CN105117605B - A kind of device and method of case prediction - Google Patents
A kind of device and method of case prediction Download PDFInfo
- Publication number
- CN105117605B CN105117605B CN201510570751.1A CN201510570751A CN105117605B CN 105117605 B CN105117605 B CN 105117605B CN 201510570751 A CN201510570751 A CN 201510570751A CN 105117605 B CN105117605 B CN 105117605B
- Authority
- CN
- China
- Prior art keywords
- influence factor
- factor
- history
- influence
- prediction
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000005303 weighing Methods 0.000 claims abstract description 84
- 238000011478 gradient descent method Methods 0.000 claims description 9
- 238000012417 linear regression Methods 0.000 description 13
- 238000004458 analytical method Methods 0.000 description 7
- 238000012731 temporal analysis Methods 0.000 description 5
- 238000000700 time series analysis Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Landscapes
- Alarm Systems (AREA)
Abstract
The invention discloses a kind of device and method of case prediction, described device includes：Determination unit, at least one influence factor is determined for receiving prediction instruction, and according to the prediction instruction；According at least one influence factor, the weighing factor of at least one influence factor is determined；Acquiring unit, for obtaining the history value of at least one influence factor；The determination unit, is additionally operable to the history value of at least one influence factor obtained according to the acquiring unit, using time series models, determines the predicted value of at least one influence factor；Predicting unit, for the weighing factor of at least one influence factor and the predicted value of at least one influence factor determined according to the determination unit, predicts crime case quantity.The present invention realizes the purpose for the accuracy for improving caseload prediction.
Description
Technical field
The present invention relates to field of computer technology, espespecially a kind of device and method of case prediction.
Background technology
With developing rapidly for science and technology, Police Information is built also fast-developing therewith.And the Accurate Prediction of caseload is past
Toward to police strength reasonable deployment, gross demand, the period with police, post setting, specialty requirement, work post distribution, region adjustment and
Civilian's displacement, auxiliary police officer's assistance etc. have very high directive significance.Local government can also be according to the change of caseload at the same time
To disclose the underlying causes of case generation, and then a very valuable reference is provided for the decision-making of government.
At present in public security industry, the main method of caseload prediction is based on time series analysis, space orientation point
What analysis carried out.It is according to the scene of history caseload, and case, predicts crime case occurred frequentlyly at a certain section
Between caseload.But as social means of transportation and information-based develop rapidly, human, financial, and material resources, information flowability are larger,
It is trans-regional, flee about to commit crimes on a large scale it is increasingly prominent.Since the above method is analyzed using time series analysis, space orientation
It is predicted, does not consider other influences factor, therefore the method for above-mentioned prediction caseload is not accurate enough.
The content of the invention
In order to solve the above technical problem, the present invention provides a kind of device and method of case prediction, it is possible to increase case
The accuracy of part quantitative forecast.
In order to reach the object of the invention, the present invention provides a kind of device of case prediction, including：Determination unit, is used for
Prediction instruction is received, and at least one influence factor is determined according to the prediction instruction；According at least one influence factor,
Determine the weighing factor of at least one influence factor；Acquiring unit, for obtaining going through at least one influence factor
History value；The determination unit, is additionally operable to the history value of at least one influence factor obtained according to the acquiring unit, profit
With time series models, the predicted value of at least one influence factor is determined；Predicting unit, for according to the definite list
The weighing factor at least one influence factor that member is determined and the predicted value of at least one influence factor, predict
Crime case quantity.
Alternatively, the determination unit, specifically for determining n history crime case number of packages amount, at least one influence factor
N history value；According to the n history crime case number of packages amount, n history value of at least one influence factor, at criticizing
Gradient descent method is managed, determines the weighing factor of at least one influence factor；N is the integer more than 0.
Alternatively, the determination unit, specifically for according to the n history crime case number of packages amount, at least one influence
N history value of factor, utilizes formula：Determine at least one influence
The weighing factor of factor；Wherein, j is integer more than 0, θjRepresenting the weighing factor of j-th of influence factor, α represents gradient,
Represent i-th of history crime case number of packages amount；hθ(x(i)) represent i-th of the crime case quantity predicted,Represent j-th of influence
I-th of history value of factor, n represent the number of history crime case number of packages amount, and represent the number of the history value of influence factor, i
For the integer more than 0, and less than or equal to n.
Alternatively, the predicting unit, is weighed specifically for the influence according at least one influence factor determined
The predicted value of weight and at least one influence factor, utilizes formula：Predict
Crime case quantity；Wherein, y and hθ(x) the crime case quantity of prediction, θ are represented0Expression system constant, θjRepresent influence because
Plain xjWeighing factor, xjRepresent j-th of influence factor, m is the number of influence factor, and ε represents random error.
