Summary of the invention
This specification one or more embodiment describes a kind of determining product peers amount in the distribution of multiple regions
The method and device of data can determine the product peers amount of each region based on Computerized Information Processing Tech, can be with
So that realizing higher conversion rate of products in the case that ensemble average peers amount meets constraint condition.
According in a first aspect, provide a kind of method of the distributed data of determining product peers amount in multiple regions,
Include:
Obtain corresponding with pre-set product multiple regions to be put data acquisition system, including with the first region pair
The first data acquisition system answered, first data acquisition system include the user information of multiple potential users in the first region, each
The user information of potential user includes that each potential user carries out with the pre-set product respectively under multiple peers amounts
The exchange probability of exchange;
Based on each potential user in the first data acquisition system under the first peers amount corresponding exchange probability, determine described in
First region corresponding conversion ratio under the first peers amount;First peers amount is to appoint in the multiple peers amount
One reciprocity stock number;
It is respectively corresponded under the multiple peers amount based on the multiple peers amount and first region
Conversion ratio, construct the first broken line;
From first broken line between minimum peers amount coordinate points and other larger peers amount coordinate points tiltedly
Optimal slope is determined in rate, and using the corresponding larger peers amount of optimal slope as candidate's equity money of first region
Source amount;
Candidate peers amount based on the multiple region determines average peers amount;
When the average peers amount is greater than preset threshold, the candidate peers amount for exporting the multiple region is made
For the distributed data.
In some embodiments, each potential user under multiple peers amounts respectively with the pre-set product into
The exchange probability of row exchange is that the user characteristics of the potential user and each peers amount are inputted exchange probabilistic forecasting respectively
Model is predicted to obtain;
The exchange Probabilistic Prediction Model is trained to obtain based on the historical behavior information to multiple sample of users;It is described
The historical behavior information of sample of users includes user characteristics, history peers amount, switch label.
In some embodiments, the user characteristics include, and user property feature and use the default production with user
The related user's scene characteristic of the scene of product.
In some embodiments, the abscissa of first broken line corresponds to the multiple peers amount, ordinate pair
The corresponding conversion ratio under the multiple peers amount of first region described in Ying Yu, the optimal slope are described first
Minimum slope on broken line between minimum peers amount coordinate points and other larger peers amount coordinate points in slope.
In some embodiments, the abscissa of first broken line corresponds to first region in the multiple equity money
Corresponding conversion ratio under the amount of source, ordinate correspond to the multiple peers amount, and the optimal slope is described first
Greatest gradient on broken line between minimum peers amount coordinate points and other larger peers amount coordinate points in slope.
In some embodiments, the candidate peers amount based on the multiple region determines average peers
Amount includes:
The size of population of candidate peers amount and each region based on the region the multiple region Zhong Ge, really
Allocate equal peers amount.
In some embodiments, the candidate peers amount based on the multiple region determines average peers
Amount includes:
Based on the candidate peers amount of the region the multiple region Zhong Ge and the quantity of potential user, average pair is determined
Etc. stock numbers.
In some embodiments, the method also includes:
It is corresponding multiple optimal from the multiple region when average peers amount is less than or equal to the preset threshold
Global optimum's slope is determined in slope, and global optimum's slope is corresponded into region as region to be optimized, it will be described in the region
The corresponding candidate peers amount of global optimum's slope is as peers amount to be optimized;
On the corresponding broken line in the region to be optimized, from the peers amount coordinate points to be optimized and be higher than it is described to
Optimize between the peers amount coordinate points of peers amount in slope, determines new optimal slope, and will the newly optimal slope
New candidate peers amount of the corresponding larger peers amount as the region to be optimized;
Based on the candidate peers amount of the region except the new candidate peers amount and the region to be optimized, more
Newly be averaged peers amount;
Above step is repeated, until updated average peers amount is greater than the preset threshold.
