CN112036662A - Method for establishing regional flow prediction model and regional flow prediction method - Google Patents
Method for establishing regional flow prediction model and regional flow prediction method Download PDFInfo
- Publication number
- CN112036662A CN112036662A CN202010947047.4A CN202010947047A CN112036662A CN 112036662 A CN112036662 A CN 112036662A CN 202010947047 A CN202010947047 A CN 202010947047A CN 112036662 A CN112036662 A CN 112036662A
- Authority
- CN
- China
- Prior art keywords
- input
- flow
- output table
- output
- area
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 93
- 230000003993 interaction Effects 0.000 claims abstract description 71
- 230000008569 process Effects 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 238000004519 manufacturing process Methods 0.000 claims description 6
- 238000007689 inspection Methods 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 abstract description 3
- 230000000295 complement effect Effects 0.000 abstract 1
- 238000011161 development Methods 0.000 description 3
- 230000002452 interceptive effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 239000004566 building material Substances 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000005034 decoration Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Abstract
The invention is suitable for the technical field of data prediction optimization, and provides a method for establishing a regional flow prediction model and a regional flow prediction method, wherein the method for establishing the regional flow prediction model comprises the following steps: in a reference annual input-output table consisting of N areas, historical flow interaction quantity between an area i and N-1 arbitrary areas j and the historical flow interaction quantity of the N-1 areas j in an extreme value range constrained by a closed economic system are randomly generated to complement the reference annual input-output table; and obtaining a complete target annual input-output table by utilizing the RAS method and combining the given value of the (N + 1) th row and the given value of the (N + 1) th column in the target annual input-output table and the supplemented reference annual input-output table, and constructing and obtaining a multiple nonlinear regression model, namely a regional flow prediction model, through the input-output relation of the input-output data set. The regional flow prediction model provided by the invention has the advantages of less required historical data, high reliability and small error.
Description
Technical Field
The invention relates to the technical field of data prediction optimization, in particular to a method for establishing a regional flow prediction model and a regional flow prediction method.
Background
Accurate regional flow prediction has important significance for the development of industries such as logistics, input-output analysis, prediction of population flow among regions and the like. For example, in the logistics industry, accurate regional flow prediction can provide accurate scientific basis for formulating regional logistics development policies, determining the construction scale of logistics infrastructure, planning road traffic networks and the like. Obviously, the aspects all play a vital role in the healthy development of the logistics industry and the social economy.
Currently, the quantitative models commonly used for regional flow prediction include trend and time series models, system dynamics models, input-output analysis models, and the like. These models predict regional flows from different angles. However, due to various reasons such as insufficient availability of historical data, too complicated and difficult realization of the method, and insufficient degree of understanding of the method by a practitioner, the currently used regional flow prediction method is estimated or predicted by a simple method depending on subjective experience of the practitioner to a considerable extent, and thus the prediction result has large error and low reliability.
Disclosure of Invention
The invention mainly aims to provide a method for establishing a regional flow prediction model and a regional flow prediction method, which are used for solving the problems of large error and low reliability of prediction results because the existing regional flow prediction mode depends on subjective prediction of practitioners.
In order to achieve the above object, a first aspect of embodiments of the present invention provides a method for building a regional flow prediction model, including:
s1, forming a benchmark annual input-output table consisting of N areas based on a closed economic system comprising the N areas; the input-output table of the reference year is N +1 rows and N +1 columns, the value of the N +1 row is the sum of all the columns, and the sum of the jth column represents the total output flow of the area j in the reference year; the value of the N +1 th column is the sum of each row, the sum of the ith row represents the total input flow of the area i in the reference year, and N is a positive integer;
s2, acquiring historical flow interaction quantity between the area i and the N-1 areas j in the reference year input-output table; historical flow interaction quantity between the area i and the N-1 areas j comprises input flow of the area i to the N-1 areas j and output flow of the N-1 areas j to the area i, wherein i and j are positive integers smaller than or equal to N;
s3, filling the input flow of the area i into the N-1 areas j into the ith row of the reference annual input-output table, and filling the output flow of the N-1 areas j to the area i into the ith column of the reference annual input-output table;
s4, randomly generating historical flow interaction quantities of N-1 areas j in an extreme value range constrained by the closed economic system, and complementing the benchmark annual input-output table by using the random historical flow interaction quantities to obtain a complemented benchmark annual input-output table; the historical traffic interaction quantity of the N-1 regions j comprises the historical traffic interaction quantity of any two regions in the N-1 regions j;
s5, obtaining a given value of the N +1 th row and a given value of the N +1 th column in the target annual input-output table according to the value of the N +1 th row and the value of the N +1 th column in the reference annual input-output table and a preset flow rate increase rate, wherein the given value of the N +1 th row in the target annual input-output table is the sum of all the columns, and the sum of the j th column represents the total output flow rate of the area j in the target year; the given value of the N +1 th column is the sum of all rows, and the sum of the ith row represents the total input flow of the area i in the target year;
s6, updating the supplemented standard annual input-output table by combining the given value of the (N + 1) th row and the given value of the (N + 1) th column in the target annual input-output table by using an RAS method to obtain a complete target annual input-output table;
s7, repeating S4-S6 to obtain an input and output data set generated in the process of updating the supplemented reference annual input and output table by using an RAS method to obtain a complete target annual input and output table; and constructing a multiple nonlinear regression model by fitting the input-output relation of the input-output data set, and taking the multiple nonlinear regression model as a regional flow prediction model.