Alternatively, at least one Guessing Factors are carried in prediction instruction；The determination unit, specifically for obtaining n
History crime case number of packages amount；It is true at least one Guessing Factors according to the n history crime case number of packages amount of acquisition
Fixed at least one influence factor.
Further, the present invention provides a kind of method of case prediction, including：Prediction instruction is received, and according to described
Prediction instruction determines at least one influence factor；According at least one influence factor, determine at least one influence because
The weighing factor of element；Obtain the history value of at least one influence factor, and according at least one influence of acquisition because
The history value of element, using time series models, determines the predicted value of at least one influence factor；N is whole more than 0
Number；According to the weighing factor at least one influence factor determined and the predicted value of at least one influence factor,
Predict crime case quantity.
Alternatively, it is described according at least one influence factor, determine that the influence of at least one influence factor is weighed
Include again：Determine n history crime case number of packages amount, n history value of at least one influence factor；According to the n historical offender
Guilty caseload, n history value of at least one influence factor, using batch processing gradient descent method, determines described at least one
The weighing factor of influence factor.
Alternatively, described according to the n history crime case number of packages amount, the n history values of at least one influence factor, utilize
Batch processing gradient descent method, determining the weighing factor of at least one influence factor includes：According to the n history crime case
Number of packages amount, n history value of at least one influence factor, utilizes formula：
Determine the weighing factor of at least one influence factor；Wherein, j is integer more than 0, θjRepresent j-th influence factor
Weighing factor, α represent gradient,Represent i-th of history crime case number of packages amount；hθ(x(i)) represent i-th of crime case predicted
Quantity,Represent i-th of history value of j-th of influence factor, n represents the number of history crime case number of packages amount, and represents to influence
The number of the history value of factor, i are the integer more than 0, and less than or equal to n.
Alternatively, the weighing factor at least one influence factor that the basis is determined and at least one shadow
The predicted value of the factor of sound, predicting crime case quantity includes：According to the influence at least one influence factor determined
The predicted value of weight and at least one influence factor, utilizes formula：Prediction
Go out crime case quantity；Wherein, y and hθ(x) the crime case quantity of prediction, θ are represented0Expression system constant, θjRepresent to influence
Factor xjWeighing factor, xjRepresent j-th of influence factor, m is the number of influence factor, and ε represents random error.
Alternatively, at least one Guessing Factors are carried in the prediction instruction；It is described to be instructed according to the prediction, determine
At least one influence factor includes：Obtain n history crime case number of packages amount；According to the n history crime case number of packages of acquisition
Amount, determines at least one influence factor at least one Guessing Factors.
Compared with prior art, the present invention includes determination unit, is instructed for receiving prediction instruction, and according to the prediction
Determine at least one influence factor；According at least one influence factor, the influence of at least one influence factor is determined
Weight；Acquiring unit, for obtaining the history value of at least one influence factor；The determination unit, is additionally operable to according to institute
State the history value of at least one influence factor of acquiring unit acquisition, using time series models, determine it is described at least
The predicted value of one influence factor；Predicting unit, for according at least one influence that the determination unit is determined because
The weighing factor of element and the predicted value of at least one influence factor, predict crime case quantity.In this way, case prediction
Device first determines influence factor when carrying out the prediction of crime case data, and then determines the weighing factor of influence factor,
And the predicted value of influence factor is obtained according to the method for time series models, so as to the weighing factor according to influence factor and
The predicted value of influence factor carries out the prediction of crime case quantity.As, the device of case prediction is the shadow according to crime case
The factor of sound carries out the prediction of crime case, rather than just quantity based on Time Series Analysis Forecasting crime case, therefore,
The present invention improves the accuracy of caseload prediction.
Other features and advantages of the present invention will be illustrated in the following description, also, partly becomes from specification
Obtain it is clear that or being understood by implementing the present invention.The purpose of the present invention and other advantages can be by specification, rights
Specifically noted structure is realized and obtained in claim and attached drawing.
Brief description of the drawings
Attached drawing is used for providing further understanding technical solution of the present invention, and a part for constitution instruction, with this
The embodiment of application is used to explain technical scheme together, does not form the limitation to technical solution of the present invention.
Fig. 1 is a kind of structure diagram of the device of case prediction provided in an embodiment of the present invention；
Fig. 2 is a kind of flow diagram of the method for case prediction provided in an embodiment of the present invention.
Embodiment
For the object, technical solutions and advantages of the present invention are more clearly understood, below in conjunction with attached drawing to the present invention
Embodiment be described in detail.It should be noted that in the case where there is no conflict, in the embodiment and embodiment in the application
Feature can mutually be combined.
Step shown in the flowchart of the accompanying drawings can be in the computer system of such as a group of computer-executable instructions
Perform.Also, although logical order is shown in flow charts, in some cases, can be with suitable different from herein
Sequence performs shown or described step.
