CN106294763A - A kind of system modeling method and device - Google Patents
A kind of system modeling method and device Download PDFInfo
- Publication number
- CN106294763A CN106294763A CN201610657290.6A CN201610657290A CN106294763A CN 106294763 A CN106294763 A CN 106294763A CN 201610657290 A CN201610657290 A CN 201610657290A CN 106294763 A CN106294763 A CN 106294763A
- Authority
- CN
- China
- Prior art keywords
- subset
- regression model
- mean square
- square deviation
- parameter
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/211—Schema design and management
- G06F16/212—Schema design and management with details for data modelling support
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Complex Calculations (AREA)
Abstract
The embodiment of the invention discloses a kind of system modeling method and device, be applied to data analysis and processing technology field.The method comprise the steps that acquisition M from the set containing N number of parameterfIndividual parameter forms the first subset；P the second subset is extracted from the supplementary set of the first subset；P regression model is set up according to the first subset and P the second subset；The ranking results of the mean square deviation according to regression model each in P regression model is that the parameter in the second subset composes weights, then weights and maximum parameter in P the second subset are added in the first subset and obtain destination subset, finally the regression model precision in destination subset meets requirement, or when the number of parameters in destination subset is equal to predetermined value, using the regression model of destination subset as goal regression model.Implement the embodiment of the present invention, by selecting the parameter of predetermined value number to be modeled in all parameters, not only reduce the burden that system prediction calculates, improve the robustness of model simultaneously.
Description
Technical field
The present invention relates to data analysis and processing technology field, particularly relate to a kind of system modeling method and device.
Background technology
It is the basis of research and analysis system that model is set up, and is also one of the key technology of data analysis and process.Only
Establishing corresponding system model, could be further analyzed system and process, reliability and the degree of accuracy of model are
Follow-up signal analysis and process work have meaning and the important prerequisite of value and guarantee.Model can be built according to the mechanism of system
It is vertical, it is also possible to by the experiment of system or statistical disposition are obtained.For multiparameter complication system, the variable mistake affected due to it
Many, its Analysis on Mechanism is often difficult to obtain clear model accurately, along with the development of computer technology, is set up by experimental data
The method of model is found broad application.
At present, relatively common multiple linear regression model is typically all used as system model, data to be analyzed
And process, but, in actual use, it is not completely independent between parameters, between parameter, there may be linear pass
System and make partial parameters redundancy, add in multiparameter complication system, the number of parameter is the biggest, not only increases and is
The burden that system prediction calculates, also results in Expired Drugs and occurs so that the model robustness of foundation is poor.
Summary of the invention
Embodiments provide a kind of system modeling method and device, by selecting predetermined value in all parameters
The parameter of number is modeled, and not only reduces the burden that system prediction calculates, improves the robustness of model simultaneously.
Embodiment of the present invention first aspect discloses a kind of system modeling method, including:
Obtain the complete or collected works comprising N number of parameter, from described complete or collected works, choose MfIndividual parameter obtains the first subset；
Extracting P the second subset from the supplementary set of described first subset, the number of parameters that each second subset comprises is
Mr, and MrWith MfAnd equal to predetermined value；
Described first subset and each second subset are respectively combined and obtain P the 3rd subset, according to described P the 3rd son
P regression model set up respectively by collection；
Calculate the mean square deviation of each regression model in described P regression model, and to calculated described mean square
Difference sorts；
According to the ranking results of described mean square deviation, the parameter in described second subset is composed weights, wherein, described mean square
The weights of the parameter in described second subset of the least correspondence of difference are the biggest；
Calculate the sum of the weights of each parameter in described P the second subset, target component is added in described first subset
Obtaining destination subset, wherein, described target component is weights and maximum parameter；
If the mean square deviation of the regression model of described destination subset is less than the first predetermined threshold value, or, described destination subset
In number of parameters equal to described predetermined value, then using the regression model of described destination subset as goal regression model.
As the optional embodiment of one, described using the regression model of described destination subset as goal regression model it
Before, described method also includes:
Judge whether the mean square deviation of the regression model of described destination subset is less than the first predetermined threshold value, and, it is judged that institute
Whether state the number of parameters in destination subset equal to described predetermined value.
As the optional embodiment of one, the mean square deviation of the described regression model judging described destination subset is the least
In the first predetermined threshold value, and, it is judged that after whether the number of parameters in described destination subset is equal to described predetermined value, described side
Method also includes:
If the mean square deviation of the regression model of described destination subset is more than the first predetermined threshold value, and, described destination subset
In number of parameters less than described predetermined value, then perform to extract the step of P the second subset from the supplementary set of described destination subset,
Number of parameters in the number of parameters comprised in each second subset and described destination subset and equal to described predetermined value.
As the optional embodiment of one, described using the regression model of described destination subset as goal regression model it
After, described method also includes:
Judge that whether the mean square deviation of described goal regression model is more than described first predetermined threshold value；
The most then according to the number of the parameter in the preset rules described destination subset of increase, until utilizing the ginseng after increasing
The mean square deviation of the regression model that several numbers are set up is less than described first predetermined threshold value, and, until current regression model is equal
Residual quantity between the mean square deviation of variance yields and a upper regression model is less than the second predetermined threshold value；
Determine that the number of the parameter that a described upper regression model is corresponding, and will be described upper one time for optimum predetermined value
Return model as optimum regression model.
As the optional embodiment of one, whether the described mean square deviation judging described goal regression model is more than described
After first predetermined threshold value, described method also includes:
If the mean square deviation of described goal regression model is less than described first predetermined threshold value, then reduce institute according to preset rules
State the number of parameter in destination subset, until utilizing the mean square deviation of the regression model that the number of parameters after reducing sets up to be more than
Described first predetermined threshold value, or, until the mean square deviation of mean square deviation and a upper regression model of current regression model it
Between residual quantity more than the 3rd predetermined threshold value；
Determine the number of the parameter that a described upper regression model is corresponding for optimum predetermined value, and by a described upper recurrence
Model is as optimum regression model.