Optionally, after the step S7, the method further includes:
and introducing a non-constant term model to fit the multiple nonlinear regression model to obtain an optimized regional flow prediction model.
Optionally, the step S7 includes:
calculating a first ratio of total input flow rate for zone i in the target year to total input flow rate in the target year and a second ratio of total output flow rate for zone j in the target year to total output flow rate in the target year based on the complemented reference year input-output table and the complete target year input-output table;
constraining a maximum value and a minimum value of a flow rate increase by the first ratio and the second ratio, and calculating a flow rate increase ratio based on the maximum value and the minimum value;
the multivariate nonlinear regression model takes the historical flow interaction quantity of the region i and the N-1 regions j, the maximum value and the minimum value of the flow growth rate, the flow growth rate ratio, the first ratio and the second ratio as input quantities, and takes the flow interaction quantity of the region i and the N-1 regions j in the complete target year input-output table as output quantities corresponding to the input quantities;
the input quantity and the output quantity are input and output data generated in the process of updating the supplemented reference annual input-output table by using an RAS method to obtain a complete target annual input-output table;
repeating S4-S6 to obtain an input and output data set, and fitting the input and output relation of the input and output data set;
and selecting a target fitting result through goodness-of-fit inspection, and taking variables and coefficients in the target fitting result as parameters of the multivariate nonlinear regression model to complete the construction of the multivariate nonlinear regression model.
Optionally, a first ratio of total input flow rate in target year for zone i to total input flow rate in target year and a second ratio of total output flow rate in target year for zone j to total output flow rate in target year are calculated based on the complemented reference year input-output table and the complete target year input-output table, and the formula is as follows:
wherein ,αiIs a first ratio, βiIn order to be the second ratio, the first ratio,
the total input flow of the area i in the input-output table of the reference year,
for the total throughput of region j in the input-output table for the reference year,
the total input flow of the area i in the target annual input-output table and the complete target annual input-output table,
the total production flow of the area j in the target annual input-output table and the complete target annual input-output table.
Optionally, a target fitting result is selected through goodness-of-fit inspection, and a variable and a coefficient in the target fitting result are used as parameters of the multiple nonlinear regression model to complete the construction of the multiple nonlinear regression model, wherein the formula is as follows:
wherein ,
for the regional flow to be predicted,
historical traffic interaction quantity alpha of the region i and the N-1 regions j in the input-output table for the reference yeariIs a first ratio, βiIs a second ratio, tminAs a minimum value of the rate of increase of the flow, tmaxThe maximum value of the flow rate increase, and γ is a flow rate increase ratio based on the maximum value of the flow rate increase and the minimum value of the flow rate increase.
Optionally, a non-constant term model is introduced to fit the multiple nonlinear regression model to obtain an optimized regional flow prediction model, and the formula is as follows:
wherein ,
for the regional flow to be predicted,
historical traffic interaction quantity alpha of the region i and the N-1 regions j in the input-output table for the reference yeariIs a first ratio, βiIs a second ratio, tminAs a minimum value of the rate of increase of the flow, tmaxThe maximum value of the flow rate increase, and γ is a flow rate increase ratio based on the maximum value of the flow rate increase and the minimum value of the flow rate increase.
A second aspect of the embodiments of the present invention provides a method for predicting regional traffic, including:
acquiring a benchmark annual input-output table, wherein the benchmark annual input-output table is based on a closed economic system comprising N areas and consists of N areas; the input-output table of the reference year is N +1 rows and N +1 columns, the given value of the N +1 row is the sum of all the columns, and the sum of the jth column represents the total output flow of the area j in the reference year; the given value of the N +1 th column is the sum of each row, the sum of the ith row represents the total input flow of the area i in the reference year, and N is a positive integer;
acquiring historical traffic interaction quantity of an area to be predicted based on the benchmark year input-output table, wherein the historical traffic interaction quantity of the area to be predicted comprises historical traffic interaction quantity between an area i and N-1 areas j; historical flow interaction quantity between the area i and the N-1 areas j comprises input flow of the area i to the N-1 areas j and output flow of the N-1 areas j to the area i, wherein i and j are positive integers smaller than or equal to N;
and predicting the area flow of the area to be predicted by using the area flow prediction model according to the historical flow interaction amount of the area to be predicted.
A third aspect of embodiments of the present invention provides a terminal device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the methods provided in the first and second aspects when executing the computer program.
A fourth aspect of embodiments of the present invention provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the steps of the method as provided in the first and second aspects above.
The embodiment of the invention provides a method for establishing a regional flow prediction model, which comprises the steps of randomly generating historical flow interactive quantity of N-1 regions j in an extreme value range of flow of each region in a closed economic system, complementing a reference annual input-output table by utilizing the random historical flow interactive quantity to obtain the complemented reference annual input-output table, and simultaneously obtaining a given value of an N +1 row and a given value of an N +1 column in a target annual input-output table according to a preset flow growth rate according to a given value of an N +1 row and a given value of an N +1 column in the reference annual input-output table, so that basic data of the historical flow interactive quantity between a region i and the N-1 regions j, total output flow of the region j in a reference year, total input flow of the region i in the reference year, the given value of the N +1 row and the given value of the N +1 column in the target annual input-output table are obtained, updating the supplemented standard annual input-output table by an RAS method to obtain a complete target annual input-output table; wherein the complete target annual input-output table is based on the prediction result of the supplemented reference annual input-output table. Then, on the basis, obtaining an input and output data set obtained by an RAS method through a large number of repeated experiments, and constructing a multiple nonlinear regression model by fitting the input and output relations of the input and output data set so as to express the data input and output relations of the RAS method; when the historical data is insufficient, the method provided by the embodiment of the invention can be used for establishing the regional flow prediction model to obtain a prediction result equivalent to that of the classical RAS method. In summary, the embodiment of the present invention provides a regional flow prediction model, which requires less historical data, has a prediction result equivalent to that of the classical RAS method, is high in reliability, small in error, has a definite input/output mathematical relationship, is easy to implement, and has a wide application range.
