CN110046256A - The prediction technique and device of case differentiation result - Google Patents
The prediction technique and device of case differentiation result Download PDFInfo
- Publication number
- CN110046256A CN110046256A CN201910321775.1A CN201910321775A CN110046256A CN 110046256 A CN110046256 A CN 110046256A CN 201910321775 A CN201910321775 A CN 201910321775A CN 110046256 A CN110046256 A CN 110046256A
- Authority
- CN
- China
- Prior art keywords
- result
- data
- sample
- case
- obtains
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000004069 differentiation Effects 0.000 title claims abstract description 25
- 238000012417 linear regression Methods 0.000 claims abstract description 69
- 230000000192 social effect Effects 0.000 claims abstract description 31
- 238000012549 training Methods 0.000 claims description 11
- 238000010606 normalization Methods 0.000 claims description 8
- 238000012937 correction Methods 0.000 claims description 7
- 235000013399 edible fruits Nutrition 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 abstract description 6
- 238000012545 processing Methods 0.000 abstract description 6
- 238000013499 data model Methods 0.000 abstract description 5
- 238000005516 engineering process Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 12
- 238000010586 diagram Methods 0.000 description 10
- 230000008859 change Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/18—Legal services
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Technology Law (AREA)
- General Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides the prediction technique and device of a kind of case differentiation result, is related to big data processing technology field.The described method includes: obtaining the case data of case to be detected, the case data include fair data, just data and social effect data；The case data are inputted into strong classifier model and linear regression model (LRM) respectively, the first result that the input strong classifier model exports and second that the linear regression model (LRM) exports are respectively obtained as a result, first result and second result are the judgement result for the case to be detected that prediction obtains；It is analyzed according to first result and second result, obtains prediction result.It is analyzed by case data of the data model to case to be detected, and comprehensive analysis is carried out to the result of multiple models output, obtain integrating the prediction result of multiple results, the problem lower by the prediction result accuracy manually predicted is avoided, the accuracy that prediction case differentiates result is improved.
Description
Technical field
The present invention relates to big data processing technology fields, and the prediction technique of result is differentiated in particular to a kind of case
And device.
Background technique
As social economy continues to develop, law court is also more and more to the trial of the case of economic type.But to every
During the trial result of a case is predicted, the usually trial of judge or lawyer according to itself experience, to case
As a result it is predicted.But judge or lawyer be by the limitation of itself, can not Accurate Prediction obtain corresponding result.
Therefore, a kind of method for needing prediction case differentiation result, to improve the existing accuracy manually predicted.
Summary of the invention
It is an object of the present invention in view of the deficiency of the prior art, provide a kind of prediction of case differentiation result
Method and device, to solve the problems, such as that existing discriminant approach accuracy rate is lower.
To achieve the above object, technical solution used in the embodiment of the present invention is as follows:
In a first aspect, the embodiment of the invention provides the prediction techniques that a kind of case differentiates result, which comprises
The case data of case to be detected are obtained, the case data include fair data, just data and social effect
Data；
The case data are inputted into strong classifier model and linear regression model (LRM) respectively, respectively obtain strong point of the input
The second of the first result and linear regression model (LRM) output that class device model exports is as a result, first result and described second
Result is the judgement result for the case to be detected that prediction obtains；
It is analyzed according to first result and second result, obtains prediction result.
Optionally, before the case data for obtaining case to be detected, the method also includes:
Sample data after obtaining multiple sample case normalization, the sample data includes: sample justice data, sample
Just data, sample social effect data and practical differentiation result；
Pre-set initial strong classifier model is trained according to the sample data, obtains classifier；
According to the sample predictions of multiple sample cases of classifier output as a result, and multiple sample cases
Part it is corresponding it is practical differentiate as a result, be modified to the classifier, obtain the strong classifier model.
Optionally, the sample predictions of multiple sample cases according to classifier output are as a result, and more
A sample case it is corresponding it is practical differentiate as a result, be modified to the classifier, obtain the strong classifier model, wrap
It includes:
Multiple sample predictions results and corresponding practical differentiation result are compared, comparison result is obtained；
According to the comparison result, the loss function of the classifier is calculated；
The classifier is modified according to the loss function, obtains the strong classifier model.