According to second aspect, a kind of device of the distributed data of determining product peers amount in multiple regions is provided,
Include:
Acquiring unit is configured to obtain data acquisition system corresponding with pre-set product multiple regions to be put, wherein
Including the first data acquisition system corresponding with the first region, first data acquisition system includes multiple potential users in the first region
User information, the user information of each potential user include the potential user under multiple peers amounts respectively with it is described pre-
If the exchange probability that product swaps;
It is corresponding under the first peers amount to be configured to each potential user in the first data acquisition system for first determination unit
Exchange probability, determine first region corresponding conversion ratio under the first peers amount;First peers amount is institute
State any peers amount in multiple peers amounts;
Construction unit is configured to the multiple peers amount and first region in the multiple equity money
Corresponding conversion ratio under the amount of source, constructs the first broken line;
Second determination unit is configured to from first broken line minimum peers amount coordinate points and other larger equities
Determine optimal slope between stock number coordinate points in slope, and using the corresponding larger peers amount of optimal slope as described the
The candidate peers amount of one region;
Third determination unit is configured to the candidate peers amount of the multiple region, determines average peers
Amount;
Output unit is configured to export the multiple region when the average peers amount is greater than preset threshold
Distributed data described in candidate peers amount.
According to the third aspect, a kind of computer readable storage medium is provided, computer program is stored thereon with, when described
When computer program executes in a computer, enable computer execute first aspect method.
According to fourth aspect, a kind of calculating equipment, including memory and processor are provided, is stored in the memory
Executable code, when the processor executes the executable code, the method for realizing first aspect.
According to the method and apparatus that this specification embodiment provides, calculating equipment can be used and pass through multiple Regional Distributions pair
The user information answered determines each region in conversion ratio reduction of speed compared with candidate's equity slow, in the very fast situation of peers amount speedup
Stock number judges whether the average peers amount of the candidate peers amount of each region meets constraint in conjunction with preset threshold
Condition may thereby determine that out and not only meet constraint condition, but also the product peers amount of each region with higher conversion;And
And this illustrates that the scheme that embodiment provides is that a kind of computer information processing scheme can avoid the consuming of a large amount of manpower and material resources, effect
Rate is high.
Specific embodiment
With reference to the accompanying drawing, the scheme provided this specification is described.
It is readily appreciated that, when non-grant, product demand person needs or is ready available with specific using its
The resource of value exchanges specific product with product supplier.This, which obtains resource with characteristic value, can be described as the equity money of the product
Source.Product can be products in kind, such as food, clothes etc.;Product may be virtual product, such as the card of game points, public transport
Period card etc..Resource can be currency, labor service time, article, virtual resource (such as " energy " in ant forest) etc..It is easy
Understand, when resource is currency, peers amount shows as the price of product.
In general, product supplier is intended to launch product in multiple regions.In this case, product supplier can be directed to
Different peers amounts are arranged in product different zones to be put, have higher conversion to pursue the product in each region
Rate.
It, can be real under the entity line of different geographical in merchandise sales scene by taking product is vendible commodity as an example
Body shop merchandising, such as clothes, food, electronic goods etc..Can also by electric business platform (such as Alipay etc.) towards
User's merchandising of different geographical, such as public transport period card, subway period card etc..
It, can be for differently in view of factors such as the consumption habit of different geographical, income level, the same type commodity that have
Different peers amounts, such as price is arranged in domain, so that commodity are in multiple regions conversion ratio all with higher.
In this specification embodiment, conversion ratio refers to using the number of users and product of existing Resource Exchange product up to use
The ratio of amount.Specifically in commercial articles vending scene, the number of users and commodity that conversion ratio refers to purchase commodity are up to number of users
Ratio can be used for measuring the consumers' acceptable degree of the commodity.
It is readily appreciated that, peers amount is lower, and conversion ratio can be higher.In view of factors such as costs, product is on different ground
The ensemble average peers amount in domain will meet certain constraint condition.