Drawings
Fig. 1 is a schematic flow chart illustrating an implementation of a method for building a regional flow prediction model according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of the benchmark annual input-output table obtained in step S1 in FIG. 1;
FIG. 3 is a schematic diagram of the complemented benchmark annual input-output table obtained in steps S2-S4 in FIG. 1;
FIG. 4 is a benchmark annual input table 4a obtained based on FIG. 3, and a complete target annual input-output table 4b updated based on the benchmark annual input-output table 4 a;
fig. 5 is a schematic flow chart illustrating an implementation of another method for building a regional flow prediction model according to an embodiment of the present invention;
FIG. 6 is a flowchart illustrating a detailed implementation of step S4 in FIG. 1;
fig. 7 is a schematic flow chart illustrating an implementation of the regional flow prediction method in the embodiment of the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Suffixes such as "module", "part", or "unit" used to denote elements are used herein only for the convenience of description of the present invention, and have no specific meaning in themselves. Thus, "module" and "component" may be used in a mixture.
As shown in fig. 1, an embodiment of the present invention provides a method for building a regional flow prediction model, which is used to build a prediction model with an input/output relationship similar to an RAS method (RAS method, also called a real-time correction method, or a double-scale method), that is, a regional flow prediction model, and the regional flow prediction model requires less historical data, and has a simple calculation process and a wide application range.
The method includes but is not limited to the following steps:
and S1, forming a benchmark annual input-output table consisting of N areas based on a closed economic system comprising the N areas.
The input-output table of the reference year is N +1 rows and N +1 columns, the value of the N +1 row is the sum of all the columns, and the sum of the jth column represents the total output flow of the area j in the reference year; the value of the N +1 th column is the sum of each row, the sum of the ith row represents the total input flow of the area i in the reference year, and N is a positive integer.
In particular applications, the region in the benchmark year input-output table may be a geographic region, such as A, B, C, D in a country, province, city, etc.; the method can also be used in industries such as building materials, furniture, real estate, decoration design and the like which need statistical analysis.
As shown in fig. 2, the standard annual input-output table formed in step S1 is a table having N +1 rows and N +1 columns, and the N +1 th row are denoted by SUM. In the embodiment of the invention, if any column in any row is used to represent the area i, then N-1 areas j are represented, and any one of the N-1 areas except the area i in the row or the column is called the area j.
In practical applications, when the benchmark year input-output table is used, data in each column of each row, namely historical traffic interaction amount among various areas, needs to be filled. However, in the embodiment of the present invention, the historical flow interaction amount of the region i to be input into N-1 regions j is used, and the total output flow of the region j in the reference year and the total input flow of the region i in the reference year are described in detail in the following step S2.
S2, acquiring historical flow interaction quantity between the area i and the N-1 areas j in the reference year input-output table;
the historical flow interaction quantity between the area i and the N-1 areas j comprises input flow of the area i to the N-1 areas j and output flow of the N-1 areas j to the area i, wherein i and j are positive integers smaller than or equal to N.
S3, filling the input flow of the region i to the N-1 regions j into the ith row of the input-output table of the reference year; the output flow of N-1 zones j to zone i is filled into the ith column of the benchmark annual input-output table.
In step S2, if the area in the benchmark annual input-output table is a geographic area, the historical traffic interaction volume may be the freight volume, the population movement volume, the economic movement volume, etc. of the area i and the N-1 areas j, and if the area in the benchmark annual input-output table is an industry, the historical traffic interaction volume may be the freight volume, the population movement volume, the economic movement volume, etc. of the industry i input to the industry j.
In the embodiment of the present invention, fig. 3 further exemplarily shows the benchmark annual input-output table obtained through the above steps S2 and S3, fig. 3 exemplarily takes i and j as 1 for explanation, when i and j are 1, data of historical traffic interaction volume needs to be filled in the first row and the first column of the benchmark annual input-output table, according to steps S2 and S3, in fig. 3, the historical traffic interaction volume of the area i and the N-1 areas j includes the input traffic x of the area i input into the N-1 areas j11、x12...x1NI.e. first row, N-1 production flows x of zone j to zone i21、x31..xN1I.e. the first column.
It can be appreciated that if i is 2 and j is 2, then the historical traffic interaction between the regions will be filled in the second row and the second column of the input-output table, and x12 and x21Without the need for padding.
If the number of the areas N is set to 5 and i and j are both set to 1, the historical traffic interaction quantity of the areas i and N-1 areas j in the intermediate yield table is x11、x12、x13、x14、x15、x21、x31、x41、x51, wherein ,x11An input flow rate x representing the input of the zone 1 into the zone 112An input flow rate x representing the input of the zone 1 into the zone 213An input flow rate x representing the input of the zone 1 into the zone 314Indicates the input flow rate, x, of the region 1 to the region 415Indicates the input flow rate, x, of the region 1 to the region 521Representing the throughput, x, of zone 2 to zone 131Representing the throughput, x, of zone 3 to zone 141Indicates the throughput, x, of zone 4 to zone 151Representing the throughput of zone 5 to zone 1. The value of row 6, representing the sum of the rows in columns 1 to 5, e.g. v1Denotes the sum of column 1, line 1 to line 5, i.e. x11、x21、x31、x41、x51The sum, also the total output flow of region 1, of the values in column 6, representing the sum of the columns, e.g. u, in rows 1 to 52Denotes the sum of row 2, column 1 to column 5, i.e. x21、x22、x23、x24、x25The sum is also the total input flow for zone 2.