Optionally, before the case data for obtaining case to be detected, the method also includes:
Sample data after obtaining multiple sample case normalization, the sample data includes: sample justice data, sample
Just data, sample social effect data and practical differentiation result；
Multiple initial load factors are determined according to the sample data；
Multiple initial load factors are trained by the sample data, obtain the linear regression model (LRM).
Optionally, described to be analyzed according to first result and second result, obtain prediction result, comprising:
According to corresponding first weight of first result and corresponding second weight of second result, in conjunction with described
One result and second result are calculated, and the prediction result is obtained.
Second aspect, the embodiment of the invention also provides the prediction meanss that a kind of case differentiates result, described device includes:
First obtains module, and for obtaining the case data of case to be detected, the case data include fair data, public affairs
Correction data and social effect data；
Input module, for the case data to be inputted strong classifier model and linear regression model (LRM) respectively, respectively
The second of the first result and linear regression model (LRM) output that export to the input strong classifier model is as a result, described first
It as a result is the judgement result for the case to be detected that prediction obtains with second result；
Determining module obtains prediction result for being analyzed according to first result and second result.
Optionally, described device further include:
Second obtains module, and for obtaining the sample data after multiple sample cases normalize, the sample data includes:
Sample justice data, the just data of sample, sample social effect data and practical differentiation result；
First training module, for being instructed according to the sample data to pre-set initial strong classifier model
Practice, obtains classifier；
Correction module, the sample predictions of multiple sample cases for being exported according to the classifier as a result, and
Multiple sample cases it is corresponding it is practical differentiate as a result, be modified to the classifier, obtain the strong classifier model.
Optionally, the correction module is specifically used for multiple sample predictions results and corresponding practical differentiation
As a result it is compared, obtains comparison result；According to the comparison result, the loss function of the classifier is calculated；According to described
Loss function is modified the classifier, obtains the strong classifier model.
Optionally, described device further include:
Third obtains module, and for obtaining the sample data after multiple sample cases normalize, the sample data includes:
Sample justice data, the just data of sample, sample social effect data and practical differentiation result；
Model building module, for determining multiple initial load factors according to the sample data；
Second training module is obtained for being trained by the sample data to multiple initial load factors
The linear regression model (LRM).
Optionally, the determining module is specifically used for according to corresponding first weight of first result and described second
As a result corresponding second weight is calculated in conjunction with first result and second result, obtains the prediction result.
The beneficial effects of the present invention are:
Case provided by the invention differentiates the prediction technique and device of result, includes fair number by obtaining case to be detected
According to, the case data of just data and social effect data, case data are inputted into strong classifier model and linear regression respectively
Model respectively obtains the first result that input strong classifier model exports and second that linear regression model (LRM) exports as a result, last
It is analyzed according to the first result and the second result, obtains prediction result.By data model to the case number of packages of case to be detected
Comprehensive analysis is carried out according to being analyzed, and to the result of multiple models output, the prediction result for integrating multiple results is obtained, avoids
The problem lower by the prediction result accuracy manually predicted improves the accuracy that prediction case differentiates result.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 is the flow diagram for the prediction technique that the case that one embodiment of the invention provides differentiates result；
Fig. 2 be another embodiment of the present invention provides case differentiate result prediction technique flow diagram；
Fig. 3 is the schematic diagram for the prediction meanss that the case that one embodiment of the invention provides differentiates result；
Fig. 4 be another embodiment of the present invention provides case differentiate result prediction meanss schematic diagram；
Fig. 5 is the schematic diagram for the prediction meanss that the case that further embodiment of this invention provides differentiates result；
Fig. 6 is the schematic diagram for the prediction meanss that the case that one embodiment of the invention provides differentiates result.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.
Fig. 1 is the flow diagram for the prediction technique that the case that one embodiment of the invention provides differentiates result, such as Fig. 1 institute
Show, this method comprises:
Step 101, the case data for obtaining case to be detected.
Wherein, which may include fair data, just data and social effect data, which is used for
Indicate the fair degree of decision in a case, which is used to indicate the just degree of decision in a case, and the social effect data
For indicating positive social effect or negative social effect caused by case.
