CN110135913A - Training method, shop site selecting method and the device of shop site selection model - Google Patents

Training method, shop site selecting method and the device of shop site selection model Download PDF

Info

Publication number
CN110135913A
CN110135913A CN201910416777.9A CN201910416777A CN110135913A CN 110135913 A CN110135913 A CN 110135913A CN 201910416777 A CN201910416777 A CN 201910416777A CN 110135913 A CN110135913 A CN 110135913A
Authority
CN
China
Prior art keywords
factor
shop
equation
influence
probability
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
Application number
CN201910416777.9A
Other languages
Chinese (zh)
Inventor
沈樱
张岩
李振军
林乾辰
闫嘉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Smart Footprint Data Technology Co Ltd
Original Assignee
Smart Footprint Data Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Smart Footprint Data Technology Co Ltd filed Critical Smart Footprint Data Technology Co Ltd
Priority to CN201910416777.9A priority Critical patent/CN110135913A/en
Publication of CN110135913A publication Critical patent/CN110135913A/en
Pending legal-status Critical Current

Links

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
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0204—Market segmentation
    • G06Q30/0205—Location or geographical consideration

Landscapes

  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Game Theory and Decision Science (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The embodiment of the present application provides training method, shop site selecting method and the device of a kind of shop site selection model, the training method includes: to construct multiple multiple-factor equations for being used to predict shop efficiency according at least two first influence factors of store location, and the shop efficiency reflects the business revenue ability in corresponding shop；At least one the multiple-factor equation for meeting significance test requirement is filtered out from multiple multiple-factor equations；Determine the degree of fitting of prediction the shop efficiency and practical shop efficiency of at least one multiple-factor equation；The highest multiple-factor equation of degree of fitting is determined as to the equation of the shop site selection model.The embodiment of the present application can train the shop site selection model for shop addressing through the above way, then carry out shop addressing, the problem that can preferably avoid conventional method subjectivity strong using the model trained.