S4, randomly generating historical flow interaction quantities of N-1 areas j in an extreme value range constrained by the closed economic system, and complementing the benchmark annual input-output table by using the random historical flow interaction quantities to obtain a complemented benchmark annual input-output table;
the historical traffic interaction volume of the N-1 regions j comprises the historical traffic interaction volume of any two regions in the N-1 regions j.
In step S4, the extreme value range of the closed economic system constraint is determined based on the historical data, and the extreme value of the flow rate between the regions is determined.
In the embodiment of the invention, the value of the random historical traffic interaction amount is random in the extreme value range of the constraint, and the influence of the step of updating the reference annual input-output table to obtain the target annual input-output table is small enough, so that when the historical traffic interaction amount input into the N-1 area j by the area i is not changed, the historical traffic interaction amounts of the N-1 area j are changed, and the predicted change of the traffic interaction amount values of the area i and the N-1 area j in the obtained complete target annual input-output table is small enough, namely the error is small enough, and the data has reliability.
S5, calculating the given value of the N +1 th row and the given value of the N +1 th column in the input-output table of the target year according to the value of the N +1 th row and the value of the N +1 th column in the input-output table of the reference year and the preset flow rate increase rate;
in the target year input-output table, the given value of the (N + 1) th row is the sum of all columns, and the sum of the jth column represents the total output flow of the area j in the target year; the given value of the N +1 th column is the sum of the rows, and the sum of the ith row represents the total input flow of the area i in the target year.
In the above step S5, the preset flow rate increase rate is a random number between the minimum value of the flow rate increase rate and the maximum value of the flow rate increase rate.
And S6, updating the supplemented standard annual input-output table by combining the given value of the (N + 1) th row and the given value of the (N + 1) th column in the target annual input-output table by using an RAS method to obtain a complete target annual input-output table.
According to the steps S5 and S6, the embodiment of the present invention updates the complemented benchmark annual input-output table by the RAS method according to the basic data of the historical flow interaction amount between the region i and the N-1 regions j, the total output flow of the regions j in the benchmark year, the total input flow of the region i in the benchmark year, the set value of the N +1 th row and the set value of the N +1 th column in the target annual input-output table, so as to obtain the complete target annual input-output table.
Therefore, in the embodiment of the invention, the historical flow interaction quantity of any two areas in the N-1 areas j is not needed, and only the historical flow interaction quantity of the two areas needing to be researched and analyzed is needed to be obtained. Take the above practical application as an exampleThe historical flow interaction quantity of the region i and the N-1 regions j is x11、x12、x13、x14、x15、x21、x31、x41、x51Which does not relate to x23、x24、x43Data of historical traffic interaction volume of equal area.
As shown in fig. 4, the embodiment of the present invention further shows a complemented reference annual input-output table 4a obtained based on the reference annual input-output table shown in fig. 3, that is, the reference annual input-output table after writing the regional historical traffic interaction amount in the first row and the first column, the random historical traffic interaction amount D1, and a complete target annual input-output table 4b obtained by updating based on the complemented reference annual input-output table 4 a. In Table 4a, the standard year input/output after completion of the test was used
Representing the historical traffic interaction volume of zone i and N-1 zones j, using
To
Indicating the total input flow of zone 1 into N-1 zones j,
to
Total production flow rates for zone 1 and N-1 zone j are shown in Table 4b for the complete target year input
To represent
Using as the predicted value of
To
Indicates that the total input flow rate of the region 1 to the N-1 regions j in the target annual input-output table calculated in step S5,
to
The total production flow rates of the region 1 and the N-1 regions j in the target annual input-output table calculated in step S5 are shown.
In practical applications, the above-mentioned complete target year input-output table
The area flow to be predicted in the embodiment of the invention is unknown quantity. In the above complete target year input-output table
To
To
The total input flow rate of zone 1 into N-1 zones j and the total output flow rates of zone 1 and N-1 zones j, calculated according to step S5, are indicated as known amounts. In the above-mentioned reference year input-output table
The historical traffic interaction quantity in the embodiment of the invention is a known quantity.
S7, repeating S4-S6 to obtain an input and output data set generated in the process of updating the supplemented reference annual input and output table by using an RAS method to obtain a complete target annual input and output table; and constructing a multiple nonlinear regression model by fitting the input-output relation of the input-output data set, and taking the multiple nonlinear regression model as a regional flow prediction model.
Wherein the input-output data set is composed of elements in the ith row and ith column in a large number of repeated experiments.
In the step S7, a large amount of data is needed as a basis for constructing the multiple nonlinear regression model, and in the embodiment of the present invention, the steps S4 to S6 are repeated to obtain multiple sets of input/output data by using different and large amounts of random historical traffic interaction quantities, where the multiple sets of input/output data are all input/output data obtained when the reference annual input/output table after different completions is updated to the complete target annual input/output table by using the RAS method multiple times. Therefore, in the embodiment of the present invention, the process of constructing the multiple nonlinear regression model is a process of fitting the input-output relationship of the input-output data set.
