According to the network structure for the RBF nerve network built, and according to sample abnormal data and sample abnormal class it
Between mapping relations, determine the parameter of RBF neural, and then according to the network structure of RBF neural, and utilize input number
According to and parameter determine the output data of RBF neural, each sample abnormal class is sorted out according to output data.As it can be seen that
The technical solution data processing and data analysis capabilities powerful using computer system, can pass through the exception to abnormal data
Classification is sorted out to reach abnormal data management effect that is unified and standardizing, reflects to realize to the intelligence of abnormal class
Process is penetrated, the efficiency and accuracy of abnormal data analysis are improved, analyzing abnormal data for developer reduces cost.
The statistical method of exception information provided by the above embodiment described further below.In the above-described embodiments, radial base
Function can be Gaussian function, be also possible to other kinds of basic function.Following each embodiments are Gauss with radial basis function
Function is illustrated.
In one embodiment, the radial basis function RBF mind built using sample data as input data is being obtained
Before network structure through network, the network structure of RBF neural need to be built using sample data as input data.Assuming that institute
The network structure for the RBF neural built shares n-layer, then the network structure of RBF neural includes 1 input layer, 1
Output layer and n-2 hidden layer.Wherein, the value of n is unfettered, can be carried out that ginseng is adjusted to obtain according to actual scene.
In the present embodiment, since the network structure of RBF neural includes input layer, output layer and hidden layer,
When building the network structure of RBF neural, the interstitial content for determining input layer, output layer and hidden layer is first had to.
In one embodiment, can according to the first quantity of sample abnormal data and the second quantity of sample abnormal class,
The interstitial content of input layer, output layer and hidden layer is determined respectively, and then according to the number of nodes of input layer, output layer and hidden layer
Mesh builds the network structure of RBF neural.
Specifically, can determine that the interstitial content of input layer is equal to the value of the first quantity, the i.e. quantity of sample abnormal data;And
Determine that the interstitial content of output layer is equal to the value of the second quantity, the i.e. quantity of sample abnormal class.Assuming that sample abnormal data has
30, and this 30 sample abnormal datas are corresponding with 10 abnormal class altogether, then can determine that the interstitial content of input layer is 30,
The interstitial content of output layer is 10.The interstitial content of hidden layer can carry out that ginseng is adjusted to obtain according to actual scene.
After building the neural network of RBF neural, the parameter of RBF neural is determined.
In one embodiment, the Basis Function Center c of RBF neural can be sought according to K- means clustering method.Specifically
Referring to following steps A1-A4:
Step A1, select multiple first sample abnormal datas as initial cluster center from each sample abnormal data.
In the step, it is assumed that select m first sample abnormal data from each sample abnormal data as initial clustering
Center ci。
Step A2, according to the first distance between each sample abnormal data and initial cluster center, by each sample exception number
It is clustered in set according to being clustered respectively at least one.
In the step, Euclidean distance is can be used to characterize in first distance.The example above is continued to use, can be preset one or more
A distance range, and will be with initial cluster center ciBetween Euclidean distance be located at the sample abnormal data in same distance range
xpCluster is in same cluster set ViIn, and will be with initial cluster center ciBetween Euclidean distance be located within the scope of different distance
Sample abnormal data xpCluster is in different cluster set ViIn.
Step A3, the average value of each sample abnormal data in each cluster set is calculated, and each gather is determined according to average value
First cluster centre of class set.
Step A4, judge whether the difference between the first cluster centre and initial cluster center is located within preset difference value;
If so, determining that initial cluster center is Basis Function Center;If it is not, then continuing the first cluster centre as initial cluster center
Calculating difference, until difference is located within preset difference value.
The average value of each sample abnormal data in each cluster set can be considered new cluster centre, i.e., in the first cluster
The heart.Under normal conditions, to keep the Basis Function Center c of RBF neural more accurate, preset difference value may be set to sufficiently small
Value, i.e., close to 0.When the difference between the first cluster centre and initial cluster center is located within preset difference value, it can illustrate
The close enough initial cluster center of one cluster centre.
Whether it is located within preset difference value by comparing the difference between the first cluster centre and initial cluster center
Determine whether the first cluster centre changes, when the first cluster centre is no longer changed, obtained cluster centre
ciThe namely Basis Function Center c of RBF neural.