Embodiment of the present invention second aspect discloses a kind of system modelling device, including:
First acquiring unit, for obtaining the complete or collected works comprising N number of parameter, chooses M from described complete or collected worksfIndividual parameter obtains
One subset；
Second acquisition unit, for extracting P the second subset, each second subset bag from the supplementary set of described first subset
The number of parameters contained is Mr, and MrWith MfAnd equal to predetermined value；
Unit set up by model, obtains P the 3rd subset, root for described first subset and each second subset being respectively combined
P regression model is set up respectively according to described P the 3rd subset；
Mean square deviation processing unit is for calculating the mean square deviation of each regression model in described P regression model and right
Calculated described mean square deviation sorts；
Weights processing unit, composes power for the ranking results according to described mean square deviation to the parameter in described second subset
Value, wherein, the weights of the parameter in described second subset of the least correspondence of described mean square deviation are the biggest；
Destination subset determines unit, for calculating the sum of the weights of each parameter in described P the second subset, by target component
Adding in described first subset and obtain destination subset, wherein, described target component is weights and maximum parameter；
Object module determines unit, for presetting threshold at the mean square deviation of the regression model of described destination subset less than first
Value, or, the number of parameters in described destination subset equal to described predetermined value time, using the regression model of described destination subset as
Goal regression model.
As the optional embodiment of one, described device also includes:
First judging unit, the most pre-less than first for judging the mean square deviation of the regression model of described destination subset
If threshold value, and, it is judged that whether the number of parameters in described destination subset is equal to described predetermined value.
As the optional embodiment of one,
Described second acquisition unit, is additionally operable to the mean square deviation at the regression model of described destination subset and presets more than first
Threshold value, and, when the number of parameters in described destination subset is less than described predetermined value, extract from the supplementary set of described destination subset
P the second subset, the number of parameters comprised in each second subset and the number of parameters in described destination subset and equal to institute
State predetermined value.
As the optional embodiment of one, described device also includes:
Second judging unit, for judging whether the mean square deviation of described goal regression model presets threshold more than described first
Value；
First adjustment unit, is used for when the mean square deviation of described goal regression model is more than described predetermined threshold value, according to
Preset rules increases the number of the parameter in described destination subset, until utilizing the regression model that the number of parameters after increasing is set up
Mean square deviation less than described first predetermined threshold value, and, until the mean square deviation of current regression model returns mould with upper one
Residual quantity between the mean square deviation of type is less than the second predetermined threshold value；
First adjusts processing unit, predetermined for optimum for determining the number of parameter that a described upper regression model is corresponding
Value, and using a described upper regression model as optimum regression model.
As the optional embodiment of one, described device also includes:
Second adjustment unit, is used for when the mean square deviation of described goal regression model is less than described first predetermined threshold value,
According to the number of the parameter in the preset rules described destination subset of reduction, until utilizing the recurrence that the number of parameters after reducing is set up
The mean square deviation of model is more than described first predetermined threshold value, or, until the mean square deviation of current regression model returns with upper one
Return the residual quantity between the mean square deviation of model more than the 3rd predetermined threshold value；
Second adjusts processing unit, predetermined for optimum for determining the number of parameter that a described upper regression model is corresponding
Value, and using a described upper regression model as optimum regression model.
As can be seen from the above technical solutions, the embodiment of the present invention has the advantage that from the set containing N number of parameter
Obtain MfIndividual parameter forms the first subset, and extracts P the second subset from the supplementary set of this first subset；According to the first subset
And P the second subset sets up P regression model；The sequence knot of the mean square deviation according to regression model each in P regression model
Fruit is that the parameter in the second subset composes weights, and wherein, the weights of the parameter in the second subset of the least correspondence of mean square deviation are the biggest,
Then weights in all second subsets and maximum parameter are added in the first subset and obtain destination subset, finally at target
The mean square deviation of the regression model of collection is less than the first predetermined threshold value, or, when the number of parameters in destination subset is equal to predetermined value,
Using the regression model of destination subset as goal regression model.Implement the embodiment of the present invention, pre-by selecting in all parameters
The parameter of definite value number is modeled, and not only reduces the burden that system prediction calculates, improves the robustness of model simultaneously.
Accompanying drawing explanation
For the technical scheme being illustrated more clearly that in the embodiment of the present invention, in embodiment being described below required for make
Accompanying drawing briefly introduce, it should be apparent that, below describe in accompanying drawing be only some embodiments of the present invention, for this
From the point of view of the those of ordinary skill in field, on the premise of not paying creative work, it is also possible to obtain it according to these accompanying drawings
His accompanying drawing.
Fig. 1 is the schematic flow sheet of a kind of system modeling method disclosed in the embodiment of the present invention；
Fig. 2 is the schematic flow sheet of another kind of system modeling method disclosed in the embodiment of the present invention；
Fig. 3 is a kind of number of parameters disclosed in the embodiment of the present invention and the schematic flow sheet of model optimization method；
Fig. 4 is the structural representation of a kind of device disclosed in the embodiment of the present invention；
Fig. 5 is the structural representation of another kind of device disclosed in the embodiment of the present invention；
Fig. 6 is the entity structure schematic diagram of another kind of device disclosed in the embodiment of the present invention.
Detailed description of the invention
In order to make the object, technical solutions and advantages of the present invention clearer, below in conjunction with accompanying drawing the present invention made into
One step ground describes in detail, it is clear that described embodiment is only some embodiments of the present invention rather than whole enforcement
Example.Based on the embodiment in the present invention, those of ordinary skill in the art are obtained under not making creative work premise
All other embodiments, broadly fall into the scope of protection of the invention.
Term " first ", " second " and " the 3rd " in description and claims of this specification and above-mentioned accompanying drawing is to use
In the different object of difference, not for describing particular order.Additionally, term " includes " and they any deformation, it is intended that
Cover non-exclusive comprising.Such as contain series of steps or the process of unit, method, system, product or equipment do not limit
Due to the step listed or unit, but the most also include step or the unit do not listed, or the most also include right
In intrinsic other step of these processes, method, product or equipment or unit.
Embodiments provide a kind of system modeling method and device, by selecting predetermined value in all parameters
The parameter of number is modeled, and not only reduces the burden that system prediction calculates, improves the robustness of model simultaneously.Hereinafter will divide
It is not described in detail.
Referring to Fig. 1, Fig. 1 is the schematic flow sheet of a kind of system modeling method disclosed in the embodiment of the present invention.Wherein, figure
System modeling method shown in 1 may comprise steps of:
101: obtain the complete or collected works comprising N number of parameter, from these complete or collected works, choose MfIndividual parameter obtains the first subset；
In the embodiment of the present invention, can obtain, by experiment, the complete or collected works comprising N number of parameter, the number of N depends on system
Feature and the means of experiment and method.And choose M from these complete or collected worksfIndividual parameter obtains the first subset, wherein, MfNumber permissible
Value and empirical parameter according to N are determined.