Description

Training method, shop site selecting method and the device of shop site selection model
Technical field
This application involves data screening technical fields, training method, shop in particular to a kind of shop site selection model Spread site selecting method and device.
Background technique
In the prior art when carrying out shop addressing, often alternate location is assessed based on field research. For example, understanding bunk self-condition such as property type by experienced expert, StoreFront area, floor, shop periphery flow of the people, handing over Logical Discussing Convenience etc..Then expert scores to above-mentioned bunk self-condition, and it is higher that scoring is picked out from several alternative shops 's.
Large enterprise is needed to plan and manages layout as needed to tap new markets, conventional method is time-consuming and laborious, due to passing System method relies primarily on subjective judgement, and generation erroneous judgement probability is larger, may bring massive losses to enterprise.Do not have in the prior art There is compared with good method the addressing for carrying out shop.
Summary of the invention
In view of this, the embodiment of the present application provide the training method of shop site selection model a kind of, shop site selecting method and Device, to improve in the prior art without preferable method come the problem of carrying out shop addressing.
In a first aspect, the embodiment of the present application provides a kind of training method of shop site selection model, which comprises root It is constructed according at least two first influence factors of store location multiple for predicting the multiple-factor equation of shop efficiency, the shop Spread the business revenue ability that efficiency reflects corresponding shop；It is filtered out from multiple multiple-factor equations and meets significance test requirement At least one multiple-factor equation；Determine the prediction shop efficiency of at least one multiple-factor equation and intending for practical shop efficiency It is right；The highest multiple-factor equation of degree of fitting is determined as to the equation of the shop site selection model.
In the above-described embodiment, multiple multiple-factor equations are first constructed, then filter out symbol from multiple multiple-factor equations Close the multiple-factor equation that significance test requires.Then to the multiple-factor equation calculation prediction shop efficiency filtered out and practical shop The degree of fitting of efficiency is spread, and the high multiple-factor equation of degree of fitting is determined as the equation of shop site selection model, through the above way The shop site selection model for shop addressing can be trained, then carries out shop addressing using the model trained, it can The problem for preferably avoiding conventional method subjectivity strong.
In a possible design, each multiple-factor equation in multiple multiple-factor equations have it is corresponding at least Two the first influence factors；It is described that at least one for meeting significance test requirement is filtered out from multiple multiple-factor equations Multiple-factor equation, comprising: for each multiple-factor equation in multiple multiple-factor equations, make the multiple-factor equation fitted Assume for genuine first；For each of at least two first influence factor corresponding with each multiple-factor equation First influence factor calculates separately described first and assumes the first probability for mistake occur；If multiple-factor equation corresponding at least two First probability of the first influence factor of each of a first influence factor is respectively less than the first preset threshold, then the multiple-factor side Cheng Fuhe significance test requirement.
In the above-described embodiment, when being screened from multiple multiple-factor equations, each multiple-factor equation is made Multiple-factor equation is genuine first it is assumed that then counting respectively to the first influence factor each of corresponding to each multiple-factor equation out It calculates first and assumes the first probability for mistake occur, then screened according to the relative size of the first probability and the first preset threshold more Factor Equations.The higher multiple-factor equation of accuracy can be filtered out by way of significance test, reduce what mistake occurred Possibility.
In a possible design, after first probability for calculating separately the first hypothesis appearance mistake, The method also includes: if there are first probability to be greater than in corresponding at least two first influence factor of multiple-factor equation Or the first influence factor equal to first preset threshold, then the multiple-factor equation does not meet significance test requirement.
It in the above-described embodiment, can be by the multiple-factor equation if multiple-factor equation does not meet significance test requirement It screens out, to improve the accuracy of multiple-factor equation.
In a possible design, it is described constructed according at least two first influence factor it is multiple for predicting The multiple-factor equation of shop efficiency, comprising: calculate related coefficient between any two at least two first influence factor；Point The related coefficient for not making the first influence factor of every two at least two first influence factor is genuine multiple second false And if calculating separately the second hypothesis of each of the multiple second hypothesis and the second probability of mistake occur；If it exists less than Second probability of two preset thresholds then filters out second probability from least two first influence factor and is greater than or waits The multiple-factor equation is constructed in the first influence factor of second preset threshold.
In the above-described embodiment, then illustrate that second probability is corresponding less than the second probability of the second preset threshold if it exists Related coefficient corresponding to two the first influence factors there is correlation, therefore when construct multiple-factor equation, have related Property two influence factors be not suitable for constructing multiple-factor equation jointly, with correlation two influence factors can respectively with its He constructs multiple-factor equation at influence factor, so as to further increase the accuracy of multiple-factor equation.
In a possible design, assume occur in described each of the multiple second hypothesis second that calculates separately After second probability of mistake, the method also includes: if second probability is all larger than or is equal to second preset threshold, The first influence factor is randomly choosed from least two first influence factor constructs the multiple-factor equation.
In the above-described embodiment, if the second probability is all larger than or is equal to the second preset threshold, illustrate at least two the Therefore one influence factor does not have correlation between any two will not reduce the accurate of multiple-factor equation random combine Property.
It is described to construct multiple use according at least two first influence factors of store location in a possible design Before the multiple-factor equation of prediction shop efficiency, the method also includes: it filters out from multiple store location influence factors At least two first influence factors.
In the above-described embodiment, primary screening first can be carried out from multiple store location influence factors, filtered out pair Then the factor that shop efficiency has a significant effect selects influence factor to construct multiple-factor side from the factor having a significant effect again The efficiency of model training can be improved in journey.
It is described that at least two first influences are filtered out from multiple store location influence factors in a possible design Factor, comprising: calculate between each store location influence factor and the shop efficiency in multiple store location influence factors Related coefficient；Make in multiple store location influence factors respectively each store location influence factor and the shop efficiency it Between related coefficient be genuine multiple thirds it is assumed that and calculate separately the multiple third assume in each third assume occur The third probability of mistake；If the store location influence factor that third probability is less than third predetermined threshold value exists, from the multiple shop Bunk sets and filters out the store location influence factor that third probability is less than third predetermined threshold value in influence factor, and described first influences Factor is the store location influence factor that third probability is less than third predetermined threshold value.
In the above-described embodiment, if third probability is less than third predetermined threshold value, show the corresponding shop of third probability Influence factor is set in bunk has more apparent influence to shop efficiency, therefore, can be according to third probability and third predetermined threshold value Size filter out the first influence factor.
Second aspect, the embodiment of the present application provide a kind of shop site selecting method, which comprises according to above-mentioned instruction Practice the multiple-factor regression model that method training is completed, obtains and select shadow as multiple ends of the independent variable of the multiple-factor regression model The factor of sound；Will to addressing region division be multiple subregions, by the multiple subregion each subregion with it is the multiple It selects the corresponding truthful data of influence factor to bring the multiple-factor regression model into eventually, obtains the prediction shop of each subregion Efficiency；According to the sequence of the prediction shop efficiency from high to low, ranking is carried out to each subregion；According to the ranking And the existing store location of user, determine the addressing in new shop.
In the above-described embodiment, using model obtain prediction shop efficiency, further according to prediction shop efficiency and user Some shops can filter out the position in new shop, compared with prior art, can more efficiently and more objectively and accurately carry out The selection of shop shop location.
It is described according to the ranking and the existing store location of user in a possible design, determine new shop Addressing, comprising: according to subregion described in ranking sequential selection from high to low；Judge the subregion selected with Whether the distance between described existing store location of user is more than pre-determined distance；If so, the subregion selected is made For the addressing in the new shop；If it is not, the subregion of a ranking lower than the subregion in the ranking is obtained, it is described to compare institute The subregion for stating the low ranking of subregion is the new subregion selected, and executes the step " sub-district that judgement is selected Whether the distance between domain and the existing store location of the user are more than pre-determined distance ".
In the above-described embodiment, the sequential selection subregion according to prediction shop efficiency from high to low, and judge sub-district Whether the distance between domain and the existing store location of user are more than pre-determined distance, if if so, showing the subregion selected As shop, the business that user has shop will not influence, therefore it can be selected that if it is not, then explanation may have shop to user The business of paving impacts, therefore can skip the subregion, to the low ranking in prediction shop efficiency ratio current sub-region Subregion continues to execute Distance Judgment, until obtain ranking it is higher and will not influence user have shop business subregion.It should Mode has comprehensively considered the interests of user, so that the store location of selection benefits user to the greatest extent.
The third aspect, the embodiment of the present application provide a kind of training device of shop site selection model, the described device side of including: Journey constructs module, constructs at least two first influence factors according to store location multiple for predicting shop efficiency Multiple-factor equation, the shop efficiency reflect the business revenue ability in corresponding shop；Equation screening module, for from it is multiple it is described mostly because At least one the multiple-factor equation for meeting significance test requirement is filtered out in sub- equation；Degree of fitting determining module, for determining The degree of fitting of prediction the shop efficiency and practical shop efficiency of at least one multiple-factor equation；Equation determining module, is used for The highest multiple-factor equation of degree of fitting is determined as to the equation of the shop site selection model.