Continue to use the example above, it is assumed that initial cluster center c0, the first cluster centre is c1, judge c0With c1Between difference
Whether value is located within preset difference value, if c0With c1Between difference be located within preset difference value, then can determine c1For RBF nerve
The Basis Function Center c of network.If if c0With c1Between difference be not located within preset difference value, then by the first cluster centre c1Make
For initial cluster center, above-mentioned steps A2-A4 is continued to execute.Specifically, according to each sample abnormal data xpIn initial clustering
Heart c1Between Euclidean distance, by each sample abnormal data xpCluster clusters set V at least oneiIn, then, pass through calculating
Each cluster set ViIn each sample abnormal data xpAverage value come determine it is each cluster set the second cluster centre c2.In turn,
Judge the second cluster centre c2With initial cluster center c1Between difference whether be located within preset difference value, and according to judgement tie
Fruit determines whether the second cluster centre c2As Basis Function Center.
In one embodiment, when solving the basic function variance of RBF neural, it can first determine that each first sample is abnormal
Maximum distance between data, then according to the interstitial content of maximum distance and hidden layer between each first sample abnormal data
Calculate basic function variance.
Specifically, (1) the basic function variance δ of RBF neural can calculate according to the following formulai。
Wherein, cmaxFor each first sample abnormal data xpBetween maximum distance, h be hidden layer interstitial content.
In one embodiment, when solving the weight between hidden layer and output layer, each sample exception number can first be calculated
According to the second distance between Basis Function Center, it is then based on interstitial content, maximum distance and the second distance of hidden layer, and benefit
The weight between hidden layer and output layer is calculated with least square method.
Specifically, (2) weight w between hidden layer and output layer can calculate according to the following formula.
Wherein, cmaxFor each first sample abnormal data xpBetween maximum distance, h be hidden layer interstitial content, ciFor
The Basis Function Center of RBF neural.xp-ciFor each sample abnormal data xpWith Basis Function Center ciBetween second distance.
After the parameter for determining RBF neural, i.e., RBF is determined using the parameter of input data and RBF neural
The output data of neural network, and each sample abnormal class is sorted out according to output data.
In one embodiment, the output data of RBF neural is determined using input data and parameter, when, it can first root
Each radial direction is determined according to the second distance between sample abnormal data, basic function variance and each sample abnormal data and Basis Function Center
The value of basic function, and then RBF neural is determined according to the weight between the value and hidden layer and output layer of radial basis function
Output data.
Specifically, can first (3) determine the value R of each radial basis function according to the following formula.Wherein, each sample abnormal data
xpA corresponding radial basis function.
Wherein, δiFor the basic function variance of RBF neural, ciFor the Basis Function Center of RBF neural, xp-ciIt is each
Sample abnormal data xpWith Basis Function Center ciBetween second distance.
Then, (4) the output data y of RBF neural can be calculated according to the following formulaj。
Wherein, h is the interstitial content of hidden layer, wijWeight between hidden layer and output layer.
It, can be different to each sample according to the output data of RBF neural after the output data for determining RBF neural
Normal classification is sorted out.
It in one embodiment, can be according to default mapping mode by each output when sorting out to each sample abnormal class
Whether data are mapped as data to be sorted out, be then that identical data return each sample abnormal class according to data to be sorted out
Class.
Specifically, default mapping mode is following formula (5):
Wherein,For the data to be sorted out after mapping, yjFor the output data of RBF neural.
It, can be by identical data institute to be sorted out after output data is mapped as wait sort out data according to above-mentioned formula (5)
The corresponding sample abnormal class of corresponding output data is classified as one kind, by different wait sort out output data pair corresponding to data
The sample abnormal class answered sorts out inhomogeneity.In this way, multiple sample abnormal class can be classified as less classification, realize
The effect that abnormal data is counted and is managed.
In one embodiment, after each sample abnormal class being classified as target abnormal class, learning training can be passed through
Each sample abnormal data and its corresponding target abnormal class construct abnormal data statistical model, the abnormal data statistical model packet
Include the mapping relations between each abnormal data and target abnormal class.Wherein, specific learning training method can be used existing
Any learning training method, the present embodiment is not construed as limiting this.