Accordingly, it is considered to which arriving above-mentioned various factors for the different peers amount of different areal settings is a system
And cumbersome engineering.In particular, in practical applications, needing to carry out Dynamic Pricing for each region, it can according to specific
Readjust prices period (such as 1 day, 1 week, 1 month etc.), the product peers amount of each region is adjusted, to pursue higher conversion
Rate.Dynamic Pricing further increases the workload of peers amount setting.
For this purpose, this specification embodiment, which provides one kind, is based on Computerized Information Processing Tech fixed output quota product peers really
Measure the method in the distributed data of multiple regions.The method that equipment executes the offer of this specification embodiment is calculated, can be based on multiple
The corresponding user information of Regional Distribution determines each region in conversion ratio reduction of speed compared in slow, the very fast situation of peers amount speedup
Candidate peers amount, the average peers amount for judging the candidate peers amount of each region in conjunction with preset threshold is
It is no to meet constraint condition, it may thereby determine that out and not only meet constraint condition, but also the product pair of each region with higher conversion
Etc. stock numbers.
Next, with reference to Fig. 1, fixed output quota product peers amount really is provided to this specification embodiment and is divided in multiple regions
The method of cloth data is specifically introduced.The method can by it is any have calculating, the device of processing capacity, equipment, platform,
Device clusters execute.As shown in Figure 1, described method includes following steps: step 100 obtains and to be put more of pre-set product
The corresponding data acquisition system in a region, including the first data acquisition system corresponding with the first region, first data set
The user information including multiple potential users in the first region is closed, the user information of each potential user includes that this is each potential
User exchanges probability with what the pre-set product swapped respectively under multiple peers amounts;Step 102 is counted based on first
According to potential user each in set under the first peers amount corresponding exchange probability, determine first region first equity
Corresponding conversion ratio under stock number;First peers amount is any peers amount in the multiple peers amount;Step
104, based on the multiple peers amount and first region corresponding turn under the multiple peers amount
Rate constructs the first broken line;Step 106, minimum peers amount coordinate points and other larger equities from first broken line
Determine optimal slope between stock number coordinate points in slope, and using the corresponding larger peers amount of optimal slope as described the
The candidate peers amount of one region;Step 108, the candidate peers amount based on the multiple region determine average equity
Stock number;Step 110, when the average peers amount is greater than preset threshold, export candidate's equity of the multiple region
Stock number is as the distributed data.
Next, above steps is specifically introduced in conjunction with specific example.
Firstly, in step 100, data acquisition system corresponding with pre-set product multiple regions to be put is obtained,
In include the first data acquisition system corresponding with the first region, first data acquisition system includes multiple potential use in the first region
The user information at family, the user information of each potential user include each potential user under multiple peers amounts respectively with
The exchange probability that the pre-set product swaps.
Potential user refers to the user for potentially having demand to product.It is commodity in product, resource is currency, peers
In the case that amount is commodity price, exchange probability is referred to as purchase probability.
In some embodiments, product is specifically as follows through electric business platform (such as Alipay etc.) or solid shop/brick and mortar store sale
The public transport period card, such as the public transport period card, the subway period card etc.;Or shared bicycle period card;Etc., herein
It will not enumerate.By taking subway period card as an example, monthly card, Zhou Ka, secondary card can be divided into.Specifically by taking subway monthly card as an example, use
After family has purchased subway monthly card, when taking ground body in subway monthly card comes into force rise constantly one month, the subway monthly card can be relied on
Enjoy specific discount or preferential.
In some embodiments, region can be administrative region.In one example, it is specifically as follows a certain administrative grade
City or region, for example, can be for provincial municipality directly under the Central Government, province, autonomous region etc..In another example, region can be with
City is specifically as follows the city for meeting specified conditions, such as the size of population is more than the city of preset value (such as 1,000,000), or
Person's construction has city of subway etc..
In some embodiments, region can be geographic area.By taking the geographic area of China as an example, region is specifically as follows
It is North China, northeast, northwest, East China, Central-South, southwestern.