According to the step S7, the multivariate nonlinear regression model constructed in the embodiment of the present invention is a regional flow prediction model, and is used for expressing the data input/output relationship of the RAS method in the regional flow prediction process, so that the data input/output relationship is more definite.
Although the multiple nonlinear regression models obtained in the steps S1 to S7 can perform more accurate regional flow prediction when the historical data is limited, as shown in fig. 5, in the embodiment of the present invention, another method for establishing a regional flow prediction model is further shown, the multiple nonlinear regression model is further optimized by introducing a model without constant terms, the steps include steps S201 to S208, and the steps S201 to S207 are the same as the steps S1 to S7, and are not repeated, and the step S208 is:
and S208, introducing a non-constant term model to fit the multiple nonlinear regression model to obtain an optimized regional flow prediction model.
In the step S208, the fitting of the multivariate nonlinear regression model with the non-constant term model is introduced in a manner that the constant term is removed from the multivariate nonlinear regression model obtained in the step S7 to obtain the non-constant term model; and fitting the multiple nonlinear regression model with the constant term removed by using the input and output data set obtained in the step S7 to obtain a new multiple nonlinear regression model, i.e., an optimized regional flow prediction model.
In a specific application, the independent variable related to the infinite term model, i.e., the model, is 0, and the dependent variable (in the embodiment of the present invention, the element in the target annual input-output table) is also generally 0, and the data input-output relationship hidden under the RAS algorithm can be represented as a functional relationship passing through the origin, so that the existence of the constant term is avoided, and the estimation of the dependent variable (in the embodiment of the present invention, the element in the target annual input-output table) by the multivariate nonlinear regression model has a large deviation, so that the data in the reference annual input-output table is predicted by the regional flow prediction model obtained in the above steps S1 to S7, and the data in the more accurate complete target annual input-output table can be obtained, and a more accurate regional flow prediction result is obtained.
To sum up, the embodiment of the present invention provides a method for building a regional flow prediction model, which randomly generates historical flow interaction quantities of N-1 regions j in an extreme value range of flow of each region in a closed economic system, and supplements a reference annual input-output table by using the random historical flow interaction quantities to obtain a supplemented reference annual input-output table, and calculates a given value of an N +1 th row and a given value of an N +1 th column in a target annual input-output table according to a preset flow increase rate according to a given value of an N +1 th row and a given value of an N +1 th column in the reference annual input-output table, so as to calculate basic data according to the historical flow interaction quantities between a region i and the N-1 regions j, total output flow of the region j in a reference year, total input flow of the region i in the reference year, a given value of an N +1 th row and a given value of the N +1 th column in the target annual input-output table, updating the supplemented standard annual input-output table by an RAS method to obtain a complete target annual input-output table; the complete target year input-output table is a prediction result based on a supplemented reference year input-output table, then a multiple nonlinear regression model is constructed based on the supplemented reference year input-output table and the complete target year input-output table, and a data input-output relation of an RAS method is expressed, so that when historical data are insufficient, a prediction result equivalent to that of a classical RAS method can be obtained; a non-constant term model is introduced for fitting a multiple nonlinear regression model, so that the finally obtained regional flow prediction model is closer to the data input and output relation of the RAS method, the regional flow prediction model is provided, the needed historical data are less, the prediction result is equivalent to that of the classical RAS method, the reliability is high, the error is small, the input and output mathematical relation is clear, the implementation is easy, and the application range is large.
As shown in fig. 6, the embodiment of the present invention further illustrates a detailed implementation process of constructing the multiple nonlinear regression model in step S4, which includes the following steps:
s41, calculating a first ratio of total input flow of the area i in the target year to the total input flow in the target year and a second ratio of total output flow of the area j in the target year to the total output flow in the target year based on the complemented reference year input-output table and the complete target year input-output table;
in one embodiment, the first ratio and the second ratio are calculated from the total input flow rate for zone i and the total output flow rate for zone j in the reference annual input-output table and the full target annual input-output table by the formula:
wherein ,αiIs a first ratio, βiIn order to be the second ratio, the first ratio,
the total input flow of the area i in the input-output table of the reference year,
for the total throughput of region j in the input-output table for the reference year,
the total input flow of the area i in the target annual input-output table and the complete target annual input-output table,
the total production flow of the area j in the target annual input-output table and the complete target annual input-output table.
And S42, constraining the maximum value and the minimum value of the flow rate increase through the first ratio and the second ratio, and calculating the flow rate increase ratio based on the maximum value and the minimum value.
In a specific application, the flow growth rate can affect the errors of the multiple nonlinear regression model and the regional flow prediction model. Generally, as the range of growth rate of each region in the input-output table becomes larger, the prediction performance of the model gradually decreases.
Therefore, in the above step S42, the maximum value and the minimum value of the flow rate increase rate are constrained by the first ratio and the second ratio, and the flow rate increase rate ratio based on the maximum value and the minimum value is calculated, and the model applicability is actually calculated.
In one embodiment, in step S42, the maximum and minimum values of the flow rate increase rate are calculated as:
tmin＝min(αi,βj)，tmax＝max(αi,βj), wherein ,αiIs a first ratio, βiIs the second ratio.
And based on the data volume flow rate increase ratio of the maximum value and the minimum value, the formula is as follows:
γ＝tmax/tmin, wherein ,tminIs the minimum value of the rate of increase of the flow, and tmaxIs the maximum value of the rate of increase of the flow.