102: from the supplementary set of above-mentioned first subset, extract P the second subset, the number of parameters that each second subset comprises
It is Mr, and MrWith MfAnd equal to predetermined value；
In the embodiment of the present invention, determined containing M by step 101fAfter first subset of individual parameter, can be from the first subset
Supplementary set in extract P the second subset, Monte Carlo method can be used to randomly draw P time from the supplementary set of the first subset, and general
The M extracted each timerIndividual parameter forms the second subset, can have altogether P the second subset, wherein, MrWith MfAnd equal to predetermined
Value, this predetermined value represents the number bringing the parameter setting up regression model extracted from N number of parameter, can be by system performance
And the empirical parameter obtained determines.Wherein, the number of P determines the precision of system model, and P is the biggest, more may obtain accurate
Regression model, but amount of calculation can increase, it is therefore desirable to select suitable number of times P according to system requirements compromise.
103: above-mentioned first subset and each second subset are respectively combined and obtain P the 3rd subset, according to above-mentioned P the 3rd
Subset sets up P regression model respectively；
In the embodiment of the present invention, after obtaining P the second subset by step 102, can above-mentioned first subset and each the
Two subsets are respectively combined and obtain P the 3rd subset, and wherein, the number of parameters in the 3rd subset is equal to above-mentioned predetermined value.And according to
This P the 3rd subset sets up P regression model respectively, wherein it is possible to use method of least square to set up regression model.
104: calculate the mean square deviation of each regression model in above-mentioned P regression model, and to calculated mean square
Difference sorts；
In the embodiment of the present invention, can estimating according to experiment sample number and experimental system output valve and regression model
Evaluation calculates the mean square deviation of regression model.And calculate the mean square deviation of each regression model in P regression model successively
Value, and be ranked up by the size order of mean square deviation.
105: according to the ranking results of mean square deviation, the parameter in above-mentioned second subset is composed weights, wherein, mean square deviation
The weights of the parameter in the second subset of the least correspondence are the biggest；
In the embodiment of the present invention, after mean square deviation being ranked up by step 104, can be according to the row of mean square deviation
Parameter in second subset is composed weights by sequence result, and mean square deviation is the least, then the regression model that this mean square deviation is corresponding
The weights of the parameter in the second subset are the biggest.
106: calculate the sum of the weights of each parameter in above-mentioned P the second subset, target component is added to above-mentioned first son
Concentration obtains destination subset, and wherein, above-mentioned target component is weights and maximum parameter；
In the embodiment of the present invention, it is, after the parameter in each second subset composes weights, P can be calculated by step 105
The sum of the weights that parameters in the second subset is corresponding, and weights and maximum parameter are added in above-mentioned first subset
To destination subset.
107: if the mean square deviation of the regression model of above-mentioned destination subset is less than the first predetermined threshold value, or, above-mentioned target
Number of parameters in subset is equal to predetermined value, then using the regression model of above-mentioned destination subset as goal regression model.
In the embodiment of the present invention, after obtaining the destination subset after the first subset is updated by step 106, if root
It is less than the first predetermined threshold value according to the mean square deviation of the regression model of destination subset foundation, or, the number of parameters in destination subset
Equal to predetermined value, then using the regression model set up according to destination subset as goal regression model, wherein, the first predetermined threshold value can
To determine according to test or empirical value, for judging whether the precision of regression model meets requirement.
In the method described by Fig. 1, from the set containing N number of parameter, obtain MfIndividual parameter forms the first subset, and
P the second subset is extracted from the supplementary set of this first subset；P recurrence is set up according to the first subset and P the second subset
Model；The ranking results of the mean square deviation according to regression model each in P regression model is that the parameter in the second subset composes weights,
Wherein, the weights of the parameter in the second subset of the least correspondence of mean square deviation are the biggest, then by weights in all second subsets and
Maximum parameter adds in the first subset and obtains destination subset, finally at the mean square deviation of regression model of destination subset less than the
One predetermined threshold value, or, when the number of parameters in destination subset is equal to predetermined value, using the regression model of destination subset as target
Regression model.Implement the embodiment of the present invention, by selecting the parameter of predetermined value number to be modeled, not only in all parameters
Reduce the burden that system prediction calculates, improve the robustness of model simultaneously.
Further, referring to Fig. 2, Fig. 2 is that disclosed in the embodiment of the present invention, the flow process of another kind of system modeling method is shown
It is intended to.Wherein, the system modeling method shown in Fig. 2 may comprise steps of:
201: obtain the complete or collected works comprising N number of parameter, from these complete or collected works, choose MfIndividual parameter obtains the first subset；
202: from the supplementary set of above-mentioned first subset, extract P the second subset, the number of parameters that each second subset comprises
It is Mr, and MrWith MfAnd equal to predetermined value；
203: above-mentioned first subset and each second subset are respectively combined and obtain P the 3rd subset, according to above-mentioned P the 3rd
Subset sets up P regression model respectively；
204: calculate the mean square deviation of each regression model in above-mentioned P regression model, and to calculated mean square
Difference sorts；
205: according to the ranking results of mean square deviation, the parameter in above-mentioned second subset is composed weights, wherein, mean square deviation
The weights of the parameter in the second subset of the least correspondence are the biggest；
206: calculate the sum of the weights of each parameter in above-mentioned P the second subset, target component is added to above-mentioned first son
Concentration obtains destination subset, and wherein, above-mentioned target component is weights and maximum parameter；
207: judge whether the mean square deviation of the regression model of above-mentioned destination subset is less than the first predetermined threshold value, and, sentence
Whether the number of parameters in disconnected above-mentioned destination subset is equal to described predetermined value；
In the embodiment of the present invention, it is that the first subset interpolation target component obtains destination subset by step 201～step 206
Afterwards, it can be determined that the mean square deviation of the regression model of destination subset whether less than the first predetermined threshold value, and, it is judged that target
Whether the number of parameters concentrated is equal to predetermined value, and wherein, the first predetermined threshold value may be used to determine model and whether meets precision and want
Asking, predetermined value represents the number of parameters needed in modeling process.
208: if the mean square deviation of the regression model of above-mentioned destination subset is less than the first predetermined threshold value, or, above-mentioned target
Number of parameters in subset is equal to predetermined value, then using the regression model of above-mentioned destination subset as goal regression model；
In the embodiment of the present invention, if the regression model of destination subset meets the parameter in required precision, or destination subset
Number reaches predetermined value, then can using utilize above-mentioned destination subset to set up regression model as goal regression model.
Alternatively, if the mean square deviation of the regression model of above-mentioned destination subset is more than the first predetermined threshold value, and, above-mentioned mesh
Number of parameters in mark subset less than described predetermined value, then performs to extract P the second subset from the supplementary set of this destination subset
Step, the number of parameters in the number of parameters comprised in each second subset and this destination subset and equal to above-mentioned predetermined value.