In some embodiments, potential user can be the registration user of electric business platform (such as Alipay).It can pass through
Potential user is belonged to corresponding region by the address information in the registration information of potential user;It can also be according to potential user's line
Potential user is belonged to corresponding region by lower loco;Potential user can also be returned according to the shipping address of online shopping
Belong to corresponding region.
In some embodiments, potential user is specifically as follows the resident of each region.
It in some embodiments, can be by the corresponding potential user in the region for any region in multiple regions
User characteristics and each peers amount be input to exchange Probabilistic Prediction Model predicted, obtain potential user each right
Etc. exchange probability under stock numbers.
User characteristics may include the attributive character and user's scene characteristic of user.User property feature may include ground
Domain, age, gender, education degree, occupation etc..User's scene characteristic is to use the related spy of the scene of product to be put with user
Sign, for example, product to be put be the public transport period block in the case where, user's scene characteristic can for special time period (such as
One month) interior user is in the frequency that subway takes scene, and/or the frequency etc. of scene is taken in public transport.
It, can be by peers amount each in multiple peers amounts respectively with the user's for any potential user
User property feature, scene characteristic are spliced into a plurality of forecast sample, and forecast sample is input to exchange Probabilistic Prediction Model and is carried out
Prediction, obtains the corresponding exchange probability of each sample.In one example, referring to Fig.2,4 peers amounts can be set, respectively
For peers amount 1, peers amount 2, peers amount 3, peers amount 4, by the user property feature of any user, field
Scape feature is successively spliced into 4 samples with peers amount 1, peers amount 2, peers amount 3, peers amount 4, so
This 4 samples are input to exchange Probabilistic Prediction Model, available peers amount 1, peers amount 2, peers afterwards
Measure 3, the corresponding exchange probability of peers amount 4.
It should be noted that multiple peers amounts in this specification embodiment can be the equity money of the quasi- exchange of product
Source amount, specifically by taking commodity price as an example, commodity price can be the quasi- price sold, such as 10 yuan, 11 yuan, 12 yuan etc..?
Under some cases, peers amount can show as original peers amount multiplied by specific discount, for example including, 6 foldings, 6.5
Folding, 7 foldings, 7.5 foldings etc..
It can acquire whether multiple sample of users exchange under each history peers amount in multiple history peers amounts
The information of product, specifically can be indicated with switch label.By the corresponding switch label of each history peers amount of multiple samples,
User characteristics and each history peers amount are spliced into training sample, and carry out model training, pre- to obtain exchange probability
Survey model.Exchange Probabilistic Prediction Model can be realized by a variety of model structures, such as can be implemented as deep neural network
(deep neural network, DNN) model, GBDT model, etc..
In some embodiments, number and the exchange of product can be browsed under each peers amount with statistical sample user
The number of product, to obtain exchange probability of the sample of users under each peers amount.User characteristics are then based on, to sample
User and potential user cluster;And the exchange probability under each peers amount for mixing the sample with family is in as with it
The exchange probability under each peers amount of the user of same class cluster.The scheme that user is clustered based on user characteristics
The introduction of the prior art can be referred to, details are not described herein.
In the various embodiments described above, when potential user is that electric business platform registers user, the user characteristics of potential user can be with
But it is not limited by consumer record etc. under registration information, social record, consumption online record, line to obtain.When potential user is pair
When answering the resident of region, the user characteristics of potential user can be, but not limited to the solid shop/brick and mortar store by the region feedback information, when
Ground resident information etc. obtains.
It, can be based on region locating for each potential user, to each after obtaining the exchange probability of each potential user
Potential user classifies, to obtain the corresponding data acquisition system in each region.
Then, in a step 102, corresponding under the first peers amount based on each potential user in the first data acquisition system
Probability is exchanged, determines first region corresponding conversion ratio under the first peers amount;First peers amount is described
Any peers amount in multiple peers amounts.