S43, the multivariate nonlinear regression model takes the historical flow interaction quantity of the region i and the N-1 regions j, the maximum value and the minimum value of the flow growth rate, the flow growth rate ratio, the first ratio and the second ratio as input quantities, and takes the flow interaction quantity of the region i and the N-1 regions j in the complete target year input-output table as output quantities corresponding to the input quantities.
And updating the supplemented reference annual input-output table by using an RAS method to obtain input-output data generated in the process of obtaining a complete target annual input-output table.
Based on the steps S41-S43 and the FIG. 4, the input amount includes the input amount in the benchmark annual input-output table
αi,βj，μ,σ,tmin,tmax,γ, wherein ,
comprises that
The output comprising the complete target year input-output table
Comprises that
And in the complete target year input-output table
Is the area flow to be predicted in the embodiment of the invention.
S44, repeating S4-S6 to obtain an input and output data set, and fitting the input and output relation of the input and output data set;
s45, selecting a target fitting result through goodness-of-fit inspection, and taking variables and coefficients in the target fitting result as parameters of the multiple nonlinear regression model to complete the construction of the multiple nonlinear regression model;
and in the steps S41 to S45, the multivariate nonlinear regression model is obtained by using the reference annual input-output table and the complete target annual input-output table obtained by using the RAS method as sample data of the fitting regression model.
In the embodiment of the present invention, by using a multivariate nonlinear regression model or a regional flow prediction model obtained by constant term model optimization as described below, when a reference annual input-output table is known, in detail, historical flow interaction amounts of a region i and N-1 regions j in a region i and a region j of the reference annual input-output table are known, regional flow to be predicted can be predicted. In a particular application, expressed as a predicted unknown complete target year input-output table,
the value of (c).
In step S45, the goodness-of-fit test refers to using the goodness-of-fit parameter to evaluate the fitness of the regression equation as a whole, in this embodiment of the present invention, the model with the largest goodness-of-fit parameter is selected as the target fitting result, and the variable and coefficient of the target fitting result are used as the parameters of the multiple nonlinear regression model, and the multiple nonlinear regression model constructed at this time is close to the data input/output relationship of the RAS method.
In one embodiment, the above-mentioned multivariate nonlinear regression model obtained in step S45 has the formula:
wherein ,
for the regional flow to be predicted,
historical traffic interaction quantity alpha of the region i and the N-1 regions j in the input-output table for the reference yeariIs a first ratio, βiIs a second ratio, tminAs a minimum value of the rate of increase of the flow, tmaxThe maximum value of the flow rate increase, and γ is a flow rate increase ratio based on the maximum value of the flow rate increase and the minimum value of the flow rate increase.
In the embodiment of the invention, a multivariate nonlinear regression model is optimized through a non-constant term model, and the obtained regional flow prediction model has the formula as follows:
wherein ,
for the regional flow to be predicted,
historical traffic interaction quantity alpha of the region i and the N-1 regions j in the input-output table for the reference yeariIs a first ratio, βiIs a second ratio, tminAs a minimum value of the rate of increase of the flow, tmaxThe maximum value of the flow rate increase, and γ is a flow rate increase ratio based on the maximum value of the flow rate increase and the minimum value of the flow rate increase.
As shown in fig. 7, an embodiment of the present invention further provides a regional flow prediction method, including the following steps:
s701, acquiring a reference year input-output table, wherein the reference year input-output table is based on a closed economic system comprising N areas and consists of N areas;
the input-output table of the reference year comprises N +1 rows and N +1 columns, the value of the N +1 row is the sum of all the columns, and the sum of the j column represents the total input flow of the area i in the reference year; the value of the N +1 th column is the sum of each row, the sum of the ith row represents the total output flow of the area j in the reference year, and N is a positive integer;
s702, acquiring historical flow interaction quantity of the area to be predicted based on the input-output table;
the historical flow interaction quantity of the area to be predicted comprises input flow of an area i to N-1 areas j and output flow of the N-1 areas j to the area i, wherein i and j are positive integers smaller than or equal to N;
s703, predicting the area flow of the area to be predicted by using the area flow prediction model according to any one of the claims 1 to 7 according to the historical flow interaction amount of the area to be predicted.
In practical application, if the regional flow prediction model shown in fig. 1 is applied to the logistics industry, the regional flow to be predicted by using the regional flow prediction model is the freight volume between the region i and the N-1 regions j.
The embodiment of the present invention further provides a terminal device, which includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where when the processor executes the computer program, each step in the method for establishing the regional flow prediction model in the foregoing embodiment or each step in the method for predicting the regional flow in the foregoing embodiment is implemented.
An embodiment of the present invention further provides a storage medium, where the storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium, and when being executed by a processor, the computer program implements each step in the method for establishing the regional flow prediction model in the foregoing embodiment, or each step in the regional flow prediction method in the foregoing embodiment.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the foregoing embodiments illustrate the present invention in detail, those of ordinary skill in the art will understand that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.