In the embodiment of the present invention, if the regression model precision utilizing above-mentioned destination subset to set up is unsatisfactory for requirement, and, mesh
Number of parameters in mark subset less than predetermined value, then needs to continue as adding in destination subset parameter, at this time, it may be necessary to circulation performs
Step 202～step 207, and, when performing 202, the supplementary set of the first subset stated in step 202 represents interpolation parameter
The supplementary set of the first subset afterwards, the number of parameters in P the second subset is equal to predetermined value and the first son added after parameter
The difference of the number of parameters concentrated.Until meeting the condition in step 208, end loop, it is thus achieved that goal regression model.
Further, the selection of the number of the predetermined value in embodiment 1 and embodiment 2 is extremely important, if predetermined value
Excessive, the most not only increase amount of calculation but also there may be Expired Drugs, if predetermined value is too small, then possibly cannot obtain can
Lean on, accurate model.Therefore, the embodiment of the present invention specifically describes the method how optimizing predetermined value, refers to Fig. 3, Fig. 3 table
Show a kind of number of parameters and the schematic flow sheet of model optimization method, wherein, the number of parameters shown in Fig. 3 and model optimization method
May comprise steps of:
301: judge that whether the mean square deviation of above-mentioned goal regression model is more than the first predetermined threshold value；
In the embodiment of the present invention, goal regression model representation is according to the system modeling method root in embodiment 1 and embodiment 2
The regression model set up according to predetermined value, this predetermined value represents the number of parameters of the modeling needs chosen from N number of parameter.By step
Rapid 301 may determine that whether the precision of goal regression model meets requirement.
302: the most then according to the number of the parameter in preset rules increase destination subset, until utilizing the ginseng after increasing
The mean square deviation of the regression model that several numbers are set up is less than the first predetermined threshold value, and, until the mean square deviation of current regression model
Residual quantity between value and the mean square deviation of a upper regression model is less than the second predetermined threshold value；
303: the number determining on this parameter that regression model is corresponding is optimum predetermined value, and by this recurrence
Model is as optimum regression model；
In the embodiment of the present invention, if the precision of goal regression model is unsatisfactory for requirement, i.e. the mean square deviation of goal regression model
Value more than the first predetermined threshold value, then can increase in destination subset according to preset rules (increasing a parameter) the most every time
The number of parameter, and set up back according to the number of the parameter after increasing according to the system modeling method in embodiment 1 and embodiment 2
Return model, and judge whether the mean square deviation of this regression model is less than the first predetermined threshold value (i.e. whether precision meets requirement), with
And, whether the residual quantity between mean square deviation and the mean square deviation of a upper regression model of this regression model presets threshold less than second
Value (i.e. judges that the precision of regression model is improved the most notable), if the mean square deviation of current regression model is more than the first predetermined threshold value
(i.e. precision is unsatisfactory for requiring), and between mean square deviation and the mean square deviation of a upper regression model of current regression model
Residual quantity is more than the second predetermined threshold value (i.e. the precision of regression model is improved notable), then circulation execution increases current according to preset rules
The step of number of parameters, until utilizing the mean square deviation of the regression model of the number of parameters foundation after increasing to preset threshold less than first
Value, and, until the residual quantity between mean square deviation and the mean square deviation of a upper regression model of current regression model is less than the
Two predetermined threshold value.Then represent that "current" model precision meets requirement, and the upper model accuracy of "current" model precision improves not
Significantly.The number then determining on this parameter that regression model is corresponding is optimum predetermined value, and by this regression model
As optimum regression model.
304: if the mean square deviation of goal regression model is less than the first predetermined threshold value, then reduce target according to preset rules
The number of the parameter concentrated, until utilizing the mean square deviation of the regression model of the number of parameters foundation after reducing to preset more than first
Threshold value, or, until the residual quantity between mean square deviation and the mean square deviation of a upper regression model of current regression model is more than
3rd predetermined threshold value；
305: the number determining on this parameter that regression model is corresponding is optimum predetermined value, and by this recurrence
Model is as optimum regression model.
In the embodiment of the present invention, if the precision of goal regression model meets requirement, i.e. the mean square deviation of goal regression model
Less than the first predetermined threshold value, then can reduce the parameter in destination subset according to preset rules (reducing a parameter) the most every time
Number, and according to the system modeling method in embodiment 1 and embodiment 2 according to reduce after parameter number set up return mould
Type, and judge whether the mean square deviation of this regression model is less than the first predetermined threshold value (i.e. whether precision meets requirement), and, should
Residual quantity between mean square deviation and the mean square deviation of a upper regression model of regression model is more than the 3rd predetermined threshold value (the most i.e.
Judge that the deteriorated accuracy of regression model is the most notable), if the mean square deviation of current regression model is less than the first predetermined threshold value (i.e. essence
Degree meets requirement), or the residual quantity between mean square deviation and the mean square deviation of a upper regression model of current regression model is little
In the 3rd predetermined threshold value (i.e. the deteriorated accuracy of regression model is the most notable), then circulation performs to reduce parameter current according to preset rules
The step of number, until utilizing the mean square deviation of the regression model of the number of parameters foundation after reducing to be more than the first predetermined threshold value,
Or, until the residual quantity between mean square deviation and the mean square deviation of a upper regression model of current regression model is pre-more than the 3rd
If threshold value.Then represent that "current" model precision is unsatisfactory for requirement, or, the upper model accuracy of "current" model precision deteriorates aobvious
Write.The number then determining on this parameter that regression model is corresponding is optimum predetermined value, and is made by this regression model
For optimum regression model.
For the system modeling method in clearer explanation above-described embodiment and number of parameters and model optimization method,
The embodiment of the present invention is described further as a example by noinvasive Hemoglobin Meter.
Noinvasive Hemoglobin Meter is the instrument of a kind of noninvasive detection human body hemoglobin concentration, its principle be use the reddest
External spectrum method, according to Lambertian-Bill (lambert-beer) law, light is molten with this by its absorbed degree after solution
The parameters such as the absorptance of liquid, concentration, light path, wavelength of light are correlated with, by the light transmission tissue of multiple wave bands, through remarkable
After internal blood absorption, the degree of absorption of its light contains the concentration information of hemoglobin components in blood, thereby through
Contact between the two, can calculate total hemoglobin concentration, carboxyhemoglobin concentration and metahemoglobin concentration etc.