As described above, conversion ratio refers to the ratio using the number of users and product for having Resource Exchange product up to number of users
Value.It therefore, can be based on each potential user in the first region at first pair for the corresponding first peers amount in the first region
Etc. corresponding exchange probability under stock numbers, conversion ratio is calculated.Specifically, by each potential user in the first region in the first equity
Under stock number it is corresponding exchange probability average value, as the first region under the first peers amount corresponding conversion ratio.
With reference to aforesaid way, available each region corresponding conversion ratio under each peers amount.
Secondly, at step 104, based on the multiple peers amount and first region in the multiple equity
Corresponding conversion ratio under stock number constructs the first broken line.
It in some embodiments, can be by each peers amount in multiple peers amounts for any region
As abscissa, by the region, corresponding conversion ratio is ordinate under each peers amount, constructs the corresponding folding in the region
Line.
It in further embodiments, can also be with right under each peers amount of the region in multiple peers amounts
The conversion ratio answered is abscissa, and each peers amount is abscissa, constructs it and corresponds to broken line.
In some embodiments, the corresponding broken line in each region is the broken line on different line charts, i.e., according to the multiple
It is corresponding to construct each region for peers amount and each region corresponding conversion ratio under the multiple peers amount
Line chart.In these implementations, each region respectively corresponds a line chart, only one broken line on each line chart.
In some embodiments, the corresponding broken line in each region can be constructed respectively on same line chart.I.e. multiplely
Each region respectively corresponds the different broken lines on the same line chart in domain.
Again, in step 106, minimum peers amount coordinate points and other larger peers amounts from the first broken line
Optimal slope is determined between coordinate points in slope, and using the corresponding larger peers amount of optimal slope as first region
Candidate peers amount.
For any region, it is based on the corresponding broken line in the region, the minimum successively calculated in multiple peers amounts is right
Etc. slope between stock numbers coordinate points and other larger peers amount coordinate points, then from the multiple slopes being calculated
Determine optimal slope, and using the corresponding larger peers amount of optimal slope as the candidate peers amount of the region.
In some embodiments, the abscissa of broken line can correspond to multiple peers amounts, and ordinate can correspond to convert
Rate, then can be using the smallest slope in the multiple slopes being calculated as optimal slope.In one example, each refering to Fig. 3
The conversion ratio of one region of behavior and the corresponding relationship record of each peers amount.It is with corresponding first region of the first row
, the line chart in Fig. 3 is the corresponding line chart in the first region.Based on the line chart in Fig. 3, R11 and R12, R11 are successively calculated
With the slope between R13, R11 and R14, R11 and R15, R11 and R16, slope is minimum between R11 and R14 in obtained slope, then
It can be using the corresponding peers amount of R14 as the candidate peers amount of the first region.
In this embodiment, when peers amount larger peers amount corresponding for minimum slope, conversion ratio is reduced
It is most slow, and peers amount increase is most fast.
In some embodiments, the abscissa of broken line can correspond to conversion ratio, and ordinate can correspond to peers amount, then
It can be using maximum slope in the multiple slopes being calculated as optimal slope.In this embodiment, in this embodiment, when right
Etc. stock numbers be greatest gradient corresponding larger peers amount when making, peers amount increase is most fast, and conversion ratio reduces most
Slowly.
Then, in step 108, the candidate peers amount based on the multiple region determines average peers amount.
In some embodiments, the simple average value that can calculate the candidate peers amount of multiple regions, is averaged
Peers amountThe peers amount that is averaged can be calculated by the following formula to obtain:
Wherein, n indicates the number of region, PiIndicate the candidate peers amount of i-th of region.
It in some embodiments, can candidate peers amount based on the region the multiple region Zhong Ge and described
The size of population of each region determines average peers amountIn one example, average peers amountIt can be by following
Formula is calculated:
Wherein, n indicates the number of region, PiIndicate the candidate peers amount of i-th of region, aiIndicate i-th of region
The size of population.