In a possible design, equation screening module includes: the first hypothesis setting submodule, for for multiple more Each multiple-factor equation in Factor Equations, making the multiple-factor equation fitted is genuine first hypothesis；First probability meter Operator module, for for each of at least two first influence factor corresponding with each multiple-factor equation One influence factor calculates separately described first and assumes the first probability for mistake occur；It upchecks submodule, for working as multiple-factor First probability of the first influence factor of each of corresponding at least two first influence factor of equation is respectively less than the first default threshold When value, determine that the multiple-factor equation meets significance test requirement.
In a possible design, described device further includes that the test fails submodule, for working as multiple-factor equation pair In at least two first influence factors answered, there are the first shadows that first probability is greater than or equal to first preset threshold When the factor of sound, determine that the multiple-factor equation does not meet significance test requirement.
In a possible design, it includes: the first related coefficient submodule that equation, which constructs module, for calculate it is described extremely Related coefficient in few two the first influence factors between any two；Second probability calculation submodule, for make respectively it is described extremely The related coefficient of the first influence factor of every two is genuine multiple second it is assumed that and calculating separately in few two the first influence factors Each of the multiple second hypothesis second assumes the second probability for mistake occur；Equation constructs submodule, exists for working as Less than the second preset threshold the second probability when, it is big from least two first influence factor to filter out second probability The multiple-factor equation is constructed in or equal to the first influence factor of second preset threshold.
In a possible design, equation constructs module further include: random selection submodule, if general for described second Rate be all larger than or be equal to second preset threshold, from least two first influence factor randomly choose first influence because Element constructs the multiple-factor equation.
In a possible design, described device further includes influence factor screening module, is used for from multiple store locations At least two first influence factors are filtered out in influence factor.
In a possible design, the influence factor screening module includes: the second related coefficient submodule, based on Calculate the related coefficient between each store location influence factor and the shop efficiency in multiple store location influence factors；The Three probability calculation submodules, for make in multiple store location influence factors respectively each store location influence factor with it is described Related coefficient between the efficiency of shop is genuine multiple thirds it is assumed that and calculating separately the multiple third and each of assuming the Three assume the third probability for mistake occur；Influence factor screens submodule, for being less than third predetermined threshold value when third probability In the presence of store location influence factor, it is default less than third that third probability is filtered out from the multiple store location influence factor The store location influence factor of threshold value.
Fourth aspect, the embodiment of the present application provide a kind of shop addressing device, and described device includes: to select influence factor eventually Module is obtained, is completed for the training of the method according to any optional implementation of first aspect or first aspect more Factorial regression model obtains and selects influence factor as multiple ends of the independent variable of the multiple-factor regression model；Subregion efficiency Obtain module, for will to addressing region division be multiple subregions, by the multiple subregion each subregion and institute Stating multiple ends selects the corresponding truthful data of influence factor to bring the multiple-factor regression model into, obtains the pre- of each subregion Survey shop efficiency；Subregion ranking module, for the sequence according to the prediction shop efficiency from high to low, to each son Region carries out ranking；New shop addressing module, for determining new shop according to the ranking and the existing store location of user Addressing.
In a possible design, new shop addressing module includes region addressing submodule, for according to the ranking Subregion described in sequential selection from high to low；Distance Judgment submodule, the subregion for judging to select with it is described Whether the distance between existing store location of user is more than pre-determined distance；First result submodule, the institute for will select State addressing of the subregion as the new shop；Second result submodule is lower than the subregion in the ranking for obtaining The subregion of one ranking, the subregion of a ranking lower than the subregion are the new subregion selected, and are executed Step " the distance between described subregion and the existing store location of the user that judgement is selected whether be more than it is default away from From ".
5th aspect, the application provide a kind of electronic equipment, comprising: processor, memory and bus, the memory are deposited The executable machine readable instructions of the processor are contained, when electronic equipment operation, the processor and the storage By bus communication between device, execution first aspect or first aspect when the machine readable instructions are executed by the processor Method described in any optional implementation.
6th aspect, the application provide a kind of computer readable storage medium, store on the computer readable storage medium There is computer program, any optional realization of first aspect or first aspect is executed when which is run by processor Method described in mode.
7th aspect, the application provide a kind of computer program product, and the computer program product is transported on computers When row, so that computer executes the method in any possible implementation of first aspect or first aspect.
Above objects, features, and advantages to enable the embodiment of the present application to be realized are clearer and more comprehensible, be cited below particularly compared with Good embodiment, and cooperate appended attached drawing, it is described in detail below.
Detailed description of the invention
Illustrate the technical solutions in the embodiments of the present application or in the prior art in order to clearer, to embodiment or will show below There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this Some embodiments of application for those of ordinary skill in the art without creative efforts, can be with It obtains other drawings based on these drawings.
Fig. 1 is the flow chart of the training method of shop site selection model provided by the embodiments of the present application；
Fig. 2 is the specific flow diagram of step S120 in Fig. 1；
Fig. 3 is the flow diagram of shop site selecting method provided by the embodiments of the present application；
Fig. 4 is the schematic block diagram of the training device of shop site selection model provided by the embodiments of the present application；
Fig. 5 is the structural block diagram of electronic equipment in the embodiment of the present application；
Fig. 6 is the application scenarios schematic diagram of shop site selection model provided by the embodiments of the present application.
Specific embodiment
Before introducing the specific embodiment of the application, first the application scenarios of the application are simply introduced.
Traditional shop addressing is based primarily upon field research and combines to accumulate experience and assesses the alternate location in shop. This method is screened to alternative store location, usually by expert from the self-condition of alternative store location to shop Position is scored, and it is higher that scoring is then picked out from several alternative store locations.Alternative store location itself Condition includes the side such as property type, alternative shop area, floor, periphery flow of the people, periphery business environment and traffic convenience degree Face.This method it is subjective.
It with the development of technology and the raising to addressing importance attention rate, can be with combining geographic information system (Geographic Information System, abbreviation GIS) quickly obtains the surrounding geographical information of alternative store location, The size of population, count amount and the distance on such as periphery compete shop quantity and distance.However personnel is needed to have GIS The surrounding geographical information of certain technical foundation ability alternative store location of quick obtaining, thus it is how more efficiently and more intelligent The problem of carry out shop addressing of change is those skilled in the art's urgent need to resolve.
Drawbacks described above existing in the prior art is applicant in the structure obtained after practicing and carefully studying, Therefore, the discovery procedure of the above problem and the solution that hereinafter the embodiment of the present application is proposed regarding to the issue above, all It should be the contribution that applicant makes the application during the application.
Referring to Figure 1, Fig. 1 shows the process signal of the training method of shop site selection model provided by the embodiments of the present application Figure, it should be appreciated that method shown in FIG. 1 can be executed by electronic equipment, which can be with electricity shown in fig. 5 hereinafter Sub- equipment is corresponding, which can be the various equipment for being able to carry out this method, for example, such as personal computer, server Or network equipment etc., the embodiment of the present application is not limited to this, and is specifically comprised the following steps:
Step S110 is constructed multiple for predicting shop efficiency according at least two first influence factors of store location Multiple-factor equation.
First influence factor is to generate the factor significantly affected to shop efficiency, and the first influence factor can be inlet flow rate Quantity, resident population's quantity, bus station quantity, subway station quantity and commercial circle quantity etc..Shop efficiency reflects corresponding shop Business revenue ability, such as the turnover in shop.The independent variable of multiple-factor equation is the first influence factor, and dependent variable is shop efficiency, Multiple-factor equation can calculate shop efficiency according to the specific value of the first influence factor.
When using at least two first influence factors building multiple-factor equation, it can be achieved by the steps of:
Step S111 calculates related coefficient between any two at least two first influence factor.
Related coefficient is Pearson correlation coefficients, and Pearson correlation coefficients reflect the linearly related journey between two variables Degree, for value between [- 1,1], linearly related degree is stronger, and Pearson correlation coefficients more level off to -1 or 1.Pearson correlation Coefficient is obtained by the covariance between two variables divided by standard deviation.
For ease of description, at least two first influence factors might as well be set as five the first influence factors: first influence because Plain A, the first influence factor B, the first influence factor C, the first influence factor D, the first influence factor E, hereinafter referred to as A, B, C, D,E.Related coefficient between any two in the first influence factor is calculated, that is, calculates separately the related coefficient (quantity of A Yu B, C, D, E It is 4), the related coefficient (quantity is 3) of B and C, D, E, the related coefficient (quantity is 2) of C and D, E, the phase relation of D and E Number (quantity is 1).
Optionally, calculate two the first influence factors related coefficient when, can use two the first influence factors it Between covariance obtained divided by standard deviation.
Step S112 makes the phase relation of the first influence factor of every two at least two first influence factor respectively Number is for genuine multiple second it is assumed that and calculating separately the multiple second and each of assuming that second assumes occur the second of mistake Probability.
It is genuine second that the related coefficient, which can be made, for each related coefficient in whole related coefficients for finding out Assuming that.It connects above example to continue to illustrate, amounts in step S111 and calculated 10 (4+3+2+1) a related coefficients, then may be used Accordingly to make 10 second hypothesis.
For ease of description, it might as well be illustrated by taking the related coefficient of A and B as an example below: for the related coefficient of A and B, Second made is assumed are as follows: the related coefficient of A and B is true.Then the related coefficient for calculating A and B is that true this second is assumed The probability of existing mistake, the probability are the second above-mentioned probability.Above-mentioned 10 second vacations can be calculated separately in the same way If each second assumes the second probability for mistake occur in, it can obtains 10 the second probability.