Physiological parameter.Owing to human tissue structure is complicated, blood constituent is the most varied, therefore is difficult to set up from mechanism sufficiently accurate
Reflection light absorption degree and the model of hemoglobin concentration relation, and generally require and obtain its mould by analysis of experimental data
Type.
As a kind of preferred embodiment, the light of the selection 8 wavelength in the range of 600nm～1300nm is (respectively
It is designated as λ1、λ2...λ8) difference the most sequentially transmission finger, 8 light are turned by photoelectricity after blood absorption some light in finger respectively
Emat sensor receives and is converted to the signal of telecommunication, this signal through the process such as hardware circuit amplifications, filtering, data acquisition laggard enter
CPU becomes digital signal.Simultaneously by gathering blood and utilizing blood gas analyzer to obtain invasive hemoglobin concentration as ginseng
Examine data.By data analysis and process, set up light digital signal and the regression model of hemoglobin concentration and for estimating
With prediction hemoglobin concentration.
If the light intensity received is Ii, i=1,2...8, due to the fluctuation of finger pulse, its alternating component ACi=Δ
IiThe fluctuating margin of expression light, and flip-flopRepresent the mean intensity of light.
Order
According to Lambertian-Beer law, hemoglobin concentration value and RijClosely related, R can be set upijWith hemoglobin
The regression model of concentration value.
Set up RijPower series parameter setWhereinRepresent RijN power, in this example
Only return with the data within 4 powers, do not consider 5 powers of parameter and above Data Regression Model.Invasive vim and vigour are divided
The reference hemoglobin concentration that analysis obtains is respectively total hemoglobin concentration YtHb, carboxyhemoglobin concentration YCOHb, high ferro blood red
Protein concentration YMetHb.As a example by total hemoglobin concentration, the most only illustrate the foundation of regression model.
Obviously the element number of parameter set X is that N=224 (8*7*4) is individual, it is necessary to sets up and relatively simplifies and return mould reliably
Type.The number of parameters initializing regression model is M=30, sets up regression model as follows:
1, two subsets set of set of system parameters X are determinedfAnd setr, wherein setfElement number is Mf(setfInitial value is
Sky, Mf=0), setrIt is at set according to Monte Carlo methodfSupplementary set (X-setfThe element number randomly drawed in) is Mr=M-
MfSubset.
2, by setfAnd setrTwo subsets are combined intoAnd utilize method of least square to set upRecurrence
Model.And calculate the mean square deviation of this regression modelWherein L is the number of sample, YkWithRespectively
Represent system output value and the estimated value of regression model thereof of kth sample.
3, step 1～2 is repeated, until obtaining p regression model.By obtain p regression model press mean square deviation err from
Little to being ranked up greatly, and give corresponding setrElement in subset gives weights (the least weights of err are the biggest) respectively.Calculate institute
There is setrIn subset the weights of element and, and be X by weights and maximum element definitionmax。
4, by element XmaxUpdate to subset setfIn, even setf=setfl+{Xmax, wherein setflRepresent the last time
setfCollection.
5, step 1～4 is repeated, until setfElement number reach M, or the precision of regression model meets and requires (i.e.
Mean square deviation is less than specifying threshold value), and regression model now is defined as system model.
And the value of number of parameters M of final mask can be determined by the steps:
1, determine that M initial value is 30 according to the empirical parameter of system performance and acquisition, and set up by above-mentioned modeling method
Current regression model also obtains corresponding mean square deviation.If the precision of "current" model is unsatisfactory for requiring (such as mean square deviation
errcMore than specifying threshold value) then proceed to step 2, the precision of "current" model has met requirement and has then proceeded to step 4 else if.
2, increase number of parameters and (M can be madec=Ml+ 1, MlNumber of parameters for last time modeling), set up current McIndividual ginseng
Several regression models also calculates corresponding mean square deviation errc。
3, step 2 is repeated, until the precision of "current" model reaches requirement (such as err≤1.0), and "current" model essence
Spend more last model accuracy and improve not notable (such as errl-errc≤ 0.02, wherein errlFor last model i.e. MlIndividual parameter
The mean square deviation of regression model).Determine final number of parameters M=Ml, its corresponding regression model is final model.
4, reduce number of parameters and (M can be madec=Ml-1), current M is set upcThe regression model of individual parameter calculating are corresponding all
Variance yields errc。
5, step 4 is repeated, until the more last model accuracy of "current" model precision significantly deteriorates (such as errl-errc＞
, or "current" model precision is unsatisfactory for requirement 0.05).Determine final number of parameters M=Ml, its corresponding regression model is
Final model.
According to the method for above-mentioned modeling, the present embodiment finally establishes the regression model of 12 parameters of total hemoglobin,
Mean square deviation is 0.95.And after the same method, establishing the regression model of 20 parameters of carboxyhemoglobin, mean square deviation is
1.7,8 Partial Linear Models of metahemoglobin, mean square deviation is 0.91.
Referring to Fig. 4, Fig. 4 is the structural representation of a kind of system modelling device disclosed in the embodiment of the present invention, such as Fig. 4 institute
Showing, this device may include that
First acquiring unit 401, for obtaining the complete or collected works comprising N number of parameter, chooses M from above-mentioned complete or collected worksfIndividual parameter obtains
To the first subset；
Second acquisition unit 402, for extracting P the second subset from the supplementary set of above-mentioned first subset, each second son
The number of parameters that collection comprises is Mr, and MrWith MfAnd equal to predetermined value；
Unit 403 set up by model, for the first subset obtained by above-mentioned first acquiring unit 401 and second acquisition unit
402 each second subsets obtained are respectively combined and obtain P the 3rd subset, set up P respectively according to the above-mentioned P of stating the 3rd subset and return
Return model；
Mean square deviation processing unit 404, for calculate above-mentioned model set up unit 403 set up P regression model in each
The mean square deviation of individual regression model, and calculated mean square deviation is sorted；
Weights processing unit 405, the ranking results of the mean square deviation for obtaining according to above-mentioned mean square deviation processing unit 404
Parameter in above-mentioned second subset is composed weights, and wherein, the weights of the parameter in the second subset of the least correspondence of mean square deviation are more
Greatly；
Destination subset determines unit 406, for the above-mentioned P of weight computing given according to above-mentioned weights processing unit 405
The sum of the weights of each parameter in second subset, adds the first subset that above-mentioned first acquiring unit 401 obtains to by target component
In obtain destination subset, wherein, above-mentioned target component is weights and maximum parameters；
Object module determines unit 407, pre-less than first for the mean square deviation of the regression model in above-mentioned destination subset
If threshold value, or, when the number of parameters in above-mentioned destination subset is equal to above-mentioned predetermined value, above-mentioned target component is determined unit
The regression model of 406 destination subset determined is as goal regression model.