It in some embodiments, can be based on the candidate peers amount of the region the multiple region Zhong Ge and potential use
The quantity at family determines average peers amountIn one example, average peers amountIt can be calculated by the following formula
It obtains:
Wherein, n indicates the number of region, PiIndicate the candidate peers amount of i-th of region, ai' indicate i-th of region
The quantity of middle potential user.
Later, in step 110, when the average peers amount is greater than preset threshold, the multiple region is exported
Candidate peers amount as distributed data.
Preset threshold can be determined according to constraint condition (such as cost of product etc.).When obtaining average peers amount
When, it can be determined that whether the average peers amount is greater than preset threshold.If it is greater than preset threshold, illustrate the average equity money
Source amount meets constraint condition.The candidate peers amount that each region can be exported, using as the pre-set product eachly
The peers amount in domain.
In some embodiments, the method that this specification embodiment provides further include: step 112, when the average equity
When stock number is less than or equal to preset threshold, optimize average peers amount.With reference to Fig. 4, step 112 specifically includes following sub-step
It is rapid:
Step 1120, global optimum's slope is determined from the corresponding multiple optimal slopes in the multiple region, and will be global
Optimal slope corresponds to region as region to be optimized, by the corresponding candidate peers amount of global optimum's slope described in the region
As peers amount to be optimized.
When average peers amount is not more than preset threshold, illustrates that the average peers amount is unsatisfactory for constraint condition, need
Improve average peers amount.At this point, global optimum's slope is corresponded to the candidate peers amount of region as to be optimized right
Etc. stock numbers, preferentially optimize, can be while improving average peers amount, so that ensemble average conversion ratio reduced
It is slower.
Step 1122, on the corresponding broken line in the region to be optimized, from the peers amount coordinate points to be optimized and
Higher than between the peers amount coordinate points of the peers amount to be optimized in slope, determining new optimal slope, and will be described
Newly new candidate peers amount of the corresponding larger peers amount of optimal slope as the region to be optimized.
Specifically can be with reference to above to the introduction of step 106, details are not described herein.
Step 1124, candidate's equity based on the region except the new candidate peers amount and the region to be optimized
Stock number updates average peers amount.
Specifically can be with reference to above to the introduction of step 108, details are not described herein.
When updated average peers amount is still less than or equal to the preset threshold, step 1120- step is repeated
Rapid 1124, until updated average peers amount is greater than the preset threshold.
When updated average peers amount is greater than the preset threshold, step 110 is executed, output is the multiplely
The candidate peers amount in domain is as the distributed data.
The scheme that this specification embodiment provides can be used calculating equipment and pass through the corresponding user's letter in multiple regions
Breath determines that each region, compared with candidate peers amount slow, in the very fast situation of peers amount speedup, then is tied in conversion ratio reduction of speed
It closes preset threshold and judges whether the average peers amount of the candidate peers amount of each region meets constraint condition, so as to
To determine not only to meet constraint condition, but also the product peers amount of each region with higher conversion;Also, this illustrates reality
The scheme for applying example offer is a kind of consuming, high-efficient that computer information processing scheme can avoid a large amount of manpower and material resources.
On the other hand, the embodiment of this specification provides a kind of determining product peers amount in the distribution of multiple regions
The device 500 of data.Refering to Fig. 5, device 500 includes:
Acquiring unit 510 is configured to obtain data acquisition system corresponding with pre-set product multiple regions to be put,
In include the first data acquisition system corresponding with the first region, first data acquisition system includes multiple potential use in the first region
The user information at family, the user information of each potential user include each potential user under multiple peers amounts respectively with
The exchange probability that the pre-set product swaps;
First determination unit 520 is configured in the first data acquisition system each potential user under the first peers amount
Corresponding exchange probability determines first region corresponding conversion ratio under the first peers amount;First peers amount
For any peers amount in the multiple peers amount;
Construction unit 530 is configured to the multiple peers amount and first region the multiple right
Etc. corresponding conversion ratio under stock numbers, construct the first broken line;
Second determination unit 540, is configured to from first broken line minimum peers amount coordinate points and other are larger
Optimal slope is determined between peers amount coordinate points in slope, and using the corresponding larger peers amount of optimal slope as institute
State the candidate peers amount of the first region;
Third determination unit 550 is configured to the candidate peers amount of the multiple region, determines average equity money
Source amount;
Output unit 560 is configured to export the multiple region when the average peers amount is greater than preset threshold
Candidate peers amount as the distributed data.