Optionally, it first makes second and wrong probability occurs it is assumed that then calculating the second hypothesis again, can be examined by conspicuousness The mode tested carries out, such as can carry out in such a way that P value is examined, and the specific calculating process of significance test should not be understood To be the limitation to the application.
Step S113, if it exists less than the second probability of the second preset threshold, then from it is described at least two first influence because Second probability is filtered out in element constructs the multiple-factor more than or equal to the first influence factor of second preset threshold Equation.
Second preset threshold is the probability for examining " null hypothesis " to set up, usually 5%.
It connects above example to continue to illustrate, after calculating 10 the second probability by step S112, by 10 second The second probability of each of probability is compared with the second preset threshold.If in 10 the second probability, existing pre- less than second If the second probability of threshold value, then show that this corresponding less than the second probability of the second preset threshold second assumes reflected two A first influence factor is linearly related.If two linearly related the first influence factors construct multiple-factor equation, meeting simultaneously Reduce the accuracy of the multiple-factor equation constructed.Therefore, it when selecting the first influence factor building multiple-factor equation, should be avoided The first linearly related influence factor is selected simultaneously.
For example, in first influence factor A, B, C, D, E, if A and B is linearly related, A and B should be avoided and selected simultaneously, A Multiple-factor equation can be constructed together at least one of C, D, E, B can also be constructed together at least one of C, D, E Multiple-factor equation can also construct multiple-factor equation by any two in C, D, E.
Step S114, if second probability is all larger than or is equal to second preset threshold, from described at least two the The first influence factor is randomly choosed in one influence factor constructs the multiple-factor equation.
If the second probability of each of 10 second probability is all larger than or is equal to the second preset threshold, show the first influence Factor A, B, C, D, E are not linear correlation, therefore, any two, any three can be randomly choosed from A, B, C, D, E A, any four or all five construct multiple-factor equation.
It might as well be illustrated for constructing multiple-factor equation according to A, C, E to how constructing multiple-factor equation below:
A, the multiple-factor equation of C, E building is as follows:
Yi=β 0+ β 1 (x1)i+β2(x2)i+β3(x3)i
Wherein, YiFor the dependent variable of multiple-factor equation, such as shop efficiency；X1, x2, x3 are respectively above-mentioned A, C, E tri- First influence factor；β 0 is constant term；β 1, β 2, β 3 are parameter to be asked；I is the quantity of sample data.
Since there are three parameter beta 1 to be asked, β 2, β 3 in the multiple-factor equation, at least need three groups of sample datas Substitute into multiple-factor equation Yi=β 0+ β 1 (x1)i+β2(x2)i+β3(x3)i, so as to find out β 1, β 2, β 3.
For example, can be by (x1)1、(x2)1、(x3)1With Y1；(x1)2、(x2)2、(x3)2With Y2；(x1)3、(x2)3、(x3)3 With Y3Y is substituted into respectivelyi=β 0+ β 1 (x1)i+β2(x2)i+β3(x3)i, solve parameter beta 1, β 2, β 3.
The quantity of sample data is related with the building quantity of the first influence factor of multiple-factor equation, the quantity of sample data Should be greater than or equal to building multiple-factor equation the first influence factor quantity.Usually, the quantity of sample data is more, structure The model built out it is accurately higher.Sample data can be provided by the user of shop addressing to be carried out, and user can provide more The shop efficiency in the same type shop that group is managed at present and the shop addressing influence factor that shop efficiency may be influenced, for example, Consumption data, same type shop under consumption data, line in the population distribution of same type shop region, the line in same type shop The commercial circle quantity of region, count amount of same type shop region etc..
For example, referring to following table:
When shop for selling mobile phone carries out addressing, user can provide the shop effect in the shop of same sale mobile phone Energy, the resident population of shop region, the working population of shop region, the visiting population of shop region, shop Mobile phone purchase volume, shop institute in the average terminal price of the held terminal of the people of region, the line of the people of shop region Bus station quantity and road quantity in region.Using other first influence factors building multiple-factor equation process with it is upper Text is identical, does not just repeat them here herein.
Step S120, filtered out from multiple multiple-factor equations meet significance test requirement at least one mostly because Sub- equation.
Significance test (significance test) is in advance to the parameter or overall distribution form of totality (stochastic variable) One is made it is assumed that then judging that this assumes that whether rationally (alternative hypothesis), that is, judge the true of totality using sample information Whether truth condition and null hypothesis have significant difference).
Fig. 2 is referred to, Fig. 2 shows the specific steps flow diagrams of step S120, specifically comprise the following steps:
Step S121 makes the multiple-factor side fitted for each multiple-factor equation in multiple multiple-factor equations Journey is genuine first hypothesis.
Multiple multiple-factor equations, each multiple-factor equation can be constructed using the first influence factor building multiple-factor equation Selected first influence factor is different.For each multiple-factor equation in multiple multiple-factor equations, can make respectively The multiple-factor equation fitted is genuine first hypothesis.
Step S122, for every at least two first influence factor corresponding with each multiple-factor equation A first influence factor calculates separately described first and assumes the first probability for mistake occur.
It connects and continues to illustrate for the multiple-factor equation of literary A, C, E building:
For tri- the first influence factors of A, C, E in the multiple-factor equation, calculating separately the multiple-factor equation fitted is There is the probability of mistake in this true hypothesis, which is the first probability, and for the multiple-factor equation, the first probability has three It is a.
Step S123, judges whether the first probability of corresponding each first influence factor of multiple-factor equation is respectively less than first Preset threshold, if so, executing step S125；If it is not, executing step S124.
First preset threshold is also the probability for examining " null hypothesis " to set up, and the specific value of the first preset threshold can be with Second preset threshold is identical, can also be different from the second preset threshold.
To the multiple-factor equation of A, C, E building, due to being to calculate separately the first hypothesis to tri- the first influence factors of A, C, E There are the probability of mistake, therefore the available and multiple-factor dependence among equations three the first probability.
Three the first probability are compared with the first preset threshold respectively, are preset if three the first probability are respectively less than first Threshold value then shows that the multiple-factor equation meets significance test requirement, executes step S125；If existing in three the first probability big In or equal to the first preset threshold the first probability, then show that the multiple-factor equation is unsatisfactory for significance test requirement.
Because multiple first influence factors together when, have interaction between multiple first influence factors, about A, the first influence factor of each of tri- first influence factors of C, E, the meaning of the first probability be, if it is less than the first default threshold Value then indicates that first influence factor can effectively influence result in current herein multiple-factor equation, and it is pre- to be greater than or equal to first If threshold value indicates that first influence factor cannot effectively influence this multiple-factor equation, that is, remove more than or equal to the first preset threshold The first influence factor have minor impact to the degree of fitting of this multiple-factor equation.
First probability is the P value during P value is examined, and P value can reflect a certain first influence factor and have other the When one influence factor, if there is statistically significant predictive ability, that is, reflect first influence factor whether in multiple-factor side Contribution is generated to prediction in journey.In some cases, it may judge to delete the from multiple-factor equation using inapparent P value Whether one influence factor will not significantly reduce the predictive ability of model.For example, if the P value of first influence factor be greater than or Equal to the first preset threshold, it can be said that first influence factor predictive ability in the presence of other first influence factors It is weaker, the first influence factor that the P value is greater than or equal to the first preset threshold can be deleted, multiple-factor equation is rebuild.
Step S124 determines that the multiple-factor equation does not meet significance test requirement.
After determining that multiple-factor equation does not meet significance test requirement, which can be deleted.
Step S125 determines that the multiple-factor equation meets significance test requirement.
After determining that multiple-factor equation meets significance test requirement, retain the multiple-factor equation, it is subsequent in order to carry out Screening.
Step S130 determines the fitting of prediction the shop efficiency and practical shop efficiency of at least one multiple-factor equation Degree.
Degree of fitting (R-squared) inspection is tested to the model built, and the prediction result of the model is compared With the degree of agreement for actually occurring result.The common degree of fitting method of inspection includes residual sum of square inspection, Chi-square Test and line Property regression testing etc..
For meeting the multiple-factor equation of significance test requirement, the specific sample of multiple groups corresponding with the first influence factor is substituted into Notebook data calculates the prediction shop efficiency of multiple-factor equation, which is compared with practical shop efficiency, Calculate the degree of fitting of multiple-factor equation.
The highest multiple-factor equation of degree of fitting is determined as the equation of the shop site selection model by step S140.
Multiple multiple-factor equations are first constructed, then is filtered out from multiple multiple-factor equations and meets the more of significance test requirement Factor Equations.Then the degree of fitting of shop efficiency and practical shop efficiency, and handle are predicted the multiple-factor equation calculation filtered out The high multiple-factor equation of degree of fitting is determined as the equation of shop site selection model, can train one through the above way for shop Then the shop site selection model of addressing carries out shop addressing using the model trained, can preferably avoid conventional method master The strong problem of the property seen.
In a specific embodiment, after training the multiple-factor equation of shop site selection model, the shop that will train Paving site selection model is verified with control group, and the prediction shop efficiency that shop site selection model is exported is residual with practical shop efficiency Difference compares, and judges whether in the reasonable scope, if not in the reasonable scope, can analyze leads to the abnormal origin cause of formation.If mould The reason of type, then adjusts model；If the reason of shop data causes exception, then shop initial data is corrected, exceptional value is removed, Repeat all of above step step.
In a specific embodiment, before step S110, the method also includes: it is influenced from multiple store locations At least two first influence factors are filtered out in factor.
First influence factor can be screened from store location influence factor and be obtained.Store location influence factor is possible shadow The influence factor of shop efficiency is rung, therefore the purpose of this step is to screen from the influence factor that may influence shop efficiency to shop Paving efficiency generates the factor significantly affected.
Optionally, the first influence factor is screened from store location influence factor to be carried out in the following way:
It calculates between each store location influence factor and the shop efficiency in multiple store location influence factors Related coefficient.
It is made in multiple store location influence factors respectively between each store location influence factor and the shop efficiency Related coefficient be genuine multiple thirds it is assumed that and calculate separately the multiple third assume in each third assume to occur it is wrong Third probability accidentally.
If the store location influence factor that third probability is less than third predetermined threshold value exists, from the multiple store location shadow The store location influence factor that third probability is less than third predetermined threshold value is filtered out in the factor of sound, first influence factor is the Three probability are less than the store location influence factor of third predetermined threshold value.
Third predetermined threshold value is also the probability for examining " null hypothesis " to set up, and numerical value can be with the first preset threshold or the Two preset thresholds are identical, can also be different from the first preset threshold or the second preset threshold.
The quantity of multiple store location influence factors might as well be set as 10, be store location influence factor A, shop position respectively Set influence factor B, store location influence factor C, store location influence factor D, store location influence factor E, store location shadow Ring factor F, store location influence factor G, store location influence factor H, store location influence factor I, store location influence because Plain J, hereinafter referred to as A, B, C, D, E, F, G, H, I, J calculate separately shop for each of 10 store location influence factors Set the related coefficient between influence factor and shop efficiency in bunk.
Optionally, when calculating the related coefficient of store location influence factor and shop efficiency, it can use store location Covariance between influence factor and shop efficiency is obtained divided by standard deviation.
For the related coefficient of each store location influence factor and shop efficiency, making related coefficient is that genuine third is false If.There are 10 store location influence factors in the present embodiment, is then corresponding with 10 thirds and assumes.
Then it calculates separately each third in 10 thirds hypothesis and assumes that the probability for mistake occur, the probability are that third is general Rate.It is limitation to the application that the specific calculating process of significance test, which should not be construed,.
Multiple third probability are compared with third predetermined threshold value respectively, third probability is less than third predetermined threshold value Store location influence factor screens, as the first influence factor.For example, this 10 shops A, B, C, D, E, F, G, H, I, J In the influence factor of position, this corresponding third probability of 5 store location influence factors of A, B, C, D, E is less than third predetermined probabilities, F, this corresponding third probability of 5 store location influence factors of G, H, I, J be greater than or equal to third predetermined probabilities, then by A, B, C, D, E is screened as the first influence factor.
If third probability is less than third predetermined threshold value, show the corresponding store location influence factor of the third probability to shop Paving efficiency has more apparent influence, therefore, can filter out the first shadow according to the size of third probability and third predetermined threshold value The factor of sound.
Fig. 3 is referred to, Fig. 3 shows shop site selecting method provided by the embodiments of the present application, includes the following steps:
Step S210, according to training complete multiple-factor regression model, obtain as the multiple-factor regression model oneself Multiple ends of variable select influence factor.
Multiple-factor regression model can be shop site selection model above, after multiple-factor regression model determines, this mostly because The independent variable of sub- regression model is that the influence factor of composition multiple-factor equation can determine.The multiple-factor side constructed with A, C, E For journey, if the multiple-factor regression model that training is completed is the multiple-factor equation of A, C, E building, the whole choosing of the multiple-factor equation Influence factor is A, C, E.
Step S220 will be multiple subregions to addressing region division, by each subregion in the multiple subregion It selects the corresponding truthful data of influence factor to bring the multiple-factor regression model into the multiple end, obtains each subregion Prediction shop efficiency.
It is the region of user's shop addressing to be carried out to addressing region, which can be a biggish panel region, such as Some prefecture-level city is also possible to some city of certain prefecture-level city.Can will be to addressing region division according to preset side length Multiple subregions, for example, can with 250 meters of side length of square area will to addressing region division be multiple subregions, can also It is divided with treating addressing region with other side lengths, such as a length of 200 meters, the rectangular area that width is 150 meters selects area to band Domain is divided.It is limitation to the application that the concrete shape and size of subregion, which should not be construed, but each subregion In statistics each factor numerical value need into cross normalization or other statistical calculations, to guarantee its science in statistical analysis.
For each subregion, the specific data of influence factor are selected to substitute into multiple-factor regression model the end in subregion, Respectively obtain the prediction shop efficiency of each subregion.Influence factor will be selected to substitute into before multiple-factor regression model eventually, it can be right It selects influence factor to be standardized eventually, each end is made to select influence factor in same range scale.
Step S230 carries out ranking to each subregion according to the sequence of the prediction shop efficiency from high to low.
Step S240 determines the addressing in new shop according to the ranking and the existing store location of user.
After multiple-factor regression model determines, the independent variable of the multiple-factor regression model be constitute the influence of multiple-factor equation because Element can determine, then to each subregion in multiple subregions, select determining influence factor, then will affect factor The multiple-factor regression model is substituted into, prediction shop efficiency is obtained, it then can be existing according to the size of prediction shop efficiency and user Some store locations determine the selection of new store location, obtain prediction shop efficiency using model, further according to prediction shop effect The position that new shop can be filtered out with the existing shop of user compared with prior art can be more efficiently and more objective Accurately carry out the selection of shop shop location.
Step S240 specifically comprises the following steps:
According to subregion described in the sequential selection of the ranking from high to low.
Judge whether the distance between the subregion selected and the existing store location of the user are more than default Distance.
If so, using the subregion selected as the addressing in the new shop.
If it is not, the subregion of a ranking lower than the subregion in the ranking is obtained, it is described lower than the subregion The subregion of one ranking is the new subregion selected, and executes the step " subregion and the use that judgement is selected Whether the distance between existing store location in family is more than pre-determined distance ".
Pre-determined distance generates the minimum range for the influence vied each other between same type shop, if between two shops Distance is less than pre-determined distance, then two shops, which will generate, vies each other.The selection of pre-determined distance is related with area, such as flat Original area can take 1000 meters, can take 800 meters in knob, the specific length value of pre-determined distance, which should not be construed, is Limitation to the application.
Fig. 6 is referred to, Fig. 6 shows the application scenarios schematic diagram of shop site selecting method provided by the embodiments of the present application, Shown in Fig. 6 in addressing region, several sub-regions have been divided with the square that side length is 250 meters, wherein plot 1 is to use The existing shop in family, plot 2 are the prediction highest subregion of shop efficiency, only due to the distance between plot 2 and plot 1 250 meters, it is less than pre-determined distance, therefore skip plot 2.Plot 3 is to predict that shop efficiency is only second to the subregion in plot 2, plot 3 The distance between plot 1 is 5550 meters, is more than pre-determined distance, therefore plot 3 can be used as the position in new shop.
If user selects at one position only to open up new shop, shop is chosen this and is just completed；If user be intended to select to Position opens up new shop at few two, then is selecting the selection that will continue to carry out position at one behind position, and user continues to select When new position, the position selected can be denoted as the existing shop of user.For example, plot 4 is that prediction shop efficiency is only second to The plot in plot 3, although plot 4 is more than pre-determined distance at a distance from plot 1, due to plot 4 and the new position just selected The distance between plot 3 is less than pre-determined distance, therefore plot 4 cannot function as the position in the new shop in another place.It repeats the above process, Until meet user addressing quantitative requirement or to can not be selected in addressing region with other selection of land block distance be more than it is default away from From plot until.
According to the sequential selection subregion of prediction shop efficiency from high to low, and judge subregion and the existing shop of user Whether the distance between position is more than pre-determined distance, if if so, showing that the subregion selected as shop, will not influence use Family have shop business, therefore it can be selected that if it is not, then explanation may have shop to user business impact, because This can skip the subregion, continue to execute distance to the subregion of the low ranking in prediction shop efficiency ratio current sub-region and sentence It is disconnected, it is higher and will not influence the subregion that user has shop business until obtaining ranking.Which has comprehensively considered user's Interests, so that the store location of selection benefits user to the greatest extent.
After user sets up shop for a period of time, the data of new shop efficiency can be collected, and with new shop Setting and shop periphery situation change, old shop can also be had an impact.The data newly obtained can be input to shop In site selection model, model is optimized.Shop site selection model can also be laid out existing market and assess, and can be excellent Change existing layout, selects new suitable store location.
Shop site selecting method provided by the embodiments of the present application can reduce site assessment work, can get rid of the limitation of paving source, The universe that selection region can quickly be treated integrally is assessed, and assessment and the efficiency of decision-making are improved, and saves human cost.
Fig. 4 is referred to, Fig. 4 shows the training device of shop site selection model provided by the embodiments of the present application, it should be appreciated that should Device 400 is corresponding to Fig. 2 embodiment of the method with above-mentioned Fig. 1, is able to carry out each step that above method embodiment is related to, the dress Set 400 specific functions may refer to it is described above, it is appropriate herein to omit detailed description to avoid repeating.Device 400 wraps Include at least one operation system that can be stored in memory or be solidificated in device 400 in the form of software or firmware (firmware) Software function module in system (operating system, OS).Specifically, which includes:
Equation constructs module 410, for constructing multiple be used for according at least two first influence factors of store location Predict the multiple-factor equation of shop efficiency, the shop efficiency reflects the business revenue ability in corresponding shop.
Equation screening module 420 meets significance test requirement for filtering out from multiple multiple-factor equations At least one multiple-factor equation.
Degree of fitting determining module 430, for determining the prediction shop efficiency and reality of at least one multiple-factor equation The degree of fitting of shop efficiency.
Equation determining module 440, for the highest multiple-factor equation of degree of fitting to be determined as the shop site selection model Equation.