Seeing also Fig. 5, Fig. 5 is the structural representation of another kind of device disclosed in the embodiment of the present invention.Wherein, Fig. 5
Shown device is that device as shown in Figure 4 is optimized and obtains, and compared with the device shown in Fig. 4, said apparatus also includes:
First judging unit 408, for judging that whether the mean square deviation of the regression model of above-mentioned destination subset is less than first
Predetermined threshold value, and, it is judged that whether the number of parameters in above-mentioned destination subset is equal to above-mentioned predetermined value.
Alternatively, in the device end shown in Fig. 5,
Above-mentioned second acquisition unit 402, is additionally operable to the mean square deviation at the regression model of above-mentioned destination subset and is more than first
Predetermined threshold value, and, when the number of parameters in above-mentioned destination subset is less than above-mentioned predetermined value, from the supplementary set of above-mentioned destination subset
Extraction P the second subset, the number of parameters comprised in each second subset and the number of parameters in above-mentioned destination subset and etc.
In above-mentioned predetermined value.
Alternatively, in the arrangement as shown in fig. 5, this device can also include:
Second judging unit 409, for judging that the mean square deviation of above-mentioned goal regression model is whether first pre-more than above-mentioned
If threshold value；
First adjustment unit 410, for above-mentioned goal regression model mean square deviation more than above-mentioned state predetermined threshold value time,
According to the number of the parameter in the preset rules above-mentioned destination subset of increase, until utilizing the recurrence that the number of parameters after increasing is set up
The mean square deviation of model is less than the first predetermined threshold value, and, until the mean square deviation of current regression model returns mould with upper one
Residual quantity between the mean square deviation of type is less than the second predetermined threshold value；
First adjusts processing unit 411, for determining on this that number of parameter that regression model is corresponding is for optimum pre-
Definite value, and using on this regression model as optimum regression model.
Alternatively, in the arrangement as shown in fig. 5, this device can also include:
Second adjustment unit 412, for when the mean square deviation of above-mentioned goal regression model is less than the first predetermined threshold value, pressing
According to the number of the parameter in the preset rules above-mentioned destination subset of reduction, until utilizing the recurrence mould that the number of parameters after reducing is set up
The mean square deviation of type is more than the first predetermined threshold value, or, until the mean square deviation of current regression model and a upper regression model
Mean square deviation between residual quantity more than the 3rd predetermined threshold value；
Second adjusts processing unit 413, for determining on this that number of parameter that regression model is corresponding is for optimum pre-
Definite value, and using on this regression model as optimum regression model.
Refer to the structural representation that Fig. 6, Fig. 6 are a kind of system modelling devices that the embodiment of the present invention provides.Such as Fig. 6 institute
Showing, a kind of system modelling device 600 that the embodiment of the present invention provides may include that at least one bus 601, is connected with bus
At least one processor 602 and at least one memorizer 603 being connected with bus.
Wherein, processor 602, by bus 601, calls the code of storage in memorizer 603 and comprises for acquisition N number of
The complete or collected works of parameter, choose M from described complete or collected worksfIndividual parameter obtains the first subset；P is extracted from the supplementary set of described first subset
Second subset, the number of parameters that each second subset comprises is Mr, and MrWith MfAnd equal to predetermined value；By described first son
Collection and each second subset are respectively combined and obtain P the 3rd subset, set up P regression model respectively according to described P the 3rd subset；
Calculate the mean square deviation of each regression model in described P regression model, and calculated described mean square deviation is sorted；
According to the ranking results of described mean square deviation, the parameter in described second subset being composed weights, wherein, described mean square deviation is the least
The weights of the corresponding parameter in described second subset are the biggest；Calculate the sum of the weights of each parameter in described P the second subset, will
Target component is added in described first subset and is obtained destination subset, and wherein, described target component is weights and maximum ginseng
Number；If the mean square deviation of the regression model of described destination subset is less than the first predetermined threshold value, or, the ginseng in described destination subset
Several numbers are equal to described predetermined value, then using the regression model of described destination subset as goal regression model.
Alternatively, in some possible embodiments of the present invention, described processor 602 calls in memorizer 603 to be deposited
The code of storage, using the regression model of described destination subset as before goal regression model, be additionally operable to:
Judge whether the mean square deviation of the regression model of described destination subset is less than the first predetermined threshold value, and, it is judged that institute
Whether state the number of parameters in destination subset equal to described predetermined value.
Alternatively, in some possible embodiments of the present invention, described processor 602 calls in memorizer 603 to be deposited
The code of storage, whether the mean square deviation at the regression model judging described destination subset is less than the first predetermined threshold value, and, it is judged that
Whether the number of parameters in described destination subset, equal to after described predetermined value, is additionally operable to:
If the mean square deviation of the regression model of described destination subset is more than the first predetermined threshold value, and, described destination subset
In number of parameters less than described predetermined value, then perform to extract the step of P the second subset from the supplementary set of described destination subset,
Number of parameters in the number of parameters comprised in each second subset and described destination subset and equal to described predetermined value.
Alternatively, in some possible embodiments of the present invention, described processor 602 calls in memorizer 603 to be deposited
The code of storage, using the regression model of described destination subset as after goal regression model, be additionally operable to:
Judge that whether the mean square deviation of described goal regression model is more than described first predetermined threshold value；
The most then according to the number of the parameter in the preset rules described destination subset of increase, until utilizing the ginseng after increasing
The mean square deviation of the regression model that several numbers are set up is less than described first predetermined threshold value, and, until current regression model is equal
Residual quantity between the mean square deviation of variance yields and a upper regression model is less than the second predetermined threshold value；
Determine the number of the parameter that a described upper regression model is corresponding for optimum predetermined value, and by a described upper recurrence
Model is as optimum regression model.
Alternatively, in some possible embodiments of the present invention, described processor 602 calls in memorizer 603 to be deposited
The code of storage, after whether the mean square deviation judging described goal regression model is more than described first predetermined threshold value, is additionally operable to:
If the mean square deviation of described goal regression model is less than described first predetermined threshold value, then reduce institute according to preset rules
State the number of parameter in destination subset, until utilizing the mean square deviation of the regression model that the number of parameters after reducing sets up to be more than
Described first predetermined threshold value, or, until the mean square deviation of mean square deviation and a upper regression model of current regression model it
Between residual quantity more than the second predetermined threshold value；
Determine the number of the parameter that a described upper regression model is corresponding for optimum predetermined value, and by a described upper recurrence
Model is as optimum regression model.