In some embodiments, each potential user under multiple peers amounts respectively with the pre-set product into
The exchange probability of row exchange is that the user characteristics of the potential user and each peers amount input to exchange probability pre- respectively
Model is surveyed to be predicted to obtain;
The exchange Probabilistic Prediction Model is trained to obtain based on the historical behavior information to multiple sample of users;It is described
The historical behavior information of sample of users includes user characteristics, history peers amount, switch label.
In some embodiments, the user characteristics include, and user property feature and use the default production with user
The related user's scene characteristic of the scene of product.
In some embodiments, the abscissa of first broken line corresponds to the multiple peers amount, ordinate pair
The corresponding conversion ratio under the multiple peers amount of first region described in Ying Yu, the optimal slope are described first
Minimum slope on broken line between minimum peers amount coordinate points and other larger peers amount coordinate points in slope.
In some embodiments, the abscissa of first broken line corresponds to first region in the multiple equity money
Corresponding conversion ratio under the amount of source, ordinate correspond to the multiple peers amount, and the optimal slope is described first
Greatest gradient on broken line between minimum peers amount coordinate points and other larger peers amount coordinate points in slope.
In some embodiments, the third determination unit 550 is additionally configured to: based on the region the multiple region Zhong Ge
Candidate peers amount and the size of population of each region determine average peers amount.
In some embodiments, the third determination unit 550 is additionally configured to: based on the region the multiple region Zhong Ge
Candidate peers amount and the quantity of potential user determine average peers amount.
In some embodiments, described device further includes optimization unit (not shown), is configured that when average peers amount
When less than or equal to the preset threshold, global optimum's slope is determined from the corresponding multiple optimal slopes in the multiple region,
And global optimum's slope is corresponded into region as region to be optimized, the corresponding candidate of global optimum described in region slope is right
Etc. stock numbers as peers amount to be optimized;
On the corresponding broken line in the region to be optimized, from the peers amount coordinate points to be optimized and be higher than it is described to
Optimize between the peers amount coordinate points of peers amount in slope, determines new optimal slope, and will the newly optimal slope
New candidate peers amount of the corresponding larger peers amount as the region to be optimized;
Based on the candidate peers amount of the region except the new candidate peers amount and the region to be optimized, more
Newly be averaged peers amount;
Above step is repeated, until updated average peers amount is greater than preset threshold.
Each functional unit of device 500 can realize that details are not described herein with reference to embodiment of the method shown in FIG. 1.
On the other hand, the embodiment of this specification provides a kind of computer readable storage medium, is stored thereon with calculating
Machine program enables computer execute method shown in FIG. 1 when the computer program executes in a computer.
On the other hand, the embodiment of this specification provides a kind of calculating equipment, including memory and processor, described to deposit
It is stored with executable code in reservoir and realizes method shown in FIG. 1 when the processor executes the executable code.
Those skilled in the art are it will be appreciated that in said one or multiple examples, function described in the invention
It can be realized with hardware, software, firmware or their any combination.It when implemented in software, can be by these functions
Storage in computer-readable medium or as on computer-readable medium one or more instructions or code transmitted.
Above-described specific embodiment has carried out further the purpose of the present invention, technical scheme and beneficial effects
It is described in detail, it should be understood that being not intended to limit the present invention the foregoing is merely a specific embodiment of the invention
Protection scope, all any modification, equivalent substitution, improvement and etc. on the basis of technical solution of the present invention, done should all
Including within protection scope of the present invention.