Equation construct module 410 include: the first related coefficient submodule, for calculate it is described at least two first influence because Related coefficient in element between any two；Second probability calculation submodule, for make respectively it is described at least two first influence because The related coefficient of the first influence factor of every two is genuine multiple second it is assumed that and calculating separately the multiple second hypothesis in element Each of second assume occur mistake the second probability；Equation constructs submodule, for when in the presence of less than the second preset threshold The second probability when, second probability is filtered out from least two first influence factor more than or equal to described second First influence factor of preset threshold constructs the multiple-factor equation.Submodule is randomly choosed, if equal for second probability More than or equal to second preset threshold, the first influence factor structure is randomly choosed from least two first influence factor Build the multiple-factor equation.
Equation screening module 420 includes: the first hypothesis setting submodule, for for each of multiple multiple-factor equations Multiple-factor equation, making the multiple-factor equation fitted is genuine first hypothesis；First probability calculation submodule, for for The first influence factor of each of at least two first influence factor corresponding with each multiple-factor equation, is counted respectively It calculates described first and assumes the first probability for mistake occur；It upchecks submodule, for working as multiple-factor equation corresponding at least two When first probability of the first influence factor of each of a first influence factor is respectively less than the first preset threshold, determine it is described mostly because Sub- equation meets significance test requirement.
Described device further includes that the test fails submodule, for being influenced when multiple-factor equation corresponding at least two first In factor, there are when the first influence factor that first probability is greater than or equal to first preset threshold, determine described more Factor Equations do not meet significance test requirement.
Influence factor screening module, for filtered out from multiple store location influence factors at least two first influence because Element.
The influence factor screening module includes: the second related coefficient submodule, is influenced for calculating multiple store locations The related coefficient between each store location influence factor and the shop efficiency in factor；Third probability calculation submodule, For making the phase in multiple store location influence factors between each store location influence factor and the shop efficiency respectively Relationship number is genuine multiple thirds it is assumed that and calculating separately each third hypothesis appearance mistake in the multiple third hypothesis Third probability；Influence factor screens submodule, for being less than the store location influence factor of third predetermined threshold value when third probability In the presence of, the store location shadow that third probability is less than third predetermined threshold value is filtered out from the multiple store location influence factor The factor of sound.
The embodiment of the present application also provides a kind of shop addressing devices, it should be appreciated that the device and above-mentioned Fig. 3 embodiment of the method It is corresponding, it is able to carry out each step that above method embodiment is related to, the specific function of the device may refer to retouching above It states, it is appropriate herein to omit detailed description to avoid repeating.Device includes that at least one can be with software or firmware (firmware) Form is stored in memory or is solidificated in the software function mould in the operating system (operating system, OS) of device Block.Specifically, which includes:
Influence factor is selected to obtain module eventually, for any optional implementation institute according to first aspect or first aspect The multiple-factor regression model that the method training stated is completed obtains multiple choosings eventually of the independent variable as the multiple-factor regression model Influence factor.
Subregion efficiency obtains module, for that will be multiple subregions to addressing region division, by the multiple subregion In each subregion and the multiple end select the corresponding truthful data of influence factor to bring the multiple-factor regression model into, obtain The prediction shop efficiency of each subregion.
Subregion ranking module, for the sequence according to the prediction shop efficiency from high to low, to each sub-district Domain carries out ranking.
New shop addressing module, for determining the choosing in new shop according to the ranking and the existing store location of user Location.
New shop addressing module includes region addressing submodule, for the sequential selection institute according to the ranking from high to low State subregion；Distance Judgment submodule, the subregion for judging to select and the existing store location of the user it Between distance whether be more than pre-determined distance；First result submodule, for using the subregion selected as the new shop The addressing of paving；Second result submodule, it is described for obtaining the subregion of a ranking lower than the subregion in the ranking The subregion of a ranking lower than the subregion is the new subregion selected, and executing step, " judgement is selected described Whether the distance between subregion and the existing store location of the user are more than pre-determined distance ".
It is apparent to those skilled in the art that for convenience and simplicity of description, the device of foregoing description Specific work process, no longer can excessively be repeated herein with reference to the corresponding process in preceding method.
The application also provides a kind of electronic equipment, and Fig. 5 is the structural block diagram of the electronic equipment 500 in the embodiment of the present application, As shown in Figure 5.Electronic equipment 500 may include that processor 510, communication interface 520, memory 530 and at least one communication are total Line 540.Wherein, communication bus 540 is for realizing the direct connection communication of these components.Wherein, equipment in the embodiment of the present application Communication interface 520 be used to carry out the communication of signaling or data with other node devices.Processor 510 can be a kind of integrated electricity Road chip, the processing capacity with signal.Above-mentioned processor 510 can be general processor, including central processing unit (Central Processing Unit, abbreviation CPU), network processing unit (Network Processor, abbreviation NP) etc.；May be used also Be digital signal processor (DSP), specific integrated circuit (ASIC), ready-made programmable gate array (FPGA) or other can compile Journey logical device, discrete gate or transistor logic, discrete hardware components.It may be implemented or execute the embodiment of the present application In disclosed each method, step and logic diagram.General processor can be microprocessor or the processor 510 can also be with It is any conventional processor etc..
Memory 530 may be, but not limited to, random access memory (Random Access Memory, RAM), only It reads memory (Read Only Memory, ROM), programmable read only memory (Programmable Read-Only Memory, PROM), erasable read-only memory (Erasable Programmable Read-Only Memory, EPROM), Electricallyerasable ROM (EEROM) (Electric Erasable Programmable Read-Only Memory, EEPROM) etc.. Computer-readable instruction fetch is stored in memory 530, when the computer-readable instruction fetch is executed by the processor 510 When, electronic equipment 500 can execute each step that above-mentioned Fig. 1 is related to Fig. 3 embodiment of the method.
Electronic equipment 500 can also include storage control, input-output unit, audio unit, display unit.
The memory 530, processor 510, Peripheral Interface, input-output unit, audio unit, is shown storage control Show that each element of unit is directly or indirectly electrically connected between each other, to realize the transmission or interaction of data.For example, these elements It can be realized and be electrically connected by one or more communication bus 540 between each other.The processor 510 is for executing memory The executable module stored in 530, such as software function module or computer program that device 400 includes.
Input-output unit is used to be supplied to user input data and realizes user and the server (or local terminal) Interaction.The input-output unit may be, but not limited to, mouse and keyboard etc..
Audio unit provides a user audio interface, may include one or more microphones, one or more loudspeaking Device and voicefrequency circuit.
Display unit provided between the electronic equipment and user an interactive interface (such as user interface) or It is referred to for display image data to user.In the present embodiment, the display unit can be liquid crystal display or touch-control is aobvious Show device.It can be the capacitance type touch control screen or resistance type touch control screen of support single-point and multi-point touch operation if touch control display Deng.Single-point and multi-point touch operation is supported to refer to that touch control display can sense one or more positions on the touch control display The touch control operation setting place while generating, and the touch control operation that this is sensed transfers to processor to be calculated and handled.
Input-output unit is used to be supplied to the interaction that user input data realizes user and processing terminal.The input is defeated Unit may be, but not limited to, out, mouse and keyboard etc..
It is appreciated that structure shown in fig. 5 is only to illustrate, the electronic equipment 500 may also include more than shown in Fig. 5 Perhaps less component or with the configuration different from shown in Fig. 5.Each component shown in Fig. 5 can use hardware, software Or combinations thereof realize.
The application also provides a kind of computer readable storage medium, is stored with computer on the computer readable storage medium Program executes method described in embodiment of the method when the computer program is run by processor.
The application also provides a kind of computer program product to be made when the computer program product is run on computers It obtains computer and executes method described in embodiment of the method.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description Specific work process, no longer can excessively be repeated herein with reference to the corresponding process in preceding method.
It should be noted that all the embodiments in this specification are described in a progressive manner, each embodiment weight Point explanation is the difference from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For device class embodiment, since it is basically similar to the method embodiment, so being described relatively simple, related place ginseng See the part explanation of embodiment of the method.
In several embodiments provided herein, it should be understood that disclosed device and method can also pass through it Its mode is realized.The apparatus embodiments described above are merely exemplary, for example, the flow chart and block diagram in attached drawing are aobvious The device of multiple embodiments according to the application, architectural framework in the cards, the function of method and computer program product are shown It can and operate.In this regard, each box in flowchart or block diagram can represent one of a module, section or code Point, a part of the module, section or code includes one or more for implementing the specified logical function executable Instruction.It should also be noted that function marked in the box can also be attached to be different from some implementations as replacement The sequence marked in figure occurs.For example, two continuous boxes can actually be basically executed in parallel, they sometimes may be used To execute in the opposite order, this depends on the function involved.It is also noted that each of block diagram and or flow chart The combination of box in box and block diagram and or flow chart can be based on the defined function of execution or the dedicated of movement The system of hardware is realized, or can be realized using a combination of dedicated hardware and computer instructions.
In addition, each functional module in each embodiment of the application can integrate one independent portion of formation together Point, it is also possible to modules individualism, an independent part can also be integrated to form with two or more modules.
The above, the only specific embodiment of the application, but the protection scope of the application is not limited thereto, it is any Those familiar with the art within the technical scope of the present application, can easily think of the change or the replacement, and should all contain Lid is within the scope of protection of this application.Therefore, the protection scope of the application shall be subject to the protection scope of the claim.