In order to improve the accuracy that prediction case differentiates result, case can be carried out by model trained in advance pre-
It surveys.Since model is predicted by the case data of case, it can first obtain the case number of packages of case to be detected
According in the next steps, to be predicted according to the case data of acquisition case to be detected.
It, therefore, can be with due to during hearing and decide a case, being related to fair standard, just index and social effect index
It is fair data, just data and social effect data by the case data quantization of case to be detected.
For example, the value range of fair data, just data and social effect data is 0 to 100, if value is 0,
Indicate case trial pole do not meet the principles such as fair and just, if but value be 100, illustrate case trial comply fully with justice
The principles such as just.
It should be noted that the technical specialist that the case data of case to be detected can be related fields evaluates to obtain,
It can obtain by other means, it is not limited in the embodiment of the present invention.
Case data are inputted strong classifier model and linear regression model (LRM) by step 102 respectively, and it is strong to respectively obtain the input
First result of sorter model output and the second result of linear regression model (LRM) output.
Wherein, first result and second result are the judgement result for the case to be detected that prediction obtains.
In order to improve the accuracy of prediction result, case data can be inputted to different models, then available difference
Model output as a result, so as in the next steps, comprehensive analysis can be carried out according to each result, obtain result to the end.
Specifically, case data can be inputted to strong classifier model and linear regression model (LRM), strong classifier model respectively
Then can the weight according to corresponding to each Weak Classifier for including, to case data carry out analytical calculation, in conjunction with each weak point
The result of class device obtains the first result.
And linear regression model (LRM) then can be according to pre-set calculation formula, and combines the corresponding instruction of different case data
The load factor got is calculated, to obtain calculated result, calculated result is finally determined according to mode classification corresponding to
Classification, to obtain the second result.
It should be noted that in practical applications, case data can also be inputted to other and train obtained model, this hair
Bright embodiment to the model for predicting case result without limitation.
Step 103 is analyzed according to the first result and the second result, obtains prediction result.
It, can be according to the first result and the second prediction of result case to be detected after obtaining the first result and the second result
Differentiation result.But in order to improve the accuracy that prediction differentiates result, the first result and the second result can be carried out again
Analytical calculation, to obtain integrating the prediction result of each model output result.
For example, can be according to the first result and the corresponding weight of the second result, to the first result and the second result point
Not corresponding parameter value is calculated, the parameter value after being calculated, and determines case to be detected further according to the parameter value being calculated
The prediction result of part.
In conclusion case provided by the invention differentiates the prediction technique of result, it include public affairs by obtaining case to be detected
Case data are inputted strong classifier model and linear by the case data of flat data, just data and social effect data respectively
Regression model, respectively obtain input strong classifier model output the first result and linear regression model (LRM) output second as a result,
It is finally analyzed according to the first result and the second result, obtains prediction result.By data model to the case of case to be detected
Number of packages carries out comprehensive analysis according to being analyzed, and to the result of multiple models output, obtains the prediction result for integrating multiple results,
The problem lower by the prediction result accuracy manually predicted is avoided, prediction case is improved and differentiates the accurate of result
Property.
Fig. 2 be another embodiment of the present invention provides case differentiate result prediction technique flow diagram, such as Fig. 2 institute
Show, this method comprises:
Step 201 is trained according to sample data, obtains strong classifier model and linear regression model (LRM).
In order to accurately be predicted the case not yet adjudicated in advance, sample data can be trained, be obtained more
The disaggregated model of a type, so as in the next steps, each result that can be obtained according to training.
Since different training method training can be used to obtain different disaggregated models.Therefore, this step 201 can wrap
Step 201a and step 201b are included, is respectively trained to obtain strong classifier model and linear regression model (LRM).
Step 201a, the sample data after obtaining the normalization of multiple sample cases, and according to sample data to presetting
Initial strong classifier model be trained, obtain classifier, further according to classifier output multiple sample cases sample it is pre-
Survey as a result, and multiple sample cases it is corresponding it is practical differentiate as a result, be modified to the classifier, obtain strong classifier mould
Type.