It is understood that the function of each functional module of the system modelling device 600 of the present embodiment can be according to above-mentioned side
Method in method embodiment implements, and it implements process and is referred to the associated description of said method embodiment, herein
Repeat no more.
The embodiment of the present invention also provides for a kind of computer-readable storage medium, and wherein, this computer-readable storage medium can store journey
Sequence, this program includes the part or all of step of the system modeling method described in said method embodiment when performing.
It should be noted that for aforesaid each method embodiment, in order to be briefly described, therefore it is all expressed as one it be
The combination of actions of row, but those skilled in the art should know, and the present invention is not limited by described sequence of movement, because of
For according to the present invention, some step can use other orders or carry out simultaneously.Secondly, those skilled in the art also should know
Knowing, embodiment described in this description belongs to preferred embodiment, involved action and the module not necessarily present invention
Necessary.
In the above-described embodiments, the description to each embodiment all emphasizes particularly on different fields, and does not has the portion described in detail in certain embodiment
Point, may refer to the associated description of other embodiments.
In several embodiments provided herein, it should be understood that disclosed device, can be by another way
Realize.Such as, device embodiment described above is only schematically, and the division of the most described unit is only one
Logic function divides, actual can have when realizing other dividing mode, the most multiple unit or assembly can in conjunction with or can
To be integrated into another system, or some features can be ignored, or does not performs.Another point, shown or discussed each other
Coupling direct-coupling or communication connection can be the INDIRECT COUPLING by some interfaces, device or unit or communication connection,
Can be being electrical or other form.
The described unit illustrated as separating component can be or may not be physically separate, shows as unit
The parts shown can be or may not be physical location, i.e. may be located at a place, or can also be distributed to multiple
On NE.Some or all of unit therein can be selected according to the actual needs to realize the mesh of the present embodiment scheme
's.
It addition, each functional unit in various embodiments of the present invention can be integrated in a processing unit, it is possible to
Being that unit is individually physically present, it is also possible to two or more unit are integrated in a unit.Above-mentioned integrated
Unit both can realize to use the form of hardware, it would however also be possible to employ the form of SFU software functional unit realizes.
If described integrated unit realizes and as independent production marketing or use using the form of SFU software functional unit
Time, can be stored in a computer read/write memory medium.Based on such understanding, technical scheme is substantially
The part that in other words prior art contributed or this technical scheme completely or partially can be with the form of software product
Embodying, this computer software product is stored in a storage medium, including some instructions with so that a computer
Equipment (can be for personal computer, server or the network equipment etc.) perform the whole of method described in each embodiment of the present invention or
Part steps.And aforesaid storage medium includes: USB flash disk, read only memory (ROM, Read-Only Memory), random access memory
Memorizer (RAM, RandomAccess Memory), portable hard drive, magnetic disc or CD etc. are various can store program code
Medium.
The above, above example only in order to technical scheme to be described, is not intended to limit；Although with reference to front
State embodiment the present invention has been described in detail, it will be understood by those within the art that: it still can be to front
State the technical scheme described in each embodiment to modify, or wherein portion of techniques feature is carried out equivalent；And these
Amendment or replacement, do not make the essence of appropriate technical solution depart from the scope of various embodiments of the present invention technical scheme.
Claims (10)
1. a system modeling method, it is characterised in that including:
Obtain the complete or collected works comprising N number of parameter, from described complete or collected works, choose MfIndividual parameter obtains the first subset；
Extracting P the second subset from the supplementary set of described first subset, the number of parameters that each second subset comprises is Mr, and
MrWith MfAnd equal to predetermined value；
Described first subset and each second subset are respectively combined and obtain P the 3rd subset, according to described P the 3rd subset respectively
Set up P regression model；
Calculate the mean square deviation of each regression model in described P regression model, and to calculated described mean square deviation
Sequence；
According to the ranking results of described mean square deviation, the parameter in described second subset is composed weights, wherein, described mean square deviation
The weights of the parameter in described second subset of the least correspondence are the biggest；
Calculate the sum of the weights of each parameter in described P the second subset, target component is added in described first subset and obtain
Destination subset, wherein, described target component is weights and maximum parameter；
If the mean square deviation of the regression model of described destination subset is less than the first predetermined threshold value, or, in described destination subset
Number of parameters is equal to described predetermined value, then using the regression model of described destination subset as goal regression model.
Method the most according to claim 1, it is characterised in that the described regression model using described destination subset returns as target
Before returning model, described method also includes:
Judge whether the mean square deviation of the regression model of described destination subset is less than the first predetermined threshold value, and, it is judged that described mesh
Whether the number of parameters in mark subset is equal to described predetermined value.
Method the most according to claim 2, it is characterised in that the mean square deviation of the regression model of the described destination subset of described judgement
Value whether less than the first predetermined threshold value, and, it is judged that the number of parameters in described destination subset whether equal to described predetermined value it
After, described method also includes:
If the mean square deviation of the regression model of described destination subset is more than the first predetermined threshold value, and, in described destination subset
Number of parameters is less than described predetermined value, then perform to extract the step of P the second subset from the supplementary set of described destination subset, each
Number of parameters in the number of parameters comprised in second subset and described destination subset and equal to described predetermined value.
Method the most according to claim 3, it is characterised in that the described regression model using described destination subset returns as target
After returning model, described method also includes:
Judge that whether the mean square deviation of described goal regression model is more than described first predetermined threshold value；
The most then according to the number of the parameter in the preset rules described destination subset of increase, until utilizing the parameter after increasing
The mean square deviation of the regression model that number is set up is less than described first predetermined threshold value, and, until the mean square deviation of current regression model
Residual quantity between value and the mean square deviation of a upper regression model is less than the second predetermined threshold value；
Determine the number of the parameter that a described upper regression model is corresponding for optimum predetermined value, and by a described upper regression model
As optimum regression model.
Method the most according to claim 4, it is characterised in that whether the described mean square deviation judging described goal regression model
After described first predetermined threshold value, described method also includes:
If the mean square deviation of described goal regression model is less than described first predetermined threshold value, then reduce described mesh according to preset rules
The number of the parameter in mark subset, until utilizing the mean square deviation of the regression model of the number of parameters foundation after reducing more than described
First predetermined threshold value, or, until between mean square deviation and the mean square deviation of a upper regression model of current regression model
Residual quantity is more than the 3rd predetermined threshold value；
Determine the number of the parameter that a described upper regression model is corresponding for optimum predetermined value, and by a described upper regression model
As optimum regression model.