Claims (10)

1. a kind of training method of shop site selection model, which is characterized in that the described method includes:
Multiple multiple-factor equations for being used to predict shop efficiency are constructed according at least two first influence factors of store location, The shop efficiency reflects the business revenue ability in corresponding shop；
At least one the multiple-factor equation for meeting significance test requirement is filtered out from multiple multiple-factor equations；
Determine the degree of fitting of prediction the shop efficiency and practical shop efficiency of at least one multiple-factor equation；
The highest multiple-factor equation of degree of fitting is determined as to the equation of the shop site selection model.
2. the method according to claim 1, wherein each multiple-factor equation in multiple multiple-factor equations There is corresponding at least two first influence factor；Described filter out from multiple multiple-factor equations meets significance test It is required that at least one multiple-factor equation, comprising:
For each multiple-factor equation in multiple multiple-factor equations, making the multiple-factor equation fitted is genuine first false If；
For each of at least two first influence factor corresponding with each multiple-factor equation first influence because Element calculates separately described first and assumes the first probability for mistake occur；
If the first probability of the first influence factor of each of corresponding at least two first influence factor of multiple-factor equation is small In the first preset threshold, determine that the multiple-factor equation meets significance test requirement.
3. according to the method described in claim 2, it is characterized in that, it is described calculate separately it is described first assume occur mistake After first probability, the method also includes:
If in corresponding at least two first influence factor of multiple-factor equation, there are first probability to be greater than or equal to described the First influence factor of one preset threshold, it is determined that the multiple-factor equation does not meet significance test requirement.
4. the method according to claim 1, wherein described construct according at least two first influence factor It is multiple for predicting the multiple-factor equation of shop efficiency out, comprising:
Calculate related coefficient between any two at least two first influence factor；
It is genuine multiple for making the related coefficient of the first influence factor of every two at least two first influence factor respectively Second it is assumed that and calculate separately it is the multiple second assume each of second assume occur mistake the second probability；
If it exists less than the second probability of the second preset threshold, then filtered out from least two first influence factor described The first influence factor that second probability is greater than or equal to second preset threshold constructs the multiple-factor equation.
5. according to the method described in claim 4, it is characterized in that, it is described calculate separately it is the multiple second assume in it is every After a second assumes the second probability for mistake occur, the method also includes:
If second probability be all larger than or be equal to second preset threshold, from least two first influence factor with Machine selects the first influence factor to construct the multiple-factor equation.
6. the method according to claim 1, wherein it is described according at least two first of store location influence because Before element constructs multiple multiple-factor equations for predicting shop efficiency, the method also includes:
At least two first influence factors are filtered out from multiple store location influence factors.
7. according to the method described in claim 6, it is characterized in that, it is described filtered out from multiple store location influence factors to Few two the first influence factors, comprising:
It calculates related between each store location influence factor and the shop efficiency in multiple store location influence factors Coefficient；
The phase in multiple store location influence factors between each store location influence factor and the shop efficiency is made respectively Relationship number is genuine multiple thirds it is assumed that and calculating separately each third hypothesis appearance mistake in the multiple third hypothesis Third probability；
If third probability be less than third predetermined threshold value store location influence factor exist, from the multiple store location influence because The store location influence factor that third probability is less than third predetermined threshold value is filtered out in element, first influence factor is that third is general Rate is less than the store location influence factor of third predetermined threshold value.
8. a kind of shop site selecting method, which is characterized in that the described method includes:
According to the multiple-factor regression model that the training method training of any one of such as claim 1 to 7 shop site selection model is completed, obtain It obtains and selects influence factor as multiple ends of the independent variable of the multiple-factor regression model；
It will be multiple subregions to addressing region division, by each subregion and the multiple choosing eventually in the multiple subregion The corresponding truthful data of influence factor brings the multiple-factor regression model into, obtains the prediction shop effect of each subregion Energy；
According to the sequence of the prediction shop efficiency from high to low, ranking is carried out to each subregion；
According to the ranking and the existing store location of user, the addressing in new shop is determined.
9. according to the method described in claim 8, it is characterized in that, described according to the ranking and user existing shop position It sets, determines the addressing in new shop, comprising:
According to subregion described in the sequential selection of the ranking from high to low；
Judge whether the distance between the subregion selected and the existing store location of the user are more than pre-determined distance；
If so, using the subregion selected as the addressing in the new shop；
If it is not, the subregion of a ranking lower than the subregion in the ranking is obtained, described one lower than the subregion The subregion of ranking is the new subregion selected, and executing step, " subregion and the user that judgement is selected are existing Whether the distance between some store locations are more than pre-determined distance ".
10. a kind of training device of shop site selection model, which is characterized in that described device includes:
Equation constructs module, constructs at least two first influence factors according to store location multiple for predicting shop The multiple-factor equation of efficiency, the shop efficiency reflect the business revenue ability in corresponding shop；
Equation screening module, for filtered out from multiple multiple-factor equations meet significance test requirement at least one Multiple-factor equation；
Degree of fitting determining module, for determining the prediction shop efficiency and practical shop efficiency of at least one multiple-factor equation Degree of fitting；
Equation determining module, for the highest multiple-factor equation of degree of fitting to be determined as to the equation of the shop site selection model.
CN201910416777.9A 2019-05-20 2019-05-20 Training method, shop site selecting method and the device of shop site selection model Pending CN110135913A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910416777.9A CN110135913A (en) 2019-05-20 2019-05-20 Training method, shop site selecting method and the device of shop site selection model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910416777.9A CN110135913A (en) 2019-05-20 2019-05-20 Training method, shop site selecting method and the device of shop site selection model