An embodiment of the present invention provides a kind of device of case prediction, as shown in Figure 1, including：
Determination unit 101, for determining at least one influence factor；Determine the weighing factor of at least one influence factor.
Specifically, user, when needing to predict crime case quantity, the determination unit 101 of the device of case prediction can connect
The prediction instruction of user's transmission is received, and analytical Prediction instructs, if not parsing Guessing Factors in prediction instructs, it is determined that
At least one influence factor itself stored can be determined as predicting influence factor required during caseload by unit 101.
If parsing Guessing Factors in prediction instructs, it is, when carrying at least one Guessing Factors in prediction instruction,
Then illustrate that user needs the device of case prediction to determine at least one influence factor in Guessing Factors.At this time, determination unit
101 determine that at least one influence factor specifically includes：
Determination unit 101, specifically for obtaining n history crime case number of packages amount；According to n history crime case of acquisition
Quantity, determines at least one influence factor.
Wherein, n is the integer more than 0.
That is, user can deduce at least one Guessing Factors related with crime case quantity in advance, and will
At least one Guessing Factors related with crime case quantity deduced are carried into prediction instruction, are sent to case prediction
The determination unit 101 of device.Determination unit 101 can parse the multiple with violating of its interior carrying after prediction instruction is received
The related Guessing Factors of guilty caseload, and n history crime case number of packages amount is obtained, obtaining n history crime case number of packages
, can be according to gradually linear regression (stepwise) method in multiple linear regression analysis method, from Guessing Factors after amount
Determine at least one and relevant influence factor of crime case data volume.As, n history crime case number of packages amount is returned
One change is handled, and then at least one and relevant influence factor of crime case data volume can be determined from Guessing Factors.
In embodiments of the present invention, thus it is speculated that factor is the factor for being used to predict crime case quantity of user in predicting.Influence
Factor be the device of case prediction in Guessing Factors, being used for of determining predicts the factor of crime case quantity.
Wherein, the method that the device of case prediction obtains n history crime case number of packages amount can be from storing historical offender
Obtained in the database of guilty caseload or n history crime case number of packages amount is sent by user, and then obtain n
History crime case number of packages amount.
Further, gradually linear regression (stepwise) of the determination unit 101 in multiple linear regression analysis method
Regression equation used in method is：
Wherein, y and hθ(x) the crime case quantity of prediction, θ are represented0Expression system constant, θjRepresent influence factor xj
Weighing factor, xjRepresent j-th of influence factor, n is the number of influence factor, and ε represents random error.
Further, the weighing factor of influence factor can periodically update.In the pre- of this progress caseload
During survey, if the cycle of its renewal is not reaching to, and when the weighing factor of at least one influence factor has calculated, determination unit
101 need not recalculate the weighing factor of each influence factor, only need to store each influence according at least one influence factor
In the memory space of the weighing factor of factor, the weighing factor of each influence factor is obtained.
If reach the update cycle of the weighing factor of influence factor in the prediction of this progress caseload, or at least
When the weighing factor of one influence factor does not calculate, determination unit 101 determines the weighing factor of at least one influence factor
Specifically include：
Determination unit 101, specifically for determining n history crime case number of packages amount, n history of at least one influence factor
Value；According to n history crime case number of packages amount, n history value of at least one influence factor, using batch processing gradient descent method,
Determine the weighing factor of at least one influence factor.
Wherein, n is the integer more than 0.
Further, determination unit 101 is after at least one influence factor is determined, according at least one influence factor,
N history value of each influence factor is obtained from the memory space of the history value of storage influence factor.If do not have in memory space
There is n history value for recording at least one influence factor, determination unit 101 can be set according to the actual requirements by user at this time
N history value of at least one influence factor, and by by n history value of at least one influence factor of user setting, really
It is set to n history value of at least one influence factor.After n history value of at least one influence factor is determined, using true
N history value of at least one influence factor made, n history crime case number of packages amount, can by batch processing gradient descent method
To determine the weighing factor of at least one influence factor.
Further, determination unit 101, determine n history crime case number of packages amount, and n of at least one influence factor go through
History value；According to n history crime case number of packages amount, n history value of at least one influence factor, is declined using batch processing gradient
Method, determines that the weighing factor of at least one influence factor specifically includes：
Determination unit 101, specifically for according to n history crime case number of packages amount, n history of at least one influence factor
Value, utilizes formula：Determine the weighing factor of at least one influence factor.
Wherein, j is integer more than 0, θjRepresenting the weighing factor of j-th of influence factor, α represents gradient,Represent i-th
A history crime case number of packages amount.hθ(x(i)) represent i-th of the crime case quantity predicted,Represent the of j-th of influence factor
I history value, n represent the number of history crime case number of packages amount, and represent the number of the history value of influence factor, i be more than 0,
And the integer less than or equal to n.:=represent is iterated calculating.