6. a system modelling device, it is characterised in that including:
First acquiring unit, for obtaining the complete or collected works comprising N number of parameter, chooses M from described complete or collected worksfIndividual parameter obtains the first son
Collection；
Second acquisition unit, for extracting P the second subset from the supplementary set of described first subset, each second subset comprises
Number of parameters is Mr, and MrWith MfAnd equal to predetermined value；
Unit set up by model, obtains P the 3rd subset, according to institute for described first subset and each second subset being respectively combined
State P the 3rd subset and set up P regression model respectively；
Mean square deviation processing unit, for calculating the mean square deviation of each regression model in described P regression model, and to calculating
The described mean square deviation sequence obtained；
Weights processing unit, composes weights for the ranking results according to described mean square deviation to the parameter in described second subset,
Wherein, the weights of the parameter in described second subset of the least correspondence of described mean square deviation are the biggest；
Destination subset determines unit, for calculating the sum of the weights of each parameter in described P the second subset, target component is added
Obtaining destination subset in described first subset, wherein, described target component is weights and maximum parameter；
Object module determines unit, and the mean square deviation for the regression model in described destination subset is less than the first predetermined threshold value,
Or, when the number of parameters in described destination subset is equal to described predetermined value, using the regression model of described destination subset as mesh
Mark regression model.
Device the most according to claim 6, it is characterised in that described device also includes:
First judging unit, for judging whether the mean square deviation of the regression model of described destination subset presets threshold less than first
Value, and, it is judged that whether the number of parameters in described destination subset is equal to described predetermined value.
Device the most according to claim 7, it is characterised in that
Described second acquisition unit, is additionally operable to the mean square deviation at the regression model of described destination subset and presets threshold more than first
Value, and, when the number of parameters in described destination subset is less than described predetermined value, from the supplementary set of described destination subset, extract P
Individual second subset, the number of parameters comprised in each second subset and the number of parameters in described destination subset and equal to described
Predetermined value.
Device the most according to claim 8, it is characterised in that described device also includes:
Second judging unit, for judging that whether the mean square deviation of described goal regression model is more than described first predetermined threshold value；
First adjustment unit, for when the mean square deviation of described goal regression model is more than described predetermined threshold value, according to presetting
The number of the regular parameter increased in described destination subset, until utilizing the equal of the regression model of the number of parameters foundation after increasing
Variance yields is less than described first predetermined threshold value, and, until the mean square deviation of current regression model and a upper regression model
Residual quantity between mean square deviation is less than the second predetermined threshold value；
First adjusts processing unit, for determining that the number of parameter that a described upper regression model is corresponding is optimum predetermined value,
And using a described upper regression model as optimum regression model.
Device the most according to claim 9, it is characterised in that described device also includes:
Second adjustment unit, is used for when the mean square deviation of described goal regression model is less than described first predetermined threshold value, according to
Preset rules reduces the number of the parameter in described destination subset, until utilizing the regression model that the number of parameters after reducing is set up
Mean square deviation more than described first predetermined threshold value, or, until the mean square deviation of current regression model returns mould with upper one
Residual quantity between the mean square deviation of type is more than the 3rd predetermined threshold value；
Second adjusts processing unit, for determining that the number of parameter that a described upper regression model is corresponding is optimum predetermined value,
And using a described upper regression model as optimum regression model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610657290.6A CN106294763A (en) | 2016-08-11 | 2016-08-11 | A kind of system modeling method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610657290.6A CN106294763A (en) | 2016-08-11 | 2016-08-11 | A kind of system modeling method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106294763A true CN106294763A (en) | 2017-01-04 |
Family
ID=57669651
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610657290.6A Pending CN106294763A (en) | 2016-08-11 | 2016-08-11 | A kind of system modeling method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106294763A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112545461A (en) * | 2020-12-05 | 2021-03-26 | 深圳市美的连医疗电子股份有限公司 | Method, device and system for detecting non-invasive hemoglobin concentration value and computer readable storage medium |
-
2016
- 2016-08-11 CN CN201610657290.6A patent/CN106294763A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112545461A (en) * | 2020-12-05 | 2021-03-26 | 深圳市美的连医疗电子股份有限公司 | Method, device and system for detecting non-invasive hemoglobin concentration value and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11564612B2 (en) | Automatic recognition and classification method for electrocardiogram heartbeat based on artificial intelligence | |
| EP1739580B1 (en) | Categorization including dependencies between different category systems | |
| CN109785928A (en) | Diagnosis and treatment proposal recommending method, device and storage medium | |
| CN107578060A (en) | A kind of deep neural network based on discriminant region is used for the method for vegetable image classification | |
| CN105701120B (en) | The method and apparatus for determining semantic matching degree | |
| CN107064054A (en) | A kind of near-infrared spectral analytical method based on CC PLS RBFNN Optimized models | |
| CN107822622A (en) | Electrocardiographic diagnosis method and system based on depth convolutional neural networks | |
| CN110033022A (en) | Processing method, device and the storage medium of text | |
| CN109785959A (en) | A kind of disease code method and apparatus | |
| CN107832737A (en) | Electrocardiogram interference identification method based on artificial intelligence | |
| CN107330750A (en) | A kind of recommended products figure method and device, electronic equipment | |
| CN110363229A (en) | A kind of characteristics of human body's parameter selection method combined based on improvement RReliefF and mRMR | |
| KR20180064863A (en) | SMI automatic analysis method of hand-wrist radiation images using deep learning | |
| Ismail et al. | Discriminating the early stages of Sirex noctilio infestation using classification tree ensembles and shortwave infrared bands | |
| CN112560964A (en) | Method and system for training Chinese herbal medicine pest and disease identification model based on semi-supervised learning | |
| WO2021151295A1 (en) | Method, apparatus, computer device, and medium for determining patient treatment plan | |
| CN110111885A (en) | Attribute forecast method, apparatus, computer equipment and computer readable storage medium | |
| CN113128585B (en) | Deep neural network based multi-size convolution kernel method for realizing electrocardiographic abnormality detection and classification | |
| CN107146160A (en) | Customer insured's analysis on the health status method and server | |
| CN109711707A (en) | A kind of Ship Power Equipment synthetical condition assessment method | |
| CN107480426B (en) | Self-iteration medical record file clustering analysis system | |
| CN106294763A (en) | A kind of system modeling method and device | |
| US11462331B2 (en) | Method and system for pressure autoregulation based synthesizing of photoplethysmogram signal | |
| CN108304930A (en) | Network pruning method, apparatus and computer readable storage medium | |
| CN113034528A (en) | Target area and organ-at-risk delineation contour accuracy testing method based on image omics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170104 |