Publications (1)

Publication Number Publication Date
CN110135913A true CN110135913A (en) 2019-08-16

Family

ID=67571291

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910416777.9A Pending CN110135913A (en) 2019-05-20 2019-05-20 Training method, shop site selecting method and the device of shop site selection model

Country Status (1)

Country Link
CN (1) CN110135913A (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110837930A (en) * 2019-11-07 2020-02-25 腾讯科技（深圳）有限公司 Address selection method, device, equipment and storage medium
CN111539764A (en) * 2020-04-17 2020-08-14 南京邮电大学 Big data multiple access selection method based on submodular function
CN111744194A (en) * 2020-07-03 2020-10-09 广州要玩娱乐网络技术股份有限公司 Virtual game control method, device and server
CN111985576A (en) * 2020-09-02 2020-11-24 南宁师范大学 Shop address selection method based on decision tree
CN112184282A (en) * 2020-09-03 2021-01-05 长安大学 Cinema site selection model establishing method, cinema site selection method and cinema site selection platform
CN112308594A (en) * 2019-09-29 2021-02-02 北京京东尚科信息技术有限公司 Address selection system, method, electronic device and storage medium
CN114004661A (en) * 2021-12-29 2022-02-01 杭州铭信信息科技有限公司 Store information processing method, device, equipment and storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107563803A (en) * 2017-08-24 2018-01-09 北京工商大学 A kind of market area partition method based on cost grid
CN108733619A (en) * 2018-05-17 2018-11-02 北京林业大学 Global arbitrary forest bottom class growth prediction model quantitative estimation method
CN108918989A (en) * 2018-04-28 2018-11-30 国网上海市电力公司 A kind of military service cable residue lifetime estimation method based on multi-parameter
CN109214952A (en) * 2017-06-29 2019-01-15 辽宁工程技术大学 A kind of traveller's decision model construction method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109214952A (en) * 2017-06-29 2019-01-15 辽宁工程技术大学 A kind of traveller's decision model construction method
CN107563803A (en) * 2017-08-24 2018-01-09 北京工商大学 A kind of market area partition method based on cost grid
CN108918989A (en) * 2018-04-28 2018-11-30 国网上海市电力公司 A kind of military service cable residue lifetime estimation method based on multi-parameter
CN108733619A (en) * 2018-05-17 2018-11-02 北京林业大学 Global arbitrary forest bottom class growth prediction model quantitative estimation method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
李湘桂: "基于多元回归分析的高校周边餐厅选址模型的设计与实现", 《中国优秀硕士学位论文全文数据库-信息科技辑》 *
韩曾丽等: "财经类高校图书馆文献资源建设经费投入分析及相关性研究", 《图书馆学研究》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112308594A (en) * 2019-09-29 2021-02-02 北京京东尚科信息技术有限公司 Address selection system, method, electronic device and storage medium
CN110837930A (en) * 2019-11-07 2020-02-25 腾讯科技（深圳）有限公司 Address selection method, device, equipment and storage medium
CN110837930B (en) * 2019-11-07 2023-09-19 腾讯科技（深圳）有限公司 Site selection method, device, equipment and storage medium
CN111539764A (en) * 2020-04-17 2020-08-14 南京邮电大学 Big data multiple access selection method based on submodular function
CN111744194A (en) * 2020-07-03 2020-10-09 广州要玩娱乐网络技术股份有限公司 Virtual game control method, device and server
CN111985576A (en) * 2020-09-02 2020-11-24 南宁师范大学 Shop address selection method based on decision tree
CN111985576B (en) * 2020-09-02 2023-11-03 南宁师范大学 Shop site selection method based on decision tree
CN112184282A (en) * 2020-09-03 2021-01-05 长安大学 Cinema site selection model establishing method, cinema site selection method and cinema site selection platform
CN114004661A (en) * 2021-12-29 2022-02-01 杭州铭信信息科技有限公司 Store information processing method, device, equipment and storage medium

Similar Documents

Publication Publication Date Title
CN110135913A (en) Training method, shop site selecting method and the device of shop site selection model
Yadav et al. PyTOPS: A Python based tool for TOPSIS
Millward-Hopkins et al. Fully integrated modelling for sustainability assessment of resource recovery from waste
Kuai et al. An application of system dynamics for evaluating planning alternatives to guide a green industrial transformation in a resource-based city
Medury et al. Simultaneous network optimization approach for pavement management systems
Deichmann Geographic aspects of inequality and poverty
Liu et al. Exploring the co-evolution of the geographies of air transport aviation and corporate networks
Bruni et al. Probabilistically constrained models for efficiency and dominance in DEA
Ustun Multi-choice goal programming formulation based on the conic scalarizing function
CN106803799B (en) Performance test method and device
CN106502815A (en) A kind of abnormal cause localization method, device and computing device
Brendel et al. Information systems in the context of sustainable mobility services: A literature review and directions for future research
CN109447658A (en) The generation of anti-fraud model and application method, device, equipment and storage medium
García-Magariño et al. Agent-based simulation of real-estate transactions
Goh et al. Integration model of fuzzy AHP and life-cycle cost analysis for evaluating highway infrastructure investments
CN113516417A (en) Service evaluation method and device based on intelligent modeling, electronic equipment and medium
Lee et al. The relative impacts of selected practices on project cost and schedule
Ferreira et al. A geometric aggregation of performance indicators considering regulatory constraints: An application to the urban solid waste management
CN107145995A (en) Production environment safety prediction methods, devices and systems
CN110059083A (en) A kind of data evaluation method, apparatus and electronic equipment
CN112613983B (en) Feature screening method and device in machine modeling process and electronic equipment
CN109635996A (en) Room rate tendency look-forward model construction method, device and storage medium
Yang et al. Ecological risk assessment based on road network development analysis of Xiamen city, China
Migliaccio et al. Empirical comparison of methods for estimating location cost adjustments factors
CN106648895A (en) Data processing method and device, and terminal

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20190816