α in determination unit 101 can be pre-set according to the actual requirements.
Acquiring unit 102, for obtaining the history value of at least one influence factor.
Specifically, acquiring unit 102 can obtain required shadow from the memory space of the history value of storage influence factor
The history value of the factor of sound.The device of case prediction can obtain multiple history values of influence factor, such as obtain influence factor
20 history values.Depending on acquiring unit 102 specifically obtains several history values of influence factor according to the actual requirements, the present invention is to this
It is not restricted.
Determination unit 101, is additionally operable to the history value of at least one influence factor obtained according to acquiring unit 102, utilizes
Time series models, determine the predicted value of at least one influence factor.
Specifically, due to when carrying out the prediction of crime case quantity, it is necessary to predicted value using influence factor.Therefore,
Determination unit 101 needs first to determine the predicted value of at least one influence factor.At least one influence is obtained in acquiring unit 102
After the history value of factor, the history value at least one influence factor that determination unit 101 is obtained according to acquiring unit 102, passes through
Time series models can determine the predicted value of at least one influence factor.
Wherein, time series models be it is well known in the prior art it is a kind of obtain predicted value method, the present invention to this not
Repeat again.
Predicting unit 103, for the weighing factor at least one influence factor determined according to determination unit 101
And the predicted value of at least one influence factor, predict crime case quantity.
Specifically, predicting unit 103 determines the weighing factor and predicted value of each influence factor in determination unit 101
Afterwards, the weighing factor of each influence factor and predicted value directly can be calculated into product respectively, obtains each influence factor pair
The crime case quantity answered, and then the crime case quantity of each influence factor is subjected to summation operation, predict total crime
Caseload.
Further, predicting unit 103, specifically for according to the weighing factor of at least one influence factor determined and
The predicted value of at least one influence factor, utilizes formula：Predict crime case
Quantity.
Wherein, y and hθ(x) the crime case quantity of prediction, θ are represented0Expression system constant, θjRepresent influence factor xj
Weighing factor, xjRepresent j-th of influence factor, m is the number of influence factor, and ε represents random error.
In predicting unit 103, θ0Can be that user is pre-set according to the actual requirements.
An embodiment of the present invention provides a kind of device of case prediction, including according to history crime case number of packages amount, determine
At least one influence factor；Determine the weighing factor of at least one influence factor；According to history crime case number of packages amount, the time is utilized
Series model, determines the predicted value of at least one influence factor；Weighed according to the influence at least one influence factor determined
The predicted value of weight and at least one influence factor, predicts crime case quantity.In this way, the device of case prediction is carrying out
During the prediction of crime case data, influence factor is first determined, and then determine the weighing factor of influence factor, and according to the time
The method of series model obtains the predicted value of influence factor, so as to the weighing factor according to influence factor and influence factor
Predicted value carries out the prediction of crime case quantity.As, the device of case prediction is carried out according to the influence factor of crime case
The prediction of crime case, rather than just the quantity based on Time Series Analysis Forecasting crime case, therefore, the present invention improves
The accuracy of caseload prediction.
An embodiment of the present invention provides a kind of method of case prediction, as shown in Fig. 2, including：
Step 201, receive prediction instruction, and is instructed according to prediction and determine at least one influence factor.
Specifically, user, when needing to predict crime case quantity, the device of case prediction can receive user's transmission
Prediction instruction, at this time case prediction device can be instructed with analytical Prediction, if prediction instruct in do not parse supposition because
At least one influence factor itself stored can be determined as predicting required during caseload by element, the then device that case is predicted
Influence factor.
If parsing Guessing Factors in prediction instructs, it is, when carrying at least one Guessing Factors in prediction instruction,
Then illustrate that user needs the device of case prediction to determine at least one influence factor in Guessing Factors.At this time, case is predicted
Device determine that at least one influence factor includes：Obtain n history crime case number of packages amount；According to n history crime of acquisition
Caseload, determines at least one influence factor at least one Guessing Factors.
Wherein, n is the integer more than 0.
That is, user can deduce at least one Guessing Factors related with crime case quantity in advance, and will
At least one Guessing Factors related with crime case quantity deduced are carried into prediction instruction, are sent to case prediction
Device.The device of case prediction can parse the multiple and crime case quantity carried in it after prediction instruction is received
Related Guessing Factors, and n history crime case number of packages amount is obtained, after n history crime case number of packages amount is obtained, Ke Yigen
According to gradually linear regression (stepwise) method in multiple linear regression analysis method, at least one is determined from Guessing Factors
A and relevant influence factor of crime case data volume.As, n history crime case number of packages amount is normalized, into
And at least one and relevant influence factor of crime case data volume can be determined from Guessing Factors.
In embodiments of the present invention, thus it is speculated that factor is the factor for being used to predict crime case quantity of user in predicting.Influence
Factor be the device of case prediction in Guessing Factors, being used for of determining predicts the factor of crime case quantity.
Wherein, the method that the device of case prediction obtains n history crime case number of packages amount can be from storing historical offender
Obtained in the database of guilty caseload or n history crime case number of packages amount is sent by user, and then obtain n
History crime case number of packages amount.
Further, in step 201, how to determine the method for influence factor includes：The device of case prediction is in polynary line
The regression equation used in gradually linear regression (stepwise) method in property regression analysis is：
Wherein, y and hθ(x) the crime case quantity of prediction, θ are represented0Expression system constant, θjRepresent influence factor xj
Weighing factor, xjRepresent j-th of influence factor, n is the number of influence factor, and ε represents random error.
Gradually linear regression (stepwise) method in multiple linear regression analysis method in step 201 is existing
There is gradually linear regression (stepwise) method in the multiple linear regression analysis method in technology, details are not described herein.
Exemplary, the Guessing Factors that user can deduce in advance are：Per-capita gross domestic product (GDP, Gross
Domestic Product), education level, the level of urbanization, Gini coefficient, the size of population, if get married.And it will deduce
Guessing Factors carry to prediction instruct in, send to case prediction device.At this time, case prediction device receive it is pre-
After surveying instruction, prediction instruction is parsed, and then the Guessing Factors that user speculates in advance can be parsed.Speculate parsing
After factor, n history crime case number of packages amount can be obtained, after n history crime case number of packages amount is obtained, can be utilized polynary
Gradually linear regression (stepwise) method in linear regression analysis method, in Guessing Factors per-capita gross domestic product, by
Level of education, the level of urbanization, Gini coefficient, the size of population, if in marriage, by per-capita gross domestic product, water of receiving an education
It is flat, the level of urbanization, Gini coefficient, five Guessing Factors of the size of population be determined as with the relevant influence of crime case quantity because
Element.
Wherein, per-capita gross domestic product can be calculated according to per-capita gross domestic product index.The level of urbanization can be with
The ratio that total population is accounted for according to urban population calculates.Education level can according to university student in every 100000 population quantity into
Row calculates.Gini coefficient can be calculated according to the income groups data in national statistical yearbook.
Step 202, according at least one influence factor, determine the weighing factor of at least one influence factor.
Wherein, the weighing factor of influence factor can periodically update.In the prediction of this progress caseload,
If the cycle of its renewal is not reaching to, and when the weighing factor of at least one influence factor has calculated, the dress of case prediction
The weighing factor of each influence factor need not be recalculated by putting, and only need to store each influence according at least one influence factor
In the memory space of the weighing factor of factor, the weighing factor of each influence factor is obtained.
If reach the update cycle of the weighing factor of influence factor in the prediction of this progress caseload, or at least
When the weighing factor of one influence factor does not calculate, at least one influence need to be determined by calculation out in the device of case prediction
The weighing factor of factor.At this time, the device of case prediction determines at least one influence factor according at least one influence factor
Weighing factor includes：Determine n history crime case number of packages amount, n history value of at least one influence factor.According to n history
Crime case quantity, n history value of at least one influence factor, using batch processing gradient descent method, determines at least one shadow
The weighing factor of the factor of sound.
Wherein, the device of case prediction is after at least one influence factor is determined, according at least one influence factor, from
Store the n history value that each influence factor is obtained in the memory space of the history value of influence factor.If do not have in memory space
N history value of at least one influence factor is recorded, the device of case prediction at this time can be set according to the actual requirements by user
Put n history value of at least one influence factor, and by by n history value of at least one influence factor of user setting,
It is determined as n history value of at least one influence factor.After n history value of at least one influence factor is determined, utilize
N history value of at least one influence factor determined, n history crime case number of packages amount, by batch processing gradient descent method,
It can determine the weighing factor of at least one influence factor.
Further, the device of case prediction is a according to n history crime case number of packages amount, the n of at least one influence factor
History value, using batch processing gradient descent method, determining the weighing factor of at least one influence factor includes：
The device of case prediction is according to n history crime case number of packages amount, n history value of at least one influence factor, profit
Use formula：Determine the weighing factor of at least one influence factor.
Wherein, j represents j-th of influence factor, θjRepresenting the weighing factor of j-th of influence factor, α represents gradient,Table
Show i-th of history crime case number of packages amount.hθ(x(i)) represent i-th of the crime case quantity predicted,Represent j-th influence because
I-th of history value of element.N represents the number of history crime case number of packages amount, and represents the number of the history value of influence factor, and i is
More than 0, and the integer less than or equal to n.:=represent is iterated calculating.
Specifically, the device of case prediction is obtaining the history value of at least one influence factor and n history crime case
After number of packages amount, formula is carried it intoCalculating is iterated, until θjConvergence,
θ can be calculatedj。
α in step 202 can be pre-set according to the actual requirements.
As above described in example, the device of case prediction passes through x respectively1, x2, x3, x4, x5It is expressed as determining domestic per capita raw
Total value is produced, education level, the level of urbanization, Gini coefficient, five influence factors of the size of population, the device of case prediction can be with
After above-mentioned five influence factors are determined, the weighing factor of each influence factor need to be further obtained.In order to obtain each shadow
The weighing factor of the factor of sound, the device of case prediction need to obtain n history crime case number of packages amount, and the n of above-mentioned 5 influence factors
A history value.At this time, the device of case prediction can obtain required n in the database of storage history crime case number of packages amount
A history crime case number of packages amount.And storage influence factor history value memory space in obtain respectively it is above-mentioned 5 influence because
N history value of element.The data that the device of case prediction obtains are as shown in table 1 below.
Table 1
Wherein,Represent i-th of history crime case number of packages amount.Represent i-th of history of per-capita gross domestic product
Value,Represent i-th of history value of education level,I-th of history value of the level of urbanization,I-th of Gini coefficient
History value,I-th of history value of the size of population.I=1,2 ... ... n.
The device of case prediction is a in the n for obtaining n required history crime case number of packages amount and above-mentioned 5 influence factors
After history value, formula can be utilized5 influence factors are determined respectively
Weighing factor.
Step 203, the history value for obtaining at least one influence factor, and going through according at least one influence factor of acquisition
History value, using time series models, determines the predicted value of at least one influence factor.
Wherein, n is the integer more than 0.
Specifically, due to when carrying out the prediction of crime case quantity, it is necessary to predicted value using influence factor.Therefore,
The device of case prediction needs first to determine the predicted value of at least one influence factor.At this time, the device of case prediction can be first
The history value of at least one influence factor is obtained respectively, according to the history value of at least one influence factor of acquisition, passage time
Series model can determine the predicted value of at least one influence factor.
Wherein, time series models be it is well known in the prior art it is a kind of obtain predicted value method, the present invention to this not
Repeat again.
Further, needed for the device of case prediction can be obtained from the memory space of the history value of storage influence factor
Influence factor history value.The device of case prediction can obtain multiple history values of influence factor, for example, obtain influence because
20 history values of element.Depending on the device of case prediction specifically obtains several history values of influence factor according to the actual requirements, this
Invention is not restricted this.
Step 204, according to the pre- of the weighing factor of at least one influence factor and at least one influence factor determined
Measured value, predicts crime case quantity.
Specifically, case prediction device after the weighing factor of each influence factor and predicted value is determined, Ke Yizhi
Connect and the weighing factor of each influence factor and predicted value are calculated into product respectively, obtain the corresponding crime case of each influence factor
Number of packages amount, and then the crime case quantity of each influence factor is subjected to summation operation, calculate total crime case quantity.
Further, it is contemplated that error condition that may be present, the device of case prediction is according at least one influence determined
The predicted value of the weighing factor of factor and at least one influence factor, utilizes formula：
Predict crime case quantity.
Wherein, y and hθ(x) the crime case quantity of prediction, θ are represented0Expression system constant, θjRepresent influence factor xj
Weighing factor, xjRepresent j-th of influence factor, m is the number of influence factor, and ε represents random error.
In this step, θ0Can be that user is pre-set according to the actual requirements.
In this step, at least one influence factor has m, at this time, xjIn j value difference then xjRepresent different shadows
The factor of sound.
It should be noted that in embodiments of the present invention, after the history value and predicted value of influence factor each mean quantization
Value.
An embodiment of the present invention provides a kind of method of case prediction, including according to history crime case number of packages amount, determine
At least one influence factor；Determine the weighing factor of at least one influence factor；According to history crime case number of packages amount, the time is utilized
Series model, determines the predicted value of at least one influence factor；Weighed according to the influence at least one influence factor determined
The predicted value of weight and at least one influence factor, predicts crime case quantity.In this way, the device of case prediction is carrying out
During the prediction of crime case data, influence factor is first determined, and then determine the weighing factor of influence factor, and according to the time
The method of series model obtains the predicted value of influence factor, it is achieved thereby that weighing factor and influence factor according to influence factor
Predicted value carry out crime case quantity prediction.As, the device of case prediction be according to the influence factor of crime case into
The prediction of row crime case, rather than just the quantity based on Time Series Analysis Forecasting crime case, therefore, the present invention carries
The high accuracy of caseload prediction.
Although disclosed herein embodiment as above, the content be only readily appreciate the present invention and use
Embodiment, is not limited to the present invention.Technical staff in any fields of the present invention, is taken off not departing from the present invention
On the premise of the spirit and scope of dew, any modification and change, but the present invention can be carried out in the form and details of implementation
Scope of patent protection, still should be subject to the scope of the claims as defined in the appended claims.
Claims (6)
- A kind of 1. device of case prediction, it is characterised in that including：Determination unit, at least one influence factor is determined for receiving prediction instruction, and according to the prediction instruction；According to described At least one influence factor, determines the weighing factor of at least one influence factor；Acquiring unit, for obtaining the history value of at least one influence factor；The determination unit, is additionally operable to the history value of at least one influence factor obtained according to the acquiring unit, profit With time series models, the predicted value of at least one influence factor is determined；Predicting unit, for the weighing factor of at least one influence factor determined according to the determination unit and described The predicted value of at least one influence factor, predicts crime case quantity；Wherein, the determination unit, specifically for determining n history crime case number of packages amount, n of at least one influence factor go through History value；According to the n history crime case number of packages amount, n history value of at least one influence factor, using under batch processing gradient Drop method, determines the weighing factor of at least one influence factor；N is the integer more than 0；The determination unit, specifically for being gone through according to the n history crime case number of packages amount, n of at least one influence factor History value, utilizes formula：Determine the influence of at least one influence factor Weight；Wherein, j is integer more than 0, θjRepresenting the weighing factor of j-th of influence factor, α represents gradient,Represent i-th History crime case number of packages amount；hθ(x(i)) represent i-th of the crime case quantity predicted,Represent the i-th of j-th of influence factor A history value, n represent the number of history crime case number of packages amount, and represent the number of the history value of influence factor, i be more than 0, and Integer less than or equal to n :=represent to be iterated calculating.
- 2. the device of case prediction according to claim 1, it is characterised in thatThe predicting unit, specifically for according to the weighing factor of at least one influence factor determined and it is described at least The predicted value of one influence factor, utilizes formula：Predict crime case quantity； Wherein, y and hθ(x) the crime case quantity of prediction, θ are represented0Expression system constant, θjRepresent influence factor xjInfluence power Weight, xjRepresent j-th of influence factor, m is the number of influence factor, and ε represents random error.
- 3. the device of case prediction according to claim 1, it is characterised in thatAt least one Guessing Factors are carried in prediction instruction；The determination unit, specifically for obtaining n history crime case number of packages amount；According to the n history crime case of acquisition Number of packages amount, determines at least one influence factor at least one Guessing Factors.
- A kind of 4. method of case prediction, it is characterised in that including：Prediction instruction is received, and at least one influence factor is determined according to the prediction instruction；According at least one influence factor, the weighing factor of at least one influence factor is determined；Obtain the history value of at least one influence factor, and the history of at least one influence factor according to acquisition Value, using time series models, determines the predicted value of at least one influence factor；N is the integer more than 0；According to the weighing factor at least one influence factor determined and the predicted value of at least one influence factor, Predict crime case quantity；Wherein, described according at least one influence factor, determining the weighing factor of at least one influence factor includes：Determine n history crime case number of packages amount, n history value of at least one influence factor；According to the n history crime case number of packages amount, n history value of at least one influence factor, using under batch processing gradient Drop method, determines the weighing factor of at least one influence factor；It is described according to the n history crime case number of packages amount, the n history values of at least one influence factor, utilize batch processing gradient Descent method, determining the weighing factor of at least one influence factor includes：According to the n history crime case number of packages amount, n history value of at least one influence factor, utilizes formula：Determine the weighing factor of at least one influence factor；Wherein, j is big In 0 integer, θjRepresenting the weighing factor of j-th of influence factor, α represents gradient,Represent i-th of history crime case number of packages Amount；hθ(x(i)) represent i-th of the crime case quantity predicted,Represent i-th of history value of j-th of influence factor, n is represented The number of history crime case number of packages amount, and represent the number of the history value of influence factor, i be more than 0, it is and whole less than or equal to n Number :=represent to be iterated calculating.
- 5. the method for case according to claim 4 prediction, it is characterised in that the basis determine described at least one The predicted value of the weighing factor of a influence factor and at least one influence factor, predicting crime case quantity includes：According to the weighing factor at least one influence factor determined and the predicted value of at least one influence factor, Utilize formula：Predict crime case quantity；Wherein, y and hθ(x) represent pre- The crime case quantity of survey, θ0Expression system constant, θjRepresent influence factor xjWeighing factor, xjRepresent j-th influence because Element, m are the number of influence factor, and ε represents random error.
- 6. the method for case prediction according to claim 4, it is characterised in thatAt least one Guessing Factors are carried in the prediction instruction；It is described to be instructed according to the prediction, determine that at least one influence factor includes：Obtain n history crime case number of packages amount；According to the n history crime case number of packages amount of acquisition, at least one shadow is determined at least one Guessing Factors The factor of sound.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510570751.1A CN105117605B (en) | 2015-09-09 | 2015-09-09 | A kind of device and method of case prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510570751.1A CN105117605B (en) | 2015-09-09 | 2015-09-09 | A kind of device and method of case prediction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105117605A CN105117605A (en) | 2015-12-02 |
| CN105117605B true CN105117605B (en) | 2018-05-01 |
Family
ID=54665593
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510570751.1A Active CN105117605B (en) | 2015-09-09 | 2015-09-09 | A kind of device and method of case prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105117605B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845723B (en) * | 2017-02-06 | 2020-12-29 | 安徽新华博信息技术股份有限公司 | Criminal case occurrence prediction method |
| CN110351523B (en) * | 2019-07-22 | 2020-10-16 | 常州机电职业技术学院 | Building video monitoring system and video monitoring adjustment method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103903430A (en) * | 2014-04-14 | 2014-07-02 | 东南大学 | Dynamic fusion type travel time predicting method with multi-source and isomorphic data adopted |
-
2015
- 2015-09-09 CN CN201510570751.1A patent/CN105117605B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103903430A (en) * | 2014-04-14 | 2014-07-02 | 东南大学 | Dynamic fusion type travel time predicting method with multi-source and isomorphic data adopted |
Non-Patent Citations (3)
| Title |
|---|
| "Temporal Association, the Dynamics of Crime, and their Economic Determinants: A Time Series Econometric Model of South Africa";John M.Luiz;《Social Indicators Research》;20011231;第53卷(第1期);全文 * |
| "基于时间序列分析的边防刑事案件预测研究";杨毓等;《廊坊师范学院学报（自然科学版）》;20141031;第14卷(第5期);第7-8页 * |
| "基于贝叶斯因素-时间序列模型的罪犯地理特征仿真研究";黄冠利等;《数学的实践与认识》;20110630;第41卷(第11期);第58-67页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105117605A (en) | 2015-12-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108230049A (en) | The Forecasting Methodology and system of order | |
| CN108364085B (en) | Takeout delivery time prediction method and device | |
| Wanke et al. | An analysis of Asian airlines efficiency with two-stage TOPSIS and MCMC generalized linear mixed models | |
| CN111091196B (en) | Passenger flow data determination method and device, computer equipment and storage medium | |
| CN111259590A (en) | Construction method of concrete dam deformation safety monitoring model | |
| CN111860977A (en) | Probability prediction method and probability prediction device for short-term load | |
| Baran et al. | Joint probabilistic forecasting of wind speed and temperature using Bayesian model averaging | |
| CN110503267B (en) | Urban financial invasion case prediction system and prediction method based on space-time scale self-adaptive model | |
| CN104217258A (en) | Method for power load condition density prediction | |
| CN114493052B (en) | Multi-model fusion self-adaptive new energy power prediction method and system | |
| CN106528608A (en) | Cold and hot storage method and system for power grid GIS (Geographic Information System) data in cloud architecture | |
| CN105117605B (en) | A kind of device and method of case prediction | |
| CN115796932B (en) | Project cost prediction method and device, electronic equipment and storage medium | |
| CN105389631B (en) | More Sea-Crossing Passage distributed areas traffic accessibility analysis methods | |
| CN103870890A (en) | Prediction method for traffic flow distribution of expressway network | |
| CN115829081A (en) | Urban traffic carbon emission prediction method based on support vector regression model | |
| CN114048920A (en) | Site selection layout method, device, equipment and storage medium for charging facility construction | |
| Elnashar et al. | Modeling pavement damage and predicting fatigue cracking of flexible pavements based on a combination of deterministic method with stochastic approach using Miner’s hypothesis | |
| Li et al. | Research on the method of traffic organization and optimization based on dynamic traffic flow model | |
| CN104217296A (en) | Listed company performance comprehensive evaluation method | |
| CN110689055A (en) | Cross-scale statistical index spatialization method considering grid unit attribute grading | |
| CN106157163A (en) | A kind of grain yield short term prediction method and device | |
| CN112990530B (en) | Regional population quantity prediction method, regional population quantity prediction device, electronic equipment and storage medium | |
| CN113642886A (en) | Method, system, medium and computing device for analyzing utilization rate of planned power grid line | |
| Zhang et al. | Accuracy assessments and uncertainty analysis of spatially explicit modeling for land use/cover change and urbanization: A case in Beijing metropolitan area |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20180820 Address after: 200436 Room 411, No. three, JIANGCHANG Road, Jingan District, Shanghai, 411 Patentee after: Shanghai wave Cloud Computing Service Co., Ltd. Address before: 100085 floor 1, C 2-1, No. 2, Shang Di Road, Haidian District, Beijing. Patentee before: Electronic information industry Co.,Ltd of the tide (Beijing) |
|
| TR01 | Transfer of patent right |