After constructing abnormal data statistical model, i.e., the different of each abnormal data is analyzed using the abnormal data statistical model
Normal classification.Specifically, using abnormal data as the input data of abnormal data statistical model, by abnormal data statistical model,
The i.e. exportable abnormal data belongs to the probability of each abnormal class, and the abnormal class that wherein maximum probability may be selected is the exception number
According to corresponding abnormal class.
In the present embodiment, each exception is analyzed by constructing abnormal data statistical model, and using abnormal data statistical model
The abnormal class of data analyzes abnormal data saving to realize the effect for intelligently analyzing abnormal data for engineer
Cost.
To sum up, the specific embodiment of this theme is described.Other embodiments are in the appended claims
In range.In some cases, the movement recorded in detail in the claims can execute and still in a different order
Desired result may be implemented.In addition, process depicted in the drawing not necessarily requires the particular order shown or continuous suitable
Sequence, to realize desired result.In some embodiments, multitasking and parallel processing can be advantageous.
The above are the statistical methods for the exception information that this specification one or more embodiment provides, and are thought based on same
Road, this specification one or more embodiment also provide a kind of statistic device of exception information.
Fig. 2 is the schematic block diagram according to a kind of statistic device of exception information of one embodiment of this specification.Such as Fig. 2 institute
Show, the statistic device 200 of exception information includes:
Module 200 is obtained, for obtaining the radial basis function RBF nerve built using sample data as input data
The network structure of network;Wherein, the network structure of RBF neural includes input layer, output layer and hidden layer;Sample data packet
Include the sample abnormal class of sample abnormal data and sample abnormal data;Each sample data corresponds to respective radial basis function;
First determining module 210, for determining according to the mapping relations between sample abnormal data and sample abnormal class
The parameter of RBF neural;
Determining and classifying module 220, it is true using input data and parameter for the network structure according to RBF neural
Determine the output data of RBF neural;And each sample abnormal class is sorted out according to output data.
In one embodiment, device 200 further include:
Module is built, for obtaining the radial basis function RBF nerve built using sample data as input data
Before the network structure of network, using sample data as input data, the network structure of RBF neural is built.
In one embodiment, building module includes:
Determination unit, for according to the first quantity of sample abnormal data and the second quantity of sample abnormal class, difference
Determine the interstitial content of input layer, output layer and hidden layer;
Unit is built, for building the network of RBF neural according to the interstitial content of input layer, output layer and hidden layer
Structure.
In one embodiment, determination unit is also used to:
Determine that the interstitial content of input layer is equal to the value of the first quantity, and, determine that the interstitial content of output layer is equal to second
The value of quantity.
In one embodiment, parameter includes Basis Function Center;
Correspondingly, the first determining module 210 includes:
Selecting unit, for selecting multiple first sample abnormal datas as in initial clustering from each sample abnormal data
The heart;
Cluster cell, for according to the first distance between each sample abnormal data and initial cluster center, by each sample
Abnormal data is clustered respectively at least one cluster set;
First computing unit, for calculating the average value of each sample abnormal data in each cluster set, and according to average
Value determines the first cluster centre of each cluster set;
Judgement and determination unit, it is pre- whether the difference for judging between the first cluster centre and initial cluster center is located at
If within difference;If so, determining that initial cluster center is Basis Function Center;If it is not, then using the first cluster centre as initial
Cluster centre continues calculating difference, until difference is located within preset difference value.
In one embodiment, parameter includes basic function variance;
Correspondingly, the first determining module 210 includes:
First determination unit, for determining the maximum distance between each first sample abnormal data;
Second computing unit, for calculating basic function variance according to the interstitial content of maximum distance and hidden layer.
In one embodiment, parameter includes the weight between hidden layer and output layer;
Correspondingly, the first determining module 210 includes:
Third computing unit, for calculating the second distance between each sample abnormal data and Basis Function Center;
4th computing unit utilizes least square for interstitial content, maximum distance and second distance based on hidden layer
Method calculates the weight between hidden layer and output layer.
In one embodiment, determining and classifying module 220 includes:
Second determination unit, for determining each radial base letter according to sample abnormal data, basic function variance and second distance
Several values;
Third determination unit determines the output data of RBF neural for the value and weight according to radial basis function.
In one embodiment, determining and classifying module 220 includes:
Map unit, for each output data to be mapped as data to be sorted out according to default mapping mode;
Sort out unit, for whether being identical data according to data to be sorted out, each sample abnormal class is sorted out.
Using the device of this specification one or more embodiment, by obtaining using sample data as input data institute
The network structure for the RBF nerve network built, and according between sample abnormal data and sample abnormal class
Mapping relations, determine the parameter of RBF neural, and then according to the network structure of RBF neural, and using input data and
Parameter determines the output data of RBF neural, is sorted out according to output data to each sample abnormal class.As it can be seen that the skill
Art the scheme data processing and data analysis capabilities powerful using computer system, can pass through the abnormal class to abnormal data
Sorted out to reach abnormal data management effect that is unified and standardizing, the intelligence of abnormal class was mapped to realize
Journey improves the efficiency and accuracy of abnormal data analysis, and analyzing abnormal data for developer reduces cost.
It should be understood that the statistic device of above-mentioned exception information can be previously described for realizing
The statistical method of exception information, datail description therein should be described with method part above it is similar, it is cumbersome to avoid, it is not another herein
It repeats.
Based on same thinking, this specification one or more embodiment also provides a kind of statistics equipment of exception information,
As shown in Figure 3.The statistics equipment of exception information can generate bigger difference because configuration or performance are different, may include one
Or more than one processor 301 and memory 302, it can store one or more storage applications in memory 302
Program or data.Wherein, memory 302 can be of short duration storage or persistent storage.The application program for being stored in memory 302 can
To include one or more modules (diagram is not shown), each module may include in the statistics equipment to exception information
Series of computation machine executable instruction.Further, processor 301 can be set to communicate with memory 302, in abnormal letter
The series of computation machine executable instruction in memory 302 is executed in the statistics equipment of breath.The statistics equipment of exception information may be used also
To include one or more power supplys 303, one or more wired or wireless network interfaces 304, one or one with
Upper input/output interface 305, one or more keyboards 306.
Specifically in the present embodiment, the statistics equipment of exception information includes memory and one or more
Program, perhaps more than one program is stored in memory and one or more than one program may include one for one of them
A or more than one module, and each module may include that series of computation machine in statistics equipment to exception information is executable
Instruction, and be configured to execute this or more than one program by one or more than one processor to include for carrying out
Following computer executable instructions:
Obtain the network structure for the RBF nerve network built using sample data as input data;
Wherein, the network structure of the RBF neural includes input layer, output layer and hidden layer;The sample data includes sample
The sample abnormal class of abnormal data and the sample abnormal data;Each sample data corresponds to respective radial base letter
Number;
According to the mapping relations between the sample abnormal data and the sample abnormal class, the RBF nerve is determined
The parameter of network;
According to the network structure of the RBF neural, the RBF mind is determined using the input data and the parameter
Output data through network;And each sample abnormal class is sorted out according to the output data.
Optionally, computer executable instructions when executed, can also make the processor:
In the network structure for obtaining the RBF nerve network built using sample data as input data
Before, using the sample data as the input data, the network structure of the RBF neural is built.
Optionally, computer executable instructions when executed, can also make the processor:
According to the first quantity of the sample abnormal data and the second quantity of the sample abnormal class, institute is determined respectively
State the interstitial content of input layer, the output layer and the hidden layer;
The RBF neural is built according to the interstitial content of the input layer, the output layer and the hidden layer
Network structure.
Optionally, computer executable instructions when executed, can also make the processor:
Determine that the interstitial content of the input layer is equal to the value of first quantity, and, determine the node of the output layer
Number is equal to the value of second quantity.
Optionally, the parameter includes Basis Function Center;
Correspondingly, computer executable instructions are when executed, the processor can also be made:
Select multiple first sample abnormal datas as initial cluster center from each sample abnormal data;
It is according to the first distance between each sample abnormal data and the initial cluster center, each sample is different
Regular data is clustered respectively at least one cluster set;
The average value of each sample abnormal data in each cluster set is calculated, and is determined according to the average value
First cluster centre of each cluster set;
Judge whether the difference between first cluster centre and the initial cluster center is located within preset difference value;
If so, determining that the initial cluster center is the Basis Function Center;If it is not, then using first cluster centre as described in
Initial cluster center continues to calculate the difference, until the difference is located within the preset difference value.