Claims (9)
1. A method for creating a regional flow prediction model, comprising:
s1, forming a benchmark annual input-output table consisting of N areas based on a closed economic system comprising the N areas; the input-output table of the reference year is N +1 rows and N +1 columns, the value of the N +1 row is the sum of all the columns, and the sum of the jth column represents the total output flow of the area j in the reference year; the value of the N +1 th column is the sum of each row, the sum of the ith row represents the total input flow of the area i in the reference year, and N is a positive integer;
s2, acquiring historical flow interaction quantity between the area i and the N-1 areas j in the reference year input-output table; historical flow interaction quantity between the area i and the N-1 areas j comprises input flow of the area i to the N-1 areas j and output flow of the N-1 areas j to the area i, wherein i and j are positive integers smaller than or equal to N;
s3, filling the input flow of the area i into the N-1 areas j into the ith row of the reference annual input-output table, and filling the output flow of the N-1 areas j to the area i into the ith column of the reference annual input-output table;
s4, randomly generating historical flow interaction quantities of N-1 areas j in an extreme value range constrained by the closed economic system, and complementing the benchmark annual input-output table by using the random historical flow interaction quantities to obtain a complemented benchmark annual input-output table; the historical traffic interaction quantity of the N-1 regions j comprises the historical traffic interaction quantity of any two regions in the N-1 regions j;
s5, obtaining a given value of the N +1 th row and a given value of the N +1 th column in the target annual input-output table according to the value of the N +1 th row and the value of the N +1 th column in the reference annual input-output table and a preset flow rate increase rate, wherein the given value of the N +1 th row in the target annual input-output table is the sum of all the columns, and the sum of the j th column represents the total output flow rate of the area j in the target year; the given value of the N +1 th column is the sum of all rows, and the sum of the ith row represents the total input flow of the area i in the target year;
s6, updating the supplemented standard annual input-output table by combining the given value of the (N + 1) th row and the given value of the (N + 1) th column in the target annual input-output table by using an RAS method to obtain a complete target annual input-output table;
s7, repeating S4-S6 to obtain an input and output data set generated in the process of updating the supplemented reference annual input and output table by using an RAS method to obtain a complete target annual input and output table; and constructing a multiple nonlinear regression model by fitting the input-output relation of the input-output data set, and taking the multiple nonlinear regression model as a regional flow prediction model.
2. The method for creating a regional flow prediction model according to claim 1, wherein after step S7, the method further comprises:
and introducing a non-constant term model to fit the multiple nonlinear regression model to obtain an optimized regional flow prediction model.
3. The method for building a regional flow prediction model according to claim 1, wherein the step S7 includes:
calculating a first ratio of total input flow rate for zone i in the target year to total input flow rate in the target year and a second ratio of total output flow rate for zone j in the target year to total output flow rate in the target year based on the complemented reference year input-output table and the complete target year input-output table;
constraining a maximum value and a minimum value of a flow rate increase by the first ratio and the second ratio, and calculating a flow rate increase ratio based on the maximum value and the minimum value;
the multivariate nonlinear regression model takes the historical flow interaction quantity of the region i and the N-1 regions j, the maximum value and the minimum value of the flow growth rate, the flow growth rate ratio, the first ratio and the second ratio as input quantities, and takes the flow interaction quantity of the region i and the N-1 regions j in the complete target year input-output table as output quantities corresponding to the input quantities;
the input quantity and the output quantity are input and output data generated in the process of updating the supplemented reference annual input-output table by using an RAS method to obtain a complete target annual input-output table;
repeating S4-S6 to obtain an input and output data set, and fitting the input and output relation of the input and output data set;
and selecting a target fitting result through goodness-of-fit inspection, and taking variables and coefficients in the target fitting result as parameters of the multivariate nonlinear regression model to complete the construction of the multivariate nonlinear regression model.
4. The method of creating a regional flow prediction model of claim 3, wherein a first ratio of total input flow for region i in the target year to total input flow in the reference year and a second ratio of total output flow for region j in the target year to total output flow in the reference year are calculated based on the complemented reference annual input-output table and the full target annual input-output table, and the formula is:
wherein ,αiIs a first ratio, βiIn order to be the second ratio, the first ratio,
the total input flow of the area i in the input-output table of the reference year,
for the total throughput of region j in the input-output table for the reference year,
the total input flow of the area i in the target annual input-output table and the complete target annual input-output table,
the total production flow of the area j in the target annual input-output table and the complete target annual input-output table.
5. The method for building a regional flow prediction model according to claim 3, wherein the construction of the multivariate nonlinear regression model is completed by selecting a target fitting result through goodness-of-fit test and using variables and coefficients in the target fitting result as parameters of the multivariate nonlinear regression model, and the formula is as follows:
wherein ,
for the regional flow to be predicted,
historical traffic interaction quantity alpha of the region i and the N-1 regions j in the input-output table for the reference yeariIs a first ratio, βiIs a second ratio, tminAs a minimum value of the rate of increase of the flow, tmaxThe maximum value of the flow rate increase, and γ is a flow rate increase ratio based on the maximum value of the flow rate increase and the minimum value of the flow rate increase.
6. The method for building a regional flow prediction model according to claim 5, wherein a non-constant term model is introduced to fit the multiple nonlinear regression model to obtain an optimized regional flow prediction model, and the formula is as follows:
wherein ,
for the regional flow to be predicted,
historical traffic interaction quantity alpha of the region i and the N-1 regions j in the input-output table for the reference yeariIs a first ratio, βiIs a second ratio, tminAs a minimum value of the rate of increase of the flow, tmaxThe maximum value of the flow rate increase, and γ is a flow rate increase ratio based on the maximum value of the flow rate increase and the minimum value of the flow rate increase.