Wherein, sample data may include: sample justice data, the just data of sample, sample social effect data and reality
Border differentiates result.
Moreover, what the technical specialist that the corresponding case data of each sample case can be related fields evaluated,
It can obtain by other means, it is not limited in the embodiment of the present invention.
For example, the value range of sample justice data, the just data of sample and sample social effect data is 0 to 100,
If value be 0, then it represents that case trial pole do not meet the principles such as fair and just, if but value be 100, illustrate case try
Comply fully with the principles such as fair and just.
Specifically, can first classify to the different types of data in each sample data of acquisition, then by same class
Each data of type are normalized, the sample data after being normalized, then to the corresponding case data of each sample case
Initial weight is set, can train to obtain the corresponding weight of first Weak Classifier later, thus according to first Weak Classifier
Corresponding weight is updated the initial weight of the corresponding case data of each sample case, to obtain the first weak typing
Device.
If desired three Weak Classifiers are trained, then can train again in the manner described above and obtain two Weak Classifiers: the
Two Weak Classifiers and third Weak Classifier, so that the first Weak Classifier, the second Weak Classifier and third Weak Classifier are carried out group
It closes, to obtain classifier.
Further, in order to improve the accuracy of classifier, the classifier that can be obtained to combination be modified, thus
To strong classifier model.
It therefore, can be to multiple sample predictions results and corresponding reality during being modified to classifier
Differentiate that result is compared, obtains comparison result, further according to the comparison result, calculate the loss function of classifier, last basis
The loss function is modified classifier, obtains strong classifier model.
Specifically, classifier can predict each sample cases, and it is pre- to obtain the corresponding sample of multiple sample cases
It surveys as a result, be again compared sample predictions result with the practical differentiation result for belonging to same sample cases, obtains comparison result.
Correspondingly, the loss function of classifier can be calculated according to comparison result, so that it is determined that classifier output result
Accuracy can calculate corresponding change of gradient further according to loss function, last according to the adjustment classification of obtained change of gradient
The weight of each Weak Classifier in device, obtains strong classifier model.
Step 201b, the sample data after obtaining the normalization of multiple sample cases, and determined according to the sample data multiple
The initial load factor, then multiple initial load factors are trained by sample data, obtain linear regression model (LRM).
Wherein, sample data may include: sample justice data, the just data of sample, sample social effect data and reality
Border differentiates result.
Sample data in this step 201b is similar with the sample data in step 201a, and details are not described herein.
After the sample data to each type is normalized, linear return can be determined according to the type of sample data
Return the number of principal component in model, and the corresponding initial load factor of each principal component is arranged according to sample data, further according to first
Beginning load factor calculates sample data, obtains sample predictions as a result, and tying according to sample predictions result and practical differentiation
Fruit is compared, and obtains comparison result, to be adjusted according to comparison result to each initial load factor.
When the number that the accuracy of linear regression model (LRM) reaches certain standard or the iteration adjustment initial load factor reaches pre-
If when threshold value, linear regression model (LRM) can be generated according to the currently corresponding parameter of each load factor.
For example, may include: fair data X according to the principal component that sample data determines1, just data X2, and society's effect
Answer data X3, then the initial load factor corresponding with each principal component can be c1、c2And c3, wherein c1With X1It is corresponding, c2With X2It is right
It answers, c3With X3It corresponds to, then available initial linear regression model (LRM) y=c1*X1+c2*X2+c3*X3, y is according to sample data
Carry out the classification results of prediction classification.If obtaining the corresponding parameter of each initial load factor according to the training of each sample data
Are as follows: c1=1.2, c2=0.5, c3=0.8, then the linear regression model (LRM) that training obtains is y=1.2X1+0.5X2+0.8X3。
It should be noted that in step 201a and step 201b, it can be by pre-set normalization formula to each
The corresponding primary data of a case is normalized, thus the sample data after being normalized.Wherein, which can
With are as follows:Wherein, i is positive integer, xiFor the sample data of the i-th seed type after normalization, xi0For normalizing
Change the primary data of preceding i-th seed type, ximinFor the smallest primary data of parameter value in the primary data of the i-th seed type, ximaxFor
The smallest primary data of parameter value in the primary data of i-th seed type.
Step 202, the case data for obtaining case to be detected.
Wherein, which may include fair data, just data and social effect data.
Case data are inputted strong classifier model and linear regression model (LRM) by step 203 respectively, respectively obtain strong point of input
First result of class device model output and the second result of linear regression model (LRM) output.
Wherein, first result and second result are the judgement result for the case to be detected that prediction obtains.
Step 204, according to corresponding first weight of the first result and corresponding second weight of the second result, in conjunction with the first knot
Fruit and the second result are calculated, and prediction result is obtained.
After obtaining the first result and the second result, the first result and the second result can be calculated again, according to
Result after calculating determines prediction result.
For example, the plaintiff of case wins a lawsuit to be indicated with parameter " 1 ", plaintiff loses a lawsuit to be indicated with parameter " 2 ", and the first result is corresponding
Weight be 0.7, the corresponding weight of the second result is 0.3, and therefore, if the first result is 1, the second result is 2, then COMPREHENSIVE CALCULATING
Calculated result afterwards is 1*0.7+2*0.3=1.3, then can be by 1.3 as the prediction result finally obtained.
In conclusion case provided by the invention differentiates the prediction technique of result, it include public affairs by obtaining case to be detected
Case data are inputted strong classifier model and linear by the case data of flat data, just data and social effect data respectively
Regression model, respectively obtain input strong classifier model output the first result and linear regression model (LRM) output second as a result,
It is finally analyzed according to the first result and the second result, obtains prediction result.By data model to the case of case to be detected
Number of packages carries out comprehensive analysis according to being analyzed, and to the result of multiple models output, obtains the prediction result for integrating multiple results,
The problem lower by the prediction result accuracy manually predicted is avoided, prediction case is improved and differentiates the accurate of result
Property.
Fig. 3 is the schematic diagram for the prediction meanss that the case that one embodiment of the invention provides differentiates result, as shown in figure 3, should
Device specifically includes:
First obtains module 301, for obtaining the case data of case to be detected, the case data include fair data,
Just data and social effect data；
Input module 302, for the case data to be inputted strong classifier model and linear regression model (LRM) respectively, respectively
To the input strong classifier model export the first result and the linear regression model (LRM) output second as a result, first result and
Second result is the judgement result for the case to be detected that prediction obtains；
Determining module 303 obtains prediction result for being analyzed according to first result and second result.
Optionally, referring to fig. 4, the device further include:
Second obtains module 304, for obtaining the sample data after multiple sample cases normalize, the sample data packet
It includes: sample justice data, the just data of sample, sample social effect data and practical differentiation result；
First training module 305, for being instructed according to the sample data to pre-set initial strong classifier model
Practice, obtains classifier；
Correction module 306, the sample predictions of multiple sample cases for being exported according to the classifier are as a result, and more
A sample case it is corresponding it is practical differentiate as a result, be modified to the classifier, obtain the strong classifier model.
Optionally, the correction module 306 is specifically used for multiple sample predictions results and corresponding practical differentiation knot
Fruit is compared, and obtains comparison result；According to the comparison result, the loss function of the classifier is calculated；According to the loss function
The classifier is modified, the strong classifier model is obtained.
Optionally, referring to Fig. 5, the device further include:
Third obtains module 307, for obtaining the sample data after multiple sample cases normalize, the sample data packet
It includes: sample justice data, the just data of sample, sample social effect data and practical differentiation result；
Model building module 308, for determining multiple initial load factors according to the sample data；
Second training module 309 is somebody's turn to do for being trained by the sample data to multiple initial load factors
Linear regression model (LRM).
Optionally, the determining module 303 is specifically used for according to corresponding first weight of first result and second result
Corresponding second weight is calculated in conjunction with first result and second result, obtains the prediction result.
In conclusion case provided by the invention differentiates the prediction meanss of result, it include public affairs by obtaining case to be detected
Case data are inputted strong classifier model and linear by the case data of flat data, just data and social effect data respectively
Regression model, respectively obtain input strong classifier model output the first result and linear regression model (LRM) output second as a result,
It is finally analyzed according to the first result and the second result, obtains prediction result.By data model to the case of case to be detected
Number of packages carries out comprehensive analysis according to being analyzed, and to the result of multiple models output, obtains the prediction result for integrating multiple results,
The problem lower by the prediction result accuracy manually predicted is avoided, prediction case is improved and differentiates the accurate of result
Property.
The method that above-mentioned apparatus is used to execute previous embodiment offer, it is similar that the realization principle and technical effect are similar, herein not
It repeats again.
The above module can be arranged to implement one or more integrated circuits of above method, such as: one
Or multiple specific integrated circuits (Application Specific Integrated Circuit, abbreviation ASIC), or, one
Or multi-microprocessor (digital singnal processor, abbreviation DSP), or, one or more field programmable gate
Array (Field Programmable Gate Array, abbreviation FPGA) etc..For another example, when some above module passes through processing elements
When the form of part scheduler program code is realized, which can be general processor, such as central processing unit (Central
Processing Unit, abbreviation CPU) or it is other can be with the processor of caller code.For another example, these modules can integrate
Together, it is realized in the form of system on chip (system-on-a-chip, abbreviation SOC).
Fig. 6 is the schematic diagram for the prediction meanss that the case that one embodiment of the invention provides differentiates result, which can collect
At in the chip of terminal device or terminal device, which can be the calculating for having the forecast function of case differentiation result and sets
It is standby.
The device includes: memory 601, processor 602.
Memory 601 is for storing program, the program that processor 602 calls memory 601 to store, to execute the above method
Embodiment.Specific implementation is similar with technical effect, and which is not described herein again.
Optionally, the present invention also provides a kind of program product, such as computer readable storage medium, including program, the journeys
Sequence is when being executed by processor for executing above method embodiment.
In several embodiments provided by the present invention, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can be tied
Another system is closed or is desirably integrated into, or some features can be ignored or not executed.Another point, it is shown or discussed
Mutual coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING or logical of device or unit
Letter connection can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or the network equipment etc.) or processor (English: processor) execute this hair
The part steps of bright each embodiment the method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory
(English: Read-Only Memory, abbreviation: ROM), random access memory (English: Random Access Memory, letter
Claim: RAM), the various media that can store program code such as magnetic or disk.
Upper is only the specific embodiment of the application, but the protection scope of the application is not limited thereto, any to be familiar with sheet
Those skilled in the art within the technical scope of the present application, can easily think of the change or the replacement, and should all cover at this
Within the protection scope of application.Therefore, the protection scope of the application should be subject to the protection scope in claims.
Claims (10)
1. a kind of case differentiates the prediction technique of result, which is characterized in that the described method includes:
The case data of case to be detected are obtained, the case data include fair data, just data and social effect data；
The case data are inputted into strong classifier model and linear regression model (LRM) respectively, respectively obtain the input strong classifier
The second of the first result and linear regression model (LRM) output that model exports is as a result, first result and second result
It is the judgement result for the case to be detected that prediction obtains；
It is analyzed according to first result and second result, obtains prediction result.
2. the method as described in claim 1, which is characterized in that before the case data for obtaining case to be detected, institute
State method further include:
Sample data after obtaining the normalization of multiple sample cases, the sample data include: that sample justice data, sample are just
Data, sample social effect data and practical differentiation result；
Pre-set initial strong classifier model is trained according to the sample data, obtains classifier；
According to the sample predictions of multiple sample cases of classifier output as a result, and multiple sample cases pair
The practical differentiation answered obtains the strong classifier model as a result, be modified to the classifier.
3. method according to claim 2, which is characterized in that the multiple sample cases exported according to the classifier
The sample predictions of part as a result, and multiple sample cases it is corresponding it is practical differentiate as a result, be modified to the classifier,
Obtain the strong classifier model, comprising:
Multiple sample predictions results and corresponding practical differentiation result are compared, comparison result is obtained；
According to the comparison result, the loss function of the classifier is calculated；
The classifier is modified according to the loss function, obtains the strong classifier model.
4. the method as described in claim 1, which is characterized in that before the case data for obtaining case to be detected, institute
State method further include:
Sample data after obtaining the normalization of multiple sample cases, the sample data include: that sample justice data, sample are just
Data, sample social effect data and practical differentiation result；
Multiple initial load factors are determined according to the sample data；
Multiple initial load factors are trained by the sample data, obtain the linear regression model (LRM).
5. the method as described in Claims 1-4 is any, which is characterized in that described according to first result and described second
As a result it is analyzed, obtains prediction result, comprising:
According to corresponding first weight of first result and corresponding second weight of second result, in conjunction with first knot
Fruit and second result are calculated, and the prediction result is obtained.
6. the prediction meanss that a kind of case differentiates result, which is characterized in that described device includes:
First obtains module, and for obtaining the case data of case to be detected, the case data include fair data, just number
According to social effect data；
Input module respectively obtains institute for the case data to be inputted strong classifier model and linear regression model (LRM) respectively
The first result that input strong classifier model exports and second that the linear regression model (LRM) exports are stated as a result, first result
It is the judgement result for the case to be detected that prediction obtains with second result；
Determining module obtains prediction result for being analyzed according to first result and second result.
7. device as claimed in claim 6, which is characterized in that described device further include:
Second obtains module, and for obtaining the sample data after multiple sample cases normalize, the sample data includes: sample
The just data of fair data, sample, sample social effect data and practical differentiation result；
First training module is obtained for being trained according to the sample data to pre-set initial strong classifier model
To classifier；
Correction module, the sample predictions of multiple sample cases for being exported according to the classifier are as a result, and multiple
The sample case it is corresponding it is practical differentiate as a result, be modified to the classifier, obtain the strong classifier model.
8. device as claimed in claim 7, which is characterized in that the correction module is specifically used for pre- to multiple samples
It surveys result and corresponding practical differentiation result is compared, obtain comparison result；According to the comparison result, described point is calculated
The loss function of class device；The classifier is modified according to the loss function, obtains the strong classifier model.
9. device as claimed in claim 6, which is characterized in that described device further include:
Third obtains module, and for obtaining the sample data after multiple sample cases normalize, the sample data includes: sample
The just data of fair data, sample, sample social effect data and practical differentiation result；
Model building module, for determining multiple initial load factors according to the sample data；
Second training module obtains described for being trained by the sample data to multiple initial load factors
Linear regression model (LRM).
10. the device as described in claim 6 to 9 is any, which is characterized in that the determining module is specifically used for according to
Corresponding first weight of first result and corresponding second weight of second result, in conjunction with first result and described second
As a result it is calculated, obtains the prediction result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910321775.1A CN110046256A (en) | 2019-04-22 | 2019-04-22 | The prediction technique and device of case differentiation result |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910321775.1A CN110046256A (en) | 2019-04-22 | 2019-04-22 | The prediction technique and device of case differentiation result |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110046256A true CN110046256A (en) | 2019-07-23 |
Family
ID=67278141
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910321775.1A Pending CN110046256A (en) | 2019-04-22 | 2019-04-22 | The prediction technique and device of case differentiation result |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110046256A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050278322A1 (en) * | 2004-05-28 | 2005-12-15 | Ibm Corporation | System and method for mining time-changing data streams |
| WO2006098766A2 (en) * | 2004-09-17 | 2006-09-21 | Proximex | Incremental data fusion and decision making system and associated method |
| CN103426007A (en) * | 2013-08-29 | 2013-12-04 | 人民搜索网络股份公司 | Machine learning classification method and device |
| CN105760888A (en) * | 2016-02-23 | 2016-07-13 | 重庆邮电大学 | Neighborhood rough set ensemble learning method based on attribute clustering |
| CN106156401A (en) * | 2016-06-07 | 2016-11-23 | 西北工业大学 | Data-driven system state model on-line identification methods based on many assembled classifiers |
| CN107291688A (en) * | 2017-05-22 | 2017-10-24 | 南京大学 | Judgement document's similarity analysis method based on topic model |
| CN107330820A (en) * | 2017-08-28 | 2017-11-07 | 北京智诚律法科技有限公司 | A kind of forecasting system and method for lawsuit result |
| CN107577785A (en) * | 2017-09-15 | 2018-01-12 | 南京大学 | A kind of level multi-tag sorting technique suitable for law identification |
| CN109241285A (en) * | 2018-08-29 | 2019-01-18 | 东南大学 | A kind of device of the judicial decision in a case of auxiliary based on machine learning |
-
2019
- 2019-04-22 CN CN201910321775.1A patent/CN110046256A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050278322A1 (en) * | 2004-05-28 | 2005-12-15 | Ibm Corporation | System and method for mining time-changing data streams |
| WO2006098766A2 (en) * | 2004-09-17 | 2006-09-21 | Proximex | Incremental data fusion and decision making system and associated method |
| CN103426007A (en) * | 2013-08-29 | 2013-12-04 | 人民搜索网络股份公司 | Machine learning classification method and device |
| CN105760888A (en) * | 2016-02-23 | 2016-07-13 | 重庆邮电大学 | Neighborhood rough set ensemble learning method based on attribute clustering |
| CN106156401A (en) * | 2016-06-07 | 2016-11-23 | 西北工业大学 | Data-driven system state model on-line identification methods based on many assembled classifiers |
| CN107291688A (en) * | 2017-05-22 | 2017-10-24 | 南京大学 | Judgement document's similarity analysis method based on topic model |
| CN107330820A (en) * | 2017-08-28 | 2017-11-07 | 北京智诚律法科技有限公司 | A kind of forecasting system and method for lawsuit result |
| CN107577785A (en) * | 2017-09-15 | 2018-01-12 | 南京大学 | A kind of level multi-tag sorting technique suitable for law identification |
| CN109241285A (en) * | 2018-08-29 | 2019-01-18 | 东南大学 | A kind of device of the judicial decision in a case of auxiliary based on machine learning |
Non-Patent Citations (1)
| Title |
|---|
| 朱青等: "基于文本挖掘和自动分类的法院裁判决策支持", 《中国管理科学》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210042580A1 (en) | Model training method and apparatus for image recognition, network device, and storage medium | |
| CN110070067A (en) | The training method of video classification methods and its model, device and electronic equipment | |
| CN107766929B (en) | Model analysis method and device | |
| CN109815492A (en) | A kind of intension recognizing method based on identification model, identification equipment and medium | |
| CN110309874A (en) | Negative sample screening model training method, data screening method and data matching method | |
| CN106651057A (en) | Mobile terminal user age prediction method based on installation package sequence table | |
| KR20040037180A (en) | System and method of face recognition using portions of learned model | |
| CN110874604A (en) | Model training method and terminal equipment | |
| CN102722577B (en) | The determination method and device of index changeable weight | |
| CN111079780A (en) | Training method of space map convolution network, electronic device and storage medium | |
| CN108416032A (en) | A kind of file classification method, device and storage medium | |
| CN110046706A (en) | Model generating method, device and server | |
| CN109711401A (en) | A kind of Method for text detection in natural scene image based on Faster Rcnn | |
| CN110717554A (en) | Image recognition method, electronic device, and storage medium | |
| CN110263854A (en) | Live streaming label determines method, apparatus and storage medium | |
| CN110135505A (en) | Image classification method, device, computer equipment and computer readable storage medium | |
| CN110069959A (en) | A kind of method for detecting human face, device and user equipment | |
| CN109614987A (en) | More disaggregated model optimization methods, device, storage medium and electronic equipment | |
| CN110533459A (en) | A kind of terminal replacement probability prediction technique and device based on gradient boosting algorithm | |
| CN114783021A (en) | Intelligent detection method, device, equipment and medium for wearing of mask | |
| CN109460872B (en) | Mobile communication user loss imbalance data prediction method | |
| CN114596590A (en) | Single-monkey video motion classification method with long-tail distribution characteristic | |
| CN110377775A (en) | A kind of picture examination method and device, storage medium | |
| CN107704364B (en) | Background application management-control method, device, storage medium and electronic equipment | |
| CN109961085B (en) | Method and device for establishing flight delay prediction model based on Bayesian estimation |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190723 |
|
| RJ01 | Rejection of invention patent application after publication |