Optionally, the parameter includes basic function variance;
Correspondingly, computer executable instructions are when executed, the processor can also be made:
Determine the maximum distance between each first sample abnormal data;
The basic function variance is calculated according to the interstitial content of the maximum distance and the hidden layer.
Optionally, the parameter includes the weight between the hidden layer and the output layer;
Correspondingly, computer executable instructions are when executed, the processor can also be made:
Calculate the second distance between each sample abnormal data and the Basis Function Center;
Interstitial content, the maximum distance and the second distance based on the hidden layer, utilize least square method meter
Calculate the weight between the hidden layer and the output layer.
Optionally, computer executable instructions when executed, can also make the processor:
According to the sample abnormal data, the basic function variance and the second distance, each radial base letter is determined
Several values;
According to the value of the radial basis function and the weight, the output data of the RBF neural is determined.
Optionally, computer executable instructions when executed, can also make the processor:
Each output data is mapped as data to be sorted out according to default mapping mode;
Whether it is identical data according to the data to be sorted out, each sample abnormal class is sorted out.
This specification one or more embodiment also proposed a kind of computer readable storage medium, this is computer-readable to deposit
Storage media stores one or more programs, which includes instruction, and it is included multiple application programs which, which works as,
Electronic equipment when executing, the electronic equipment can be made to execute the statistical method of above-mentioned exception information, and be specifically used for executing:
Obtain the network structure for the RBF nerve network built using sample data as input data;
Wherein, the network structure of the RBF neural includes input layer, output layer and hidden layer;The sample data includes sample
The sample abnormal class of abnormal data and the sample abnormal data;Each sample data corresponds to respective radial base letter
Number;
According to the mapping relations between the sample abnormal data and the sample abnormal class, the RBF nerve is determined
The parameter of network;
According to the network structure of the RBF neural, the RBF mind is determined using the input data and the parameter
Output data through network;And each sample abnormal class is sorted out according to the output data.
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.Specifically, computer for example may be used
Think personal computer, laptop computer, cellular phone, camera phone, smart phone, personal digital assistant, media play
It is any in device, navigation equipment, electronic mail equipment, game console, tablet computer, wearable device or these equipment
The combination of equipment.
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 specification one or more embodiment.
It should be understood by those skilled in the art that, this specification one or more embodiment can provide for method, system or
Computer program product.Therefore, complete hardware embodiment can be used in this specification one or more embodiment, complete software is implemented
The form of example or embodiment combining software and hardware aspects.Moreover, this specification one or more embodiment can be used one
It is a or it is multiple wherein include computer usable program code computer-usable storage medium (including but not limited to disk storage
Device, CD-ROM, optical memory etc.) on the form of computer program product implemented.
This specification one or more embodiment is referring to according to the method for the embodiment of the present application, equipment (system) and meter
The flowchart and/or the block diagram of calculation machine program product describes.It should be understood that can be realized by computer program instructions flow chart and/
Or the combination of the process and/or box in each flow and/or block and flowchart and/or the block diagram in block diagram.It can
These computer program instructions are provided at general purpose computer, special purpose computer, Embedded Processor or other programmable datas
The processor of equipment is managed to generate a machine, so that holding by the processor of computer or other programmable data processing devices
Capable instruction generates for realizing in one or more flows of the flowchart and/or one or more blocks of the block diagram
The device of specified function.
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.
In a typical configuration, calculating equipment includes one or more processors (CPU), input/output interface, net
Network interface and memory.
Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/or
The forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable medium
Example.
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.
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.
This specification one or more embodiment can computer executable instructions it is general on
It hereinafter describes, such as program module.Generally, program module includes executing particular task or realization particular abstract data type
Routine, programs, objects, component, data structure etc..The application can also be practiced in a distributed computing environment, at these
In distributed computing environment, by executing task by the connected remote processing devices of communication network.In distributed computing
In environment, program module can be located 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 foregoing is merely this specification one or more embodiments, are not limited to this specification.For this
For the technical staff of field, this specification one or more embodiment can have various modifications and variations.It is all in this specification one
Any modification, equivalent replacement, improvement and so within the spirit and principle of a or multiple embodiments, should be included in this explanation
Within the scope of the claims of book one or more embodiment.