7. A regional flow prediction method is characterized by comprising the following steps:
acquiring a benchmark annual input-output table, wherein the benchmark annual input-output table is based on a closed economic system comprising N areas and consists of N areas; the input-output table of the reference year is N +1 rows and N +1 columns, the value of the N +1 row is the sum of all the columns, and the sum of the jth column represents the total output flow of the area j in the reference year; the value of the N +1 th column is the sum of each row, the sum of the ith row represents the total input flow of the area i in the reference year, and N is a positive integer;
acquiring historical traffic interaction quantity of an area to be predicted based on the benchmark year input-output table, wherein the historical traffic interaction quantity of the area to be predicted comprises historical traffic interaction quantity between an area i and N-1 areas j; historical flow interaction quantity between the area i and the N-1 areas j comprises input flow of the area i to the N-1 areas j and output flow of the N-1 areas j to the area i, wherein i and j are positive integers smaller than or equal to N;
predicting the regional flow of the region to be predicted by using the regional flow prediction model according to any one of the claims 1 to 6 according to the historical flow interaction amount of the region to be predicted.
8. A terminal device comprising a memory, a processor and a computer program stored on the memory and operable on the processor, the processor implementing the steps of the method of creating a regional flow prediction model according to any one of claims 1 to 6 or the steps of the regional flow prediction method according to claim 7 when executing the computer program.
9. A storage medium being a computer readable storage medium having a computer program stored thereon, wherein the computer program, when being executed by a processor, implements the steps of the method of creating a regional flow prediction model according to any one of claims 1 to 6 or the steps of the regional flow prediction method according to claim 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010947047.4A CN112036662B (en) | 2020-09-10 | 2020-09-10 | Method for establishing regional flow prediction model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010947047.4A CN112036662B (en) | 2020-09-10 | 2020-09-10 | Method for establishing regional flow prediction model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112036662A true CN112036662A (en) | 2020-12-04 |
| CN112036662B CN112036662B (en) | 2023-06-20 |
Family
ID=73584745
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010947047.4A Active CN112036662B (en) | 2020-09-10 | 2020-09-10 | Method for establishing regional flow prediction model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112036662B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101286933A (en) * | 2008-03-05 | 2008-10-15 | 中科院嘉兴中心微系统所分中心 | Self-adapting cluster regulating method for wireless sensor network based on flow rate |
| CN108259462A (en) * | 2017-11-29 | 2018-07-06 | 国网吉林省电力有限公司信息通信公司 | Big data Safety Analysis System based on mass network monitoring data |
| CN108491969A (en) * | 2018-03-16 | 2018-09-04 | 国家电网公司 | Spatial Load Forecasting model building method based on big data |
-
2020
- 2020-09-10 CN CN202010947047.4A patent/CN112036662B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101286933A (en) * | 2008-03-05 | 2008-10-15 | 中科院嘉兴中心微系统所分中心 | Self-adapting cluster regulating method for wireless sensor network based on flow rate |
| CN108259462A (en) * | 2017-11-29 | 2018-07-06 | 国网吉林省电力有限公司信息通信公司 | Big data Safety Analysis System based on mass network monitoring data |
| CN108491969A (en) * | 2018-03-16 | 2018-09-04 | 国家电网公司 | Spatial Load Forecasting model building method based on big data |
Non-Patent Citations (1)
| Title |
|---|
| 王新语;谢扬;: "基于模糊约束的空中交通流量预测模型研究", 交通运输研究, no. 05, pages 68 - 73 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112036662B (en) | 2023-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Keilman et al. | Why population forecasts should be probabilistic-illustrated by the case of Norway | |
| Aastveit et al. | Nowcasting GDP in real time: A density combination approach | |
| Ferrara et al. | Macroeconomic forecasting during the Great Recession: The return of non-linearity? | |
| Martínez-Miranda et al. | Double chain ladder and Bornhuetter-Ferguson | |
| Wan et al. | Frequentist model averaging for multinomial and ordered logit models | |
| Pešta et al. | Conditional least squares and copulae in claims reserving for a single line of business | |
| Tyralis et al. | On the prediction of persistent processes using the output of deterministic models | |
| CN110633859B (en) | Hydrologic sequence prediction method integrated by two-stage decomposition | |
| Lambert et al. | Global sensitivity analysis using sparse high dimensional model representations generated by the group method of data handling | |
| Martínez Miranda et al. | Double chain ladder, claims development inflation and zero-claims | |
| Rose | Numerical methods for solving optimal control problems | |
| Yukalov et al. | Self-similar structures and fractal transforms in approximation theory | |
| Hubert et al. | Detecting influential data points for the Hill estimator in Pareto-type distributions | |
| Zhang et al. | A hybrid sequential sampling strategy for sparse polynomial chaos expansion based on compressive sampling and Bayesian experimental design | |
| Mitchell et al. | Nowcasting Euro area GDP growth using Bayesian quantile regression | |
| de Graaf et al. | Efficient exposure computation by risk factor decomposition | |
| CN112036662A (en) | Method for establishing regional flow prediction model and regional flow prediction method | |
| Maleta | Methods for Determining the Impact of the Temporal Trend in the Valuation of Land Property | |
| Korn | A Simple Method for Modeling Changes over Time | |
| Knight et al. | A knowledge-based system to represent spatial reasoning for fire modelling | |
| US20230289561A1 (en) | Computer-implemented method for the generation of a mathematical model with reduced computational complexity | |
| Szekeres | Checking the Evidence for Declining Discount Rates | |
| Darus | Review on fuzzy difference equation | |
| Jallbjørn | Multi-population mortality models and scenario-based projections | |
| Kim | Event tree based sampling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |