CN109102324A

CN109102324A - Model training method, the red packet material based on model are laid with prediction technique and device

Info

Publication number: CN109102324A
Application number: CN201810764095.2A
Authority: CN
Inventors: 聂茜倩
Original assignee: Alibaba Group Holding Ltd
Current assignee: Advanced New Technologies Co Ltd; Advantageous New Technologies Co Ltd
Priority date: 2018-07-12
Filing date: 2018-07-12
Publication date: 2018-12-28
Anticipated expiration: 2038-07-12
Also published as: CN109102324B

Abstract

Disclose a kind of model training method, the red packet material based on model is laid with prediction technique and device, which includes: to obtain at least one barcode scanning neck red packet data and at least one barcode scanning payment data；For any setting businessman at least one setting businessman, according to the relevant information of the setting businessman, feature extraction is carried out to each data got respectively, obtains the corresponding training sample of the setting businessman, wherein, the training sample is using the characteristic value extracted as input value；It is trained using unsupervised learning algorithm and training sample obtained, obtain prediction model, the prediction model is equipped with the probability of red packet material as output valve to set businessman using the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data as input value.

Description

Model training method, the red packet material based on model are laid with prediction technique and device

Technical field

This specification embodiment is related to technical field of data processing more particularly to a kind of model training method, based on model Red packet material be laid with prediction technique and device.

Background technique

With the development of internet technology and intelligent terminal it is universal, the competitions between a variety of payment class application programs are also beneficial Fiercer, in order to fight for the possession share of " paying under line ", the Marketing group of certain payment class application programs proposes " barcode scanning neck The marketing strategy of red packet ", specifically, being printed on publicity material (the following letter of red packet two dimensional code by being laid in cooperation businessman shop Claim red packet material), such as post stand, brochure etc. can should by the application scan installed on mobile phone after user to shop Red packet two dimensional code obtains the red packet with certain amount of money, when pay under line by the application program next time, may be used The certain amount of money of red packet credit attracts more and more users to complete to pay under line using the application program with this.

Since region is extensive, cooperation businessman is numerous, and the Marketing group for paying class application program will usually be laid with red packet material Task meet at special extension worker be responsible for, in the related technology, in order to supervise the authenticity of red packet material laying condition, to mention The validity of high " barcode scanning neck red packet " this marketing strategy, can be by the way of manually arriving shop inspection, or passes through businessman Or extension worker takes pictures, the mode of recorded video.However, a large amount of manpower and material resources cost will be expended according to former mode； According to latter approach, then since number of pictures is limited, and the authenticity of photo needs to be queried, and causes to supervise the accurate of result Property is unable to get guarantee.

Summary of the invention

In view of the above technical problems, this specification embodiment provides a kind of model training method, the red packet object based on model Material is laid with prediction technique and device, technical solution are as follows:

According to this specification embodiment in a first aspect, providing a kind of model training method, which comprises

Obtain at least one barcode scanning neck red packet data and at least one barcode scanning payment data；

For any setting businessman at least one setting businessman, according to the relevant information of the setting businessman, respectively Feature extraction is carried out to each data got, obtains the corresponding training sample of the setting businessman, wherein the training Sample is using the characteristic value extracted as input value；

It is trained using unsupervised learning algorithm and training sample obtained, obtains prediction model, the prediction mould Type is equipped with red packet using the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data as input value to set businessman The probability of material is output valve.

According to the second aspect of this specification embodiment, a kind of red packet material laying prediction technique based on model is provided, The described method includes:

Feature extraction is carried out to accessed barcode scanning neck red packet data and barcode scanning payment data respectively, obtains the barcode scanning Lead the characteristic value of red packet data and the characteristic value of barcode scanning payment data；

By characteristic value input prediction model obtained, corresponding output valve is obtained, the prediction model is led red with barcode scanning The characteristic value of the characteristic value of bag data and barcode scanning payment data is input value, with set businessman be equipped with the probability of red packet material as Output valve.

According to the third aspect of this specification embodiment, a kind of model training apparatus is provided, described device includes:

First obtains module, for obtaining at least one barcode scanning neck red packet data and at least one barcode scanning payment data；

First extraction module, for setting any setting businessman in businessman at least one, according to the setting quotient The relevant information of family carries out feature extraction to each data got respectively, obtains the corresponding training of the setting businessman Sample, wherein the training sample is using the characteristic value extracted as input value；

Training module obtains prediction mould for being trained using unsupervised learning algorithm and training sample obtained Type, the prediction model is using the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data as input value, with setting The probability that businessman is equipped with red packet material is output valve.

According to the fourth aspect of this specification embodiment, a kind of red packet material laying prediction meanss based on model are provided, Described device includes:

Second obtains module, for obtaining at least one barcode scanning neck red packet data and at least one barcode scanning payment data；

Second extraction module, for carrying out feature respectively to accessed barcode scanning neck red packet data and barcode scanning payment data It extracts, obtains the characteristic value of the barcode scanning neck red packet data and the characteristic value of barcode scanning payment data；

Input module, for obtaining corresponding output valve, the prediction mould for characteristic value input prediction model obtained Type is equipped with red packet using the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data as input value to set businessman The probability of material is output valve.

According to the 5th of this specification embodiment aspect, a kind of computer equipment is provided, including memory, processor and deposit Store up the computer program that can be run on a memory and on a processor, wherein the processor is realized when executing described program The model training method that this specification embodiment provides.

According to the 6th of this specification embodiment aspect, a kind of computer equipment is provided, including memory, processor and deposit Store up the computer program that can be run on a memory and on a processor, wherein the processor is realized when executing described program The red packet material based on model that this specification embodiment provides is laid with prediction technique.

Technical solution provided by this specification embodiment, by obtaining at least one scanning neck red packet data and at least one Barcode scanning payment data, for any setting businessman at least one setting businessman, according to the relevant information of setting businessman, point It is other that feature extraction is carried out to each data for getting, obtain the corresponding training sample of setting businessman, wherein training sample with The characteristic value extracted is input value, is trained, is predicted using unsupervised learning algorithm and training sample obtained Model, the prediction model is using the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data as input value, with setting The probability that businessman is equipped with red packet material is output valve, subsequent, by the prediction model can red packet material to businessman be laid with Situation predicted, is handled by this kind, and can use manpower and material resources sparingly cost, while carrying out prediction based on accurate data can be with Accurate prediction result is obtained, and can be without geographical restrictions with man's activity, realization more comprehensively predicts each Set the red packet material laying condition of businessman.

It should be understood that above general description and following detailed description be only it is exemplary and explanatory, not This specification embodiment can be limited.

In addition, any embodiment in this specification embodiment does not need to reach above-mentioned whole effects.

Detailed description of the invention

In order to illustrate more clearly of this specification embodiment or technical solution in the prior art, below will to embodiment or Attached drawing needed to be used in the description of the prior art is briefly described, it should be apparent that, the accompanying drawings in the following description is only The some embodiments recorded in this specification embodiment for those of ordinary skill in the art can also be attached according to these Figure obtains other attached drawings.

Fig. 1 is the embodiment flow chart of the model training method shown in one exemplary embodiment of this specification；

Fig. 2 is the implementation of the red packet material laying prediction technique based on model shown in one exemplary embodiment of this specification Example flow chart；

Fig. 3 is a kind of embodiment block diagram for model training apparatus that one exemplary embodiment of this specification provides；

Fig. 4 is that a kind of red packet material based on model that one exemplary embodiment of this specification provides is laid with prediction meanss Embodiment block diagram；

Fig. 5 is that one kind provided by this specification embodiment more specifically calculates device hardware structural schematic diagram.

Specific embodiment

In order to make those skilled in the art more fully understand the technical solution in this specification embodiment, below in conjunction with this Attached drawing in specification embodiment is described in detail the technical solution in this specification embodiment, it is clear that described Embodiment is only a part of the embodiment of this specification, instead of all the embodiments.The embodiment of base in this manual, Those of ordinary skill in the art's every other embodiment obtained, all should belong to the range of protection.

To solve the above-mentioned problems, this specification embodiment provides a kind of model training method and the red packet object based on model Material is laid with prediction technique, as follows, and the two methods are described in detail respectively.

It is illustrated firstly, showing the model training method that following embodiment a pair of this specification embodiments provide:

Embodiment one:

It referring to Figure 1, is the embodiment flow chart of the model training method shown in one exemplary embodiment of this specification, it should Method may comprise steps of:

Step 102: obtaining at least one barcode scanning neck red packet data and at least one barcode scanning payment data.

In this specification embodiment, user is got the relevant data of red packet this behavior and be known as sweeping by scanning the two-dimensional code User pay the relevant data of this behavior and is known as barcode scanning payment data by code neck red packet data by scanning the two-dimensional code, Every generation single pass two dimensional code gets red packet behavior as a result, will generate a barcode scanning neck red packet data, every generation single pass Two dimensional code carries out payment behavior, will generate a barcode scanning payment data.

In one embodiment, above-mentioned barcode scanning neck red packet data at least may include: that red packet identifies, red packet gets position, red Packet gets moment, the red packet amount of money etc..

In one embodiment, above-mentioned barcode scanning payment data can be divided into two classes, and one type is that the barcode scanning branch of red packet is not used Data are paid, at least may include: Receiving information, location for payment, payment moment, payment amount etc.；It is another kind of to have for use The barcode scanning payment data of red packet at least may include: Receiving information, location for payment, payment moment, payment amount, red packet gold Volume, red packet mark etc..

It should be noted that above-mentioned red packet mark consists of two parts, respectively red packet code in this specification embodiment Value is numbered with red packet, wherein red packet code value can be described as the code value of two dimensional code again, that is to say, that if different user or same use The red packet red packet code value having the same that family is got in different time, by scanning same two dimensional code, and have different Red packet number, with the red packet different by different red packet number-marks；And same user or different user are by scanning not The red packet that same two dimensional code is got then has different red packet code values.

In this specification embodiment, available set period, such as June 21 0:00 to June 23 0:00 this The barcode scanning neck red packet data and barcode scanning payment data generated in period.

Step 104: for any setting businessman at least one setting businessman, being believed according to the correlation of setting businessman Breath carries out feature extraction to each data got respectively, obtains the corresponding training sample of setting businessman, wherein instruction Practice sample using the characteristic value extracted as input value.

In this specification embodiment, for convenience, there is cooperation for signing in advance with the Marketing group of application program The businessman of agreement is known as setting businessman, those settings businessman can be distributed in the different location in same city, or be distributed across Multiple cities, this specification embodiment to this with no restriction.

In this specification embodiment, it can be directed to each setting businessman, according to the relevant information of setting businessman, respectively To each data got in above-mentioned steps 102, i.e. each barcode scanning payment data and barcode scanning neck red packet data carries out special Sign is extracted, and using the characteristic value extracted as input value, obtains the corresponding training sample of setting businessman.

In this specification embodiment, two kinds of characteristic value can be set, one type is spy relevant to position Value indicative, another kind of is to check and write off the relevant characteristic value of situation to red packet, is based on the two kinds of characteristic value, above-mentioned setting businessman's Relevant information at least may include: the location information for setting businessman, the Receiving information for setting businessman.

(1) it is directed to characteristic value relevant to position, the process for extracting characteristic value is described:

It is possible, firstly, to will be in the location information and each barcode scanning neck red packet data that set businessman using GeoHash algorithm Red packet get position and matched, determine in the specified range of setting businessman, such as got red packet in 1 kilometer range Total number, seen from the above description, every generation single pass two dimensional code get red packet behavior, will generate a barcode scanning neck red packet number According to namely red packet of every generation get event, then generate barcode scanning and lead red packet data, then, it is above-mentioned setting businessman's The total number that red packet is got in specified range is then in the specified range of setting businessman, and the red packet of generation gets time of event Several namely red packet gets the quantity that position is located at the barcode scanning neck red packet data of the specified range.

Similar, can use GeoHash algorithm will be in the location information and each barcode scanning payment data that set businessman Location for payment matched, determine that total stroke count of payment transaction namely location for payment are specified positioned at this in the specified range The quantity of the barcode scanning payment data of range.

It is subsequent, then it can further calculate out the ratio between total stroke count of the total number and payment transaction of being got red packet Value, gets accounting for the ratio as the red packet in specified range.

Secondly, can set the location information of businessman with other according to the location information of setting businessman, determine to set at this Determine in the specified range of businessman, businessman's quantity of possessed other settings businessman, it will be appreciated by persons skilled in the art that Businessman's concentration that businessman's quantity can reflect out in the specified range for setting businessman in turn can be further The ratio between the total number for being got red packet and businessman's quantity is calculated, using the ratio as the red packet in the specified range Get closeness.

It, then can be by the above-mentioned total number for being got red packet, total stroke count, red of payment transaction in this specification embodiment Packet gets accounting, businessman's quantity and red packet and gets closeness as the characteristic value extracted.

In addition, the characteristic value for multiple specified ranges can also be got in this specification embodiment, for example, also The available red packet within the scope of 10 meters, 30 meters, 100 meters, 300 meters or 500 meters of setting businessman gets accounting, this explanation Book embodiment to this with no restriction.

(2) the relevant characteristic value of situation is checked and write off for red packet, the process for extracting characteristic value is described:

In this specification embodiment, for one setting businessman, can determine first consumed in setting businessman it is red Packet, specifically, the Receiving information in the Receiving information of setting businessman and each barcode scanning payment data can be matched, The consistent barcode scanning payment data of Receiving information for getting Receiving information and setting businessman will acquire for convenience Barcode scanning payment data be known as the corresponding barcode scanning payment data of setting businessman；Further, from the corresponding barcode scanning of setting businessman It is filtered out in payment data using the barcode scanning payment data for having red packet, that is, includes the barcode scanning payment data of red packet mark, in order to retouch It states conveniently, the barcode scanning payment data including red packet mark filtered out is determined as target barcode scanning payment data, so far, then really It defines and carries out payment transaction in setting businessman, and use the target barcode scanning payment data for having red packet；It is subsequent, then it can basis The target barcode scanning payment data, which is extracted, checks and writes off the relevant characteristic value of situation to red packet, may include: red packet transaction accounting, red packet The transaction for checking and writing off the contribution degree, best red packet code value of checking and writing off average time interval, best red packet code value apart from dispersion, red packet accounts for Than etc..

Wherein, i: red packet transaction accounting can be extracted by following procedure:

Can calculate the quantity of target barcode scanning payment data barcode scanning payment data corresponding with setting businessman data it Between ratio, the ratio can become setting businessman red packet trade accounting.

Ii: red packet, which is checked and write off, to be checked and write off average time interval apart from dispersion, red packet and can be extracted by following procedure:

For any bar target barcode scanning payment data, first can by the red packet in the target barcode scanning payment data identify with Red packet mark in each barcode scanning neck red packet data is matched, and will be identified with target barcode scanning payment data with identical red packet Barcode scanning neck red packet data lead red packet data as its corresponding barcode scanning.

Further, it is possible to be led in red packet data according to the corresponding barcode scanning of the payment moment in target barcode scanning payment data Red packet get constantly determine red packet check and write off time interval, it will be appreciated by persons skilled in the art that the red packet checks and writes off the time Interval then reflects red packet from getting to consumption every duration；Correspondingly, can be according to the payment in target barcode scanning payment data Red packet in the corresponding barcode scanning neck red packet data in position gets position and determines that red packet checks and writes off distance interval, those skilled in the art Member it is understood that the red packet check and write off distance interval then reflect red packet get position and consumption position institute's gauge from.

Seen from the above description, for each target barcode scanning payment data, a red packet can be calculated and check and write off the time Distance interval is checked and write off with red packet in interval, subsequent, in this specification embodiment, then can calculate all red packets and check and write off between the time Every average value, check and write off average time interval using the average value as red packet；Correspondingly, can calculate all red packets check and write off away from Sow discord every variance, check and write off using the variance as red packet apart from dispersion.

Iii: contribution degree, the transaction accounting of best red packet code value of best red packet code value can be extracted by following procedure:

Seen from the above description, the red packet red packet code value having the same got by scanning same two dimensional code, that , the red packet of red packet code value having the same checks and writes off situation then and can reflect the laying condition of its corresponding two dimensional code, That is the laying condition of red packet material, for example, the red packet of red packet code value having the same to check and write off situation better, then can show The paving location of its corresponding two dimensional code is better, and user's receiving degree is better.

Based on this, in this specification embodiment, can be identified according to red packet in red packet code value to target barcode scanning pay Data are grouped, wherein the red packet code value that red packet identifies in the target barcode scanning payment data of each grouping is identical, different grouping Target barcode scanning payment data in red packet mark red packet code value it is not identical, that is, the target barcode scanning in the same grouping is paid Data are related to the same two dimensional code, and a grouping then corresponds to a red packet code value.

Later, the sequence according to the quantity of target barcode scanning payment data in each grouping from high to low, it is corresponding to each grouping Red packet code value is ranked up, and the red packet code value of the red packet code value to rank first and position of being number two is determined according to ranking results, In this specification embodiment, for convenience, the red packet code value to rank first is known as the first red packet code value, will be number two The red packet code value of position is known as the second red packet code value, it will be appreciated by persons skilled in the art that in the corresponding grouping of red packet code value The quantity of target barcode scanning payment data is more, then it represents that the red packet with the red packet code value to check and write off situation better, be based on this, this In specification embodiment, the first red packet code value that this is ranked first is determined as best red packet code value.

It is subsequent, it can determine the quantity of the target barcode scanning payment data with the first red packet code value, respectively for the side of description Just, which is known as the first quantity, and determines the quantity with the target barcode scanning payment data of the second red packet code value, in order to retouch It states conveniently, which is known as the second quantity.

It is subsequent, by the ratio between the quantity of the first quantity barcode scanning payment data corresponding with setting businessman, as most The transaction accounting of good red packet code value, and using the ratio between the first quantity and the second quantity as the contribution of best red packet code value Degree.

Based on the description in step 104, the characteristic value in this specification embodiment can be as described in Table 1:

Table 1

Step 106: it is trained using unsupervised learning algorithm and training sample obtained, obtains prediction model, it should Prediction model is using the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data as input value, to set businessman's laying The probability for having red packet material is output valve.

By the associated description of above-mentioned steps it is found that in this specification embodiment, acquired training sample only has defeated Enter value, have no label value, therefore, in this specification embodiment, unsupervised learning algorithm, such as PCA can be used (Principal Component Analysis, principal component analysis) algorithm is trained training sample obtained, obtains Prediction model, the prediction model using barcode scanning neck red packet data characteristic value and barcode scanning payment data characteristic value as input value, with It sets businessman and is equipped with the probability of red packet material as output valve.

In addition, the description in 104 is it is found that be directed to each setting businessman, available one and red packet through the above steps The relevant sequence of code value, then, in this specification embodiment, it can also predict what setting businessman was laid with according to the sequence The two dimensional code code value of red packet material is spread for example, best red packet code value described in above-mentioned steps 104 is predicted as setting businessman If red packet material two dimensional code code value, as a result, the output valve of above-mentioned prediction model can also include setting businessman be laid with it is red The two dimensional code code value of packet material.

So far, the associated description of embodiment one is completed.

Prediction is laid with to the red packet material based on model that this specification embodiment provides secondly, showing following embodiments two Method is illustrated:

Embodiment two:

Fig. 2 is referred to, is that the red packet material based on model shown in one exemplary embodiment of this specification is laid with prediction side The embodiment flow chart of method, this method may comprise steps of:

Step 202: obtaining at least one barcode scanning neck red packet data and at least one barcode scanning payment data.

Step 204: feature extraction being carried out to accessed barcode scanning neck red packet data and barcode scanning payment data respectively, is obtained Barcode scanning leads the characteristic value of red packet data and the characteristic value of barcode scanning payment data.

The detailed description of step 202 and step 204 may refer to the associated description of above-described embodiment one kind, and this specification is real It applies example and this is no longer described in detail.

Step 206: by characteristic value input prediction model obtained, obtaining corresponding output valve, the prediction model is to sweep The characteristic value of code neck red packet data and the characteristic value of barcode scanning payment data are input value, are equipped with red packet material to set businessman Probability is output valve.

So far, the associated description of embodiment two is completed.

Technical solution provided by this specification embodiment leads red packet data and at least one by obtaining at least one barcode scanning Barcode scanning payment data carries out feature extraction to accessed barcode scanning neck red packet data and barcode scanning payment data respectively, obtains The characteristic value of barcode scanning neck red packet data and the characteristic value of barcode scanning payment data obtain characteristic value input prediction model obtained To corresponding output valve, which is input with the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data Value, to set probability that businessman is equipped with red packet material as output valve, may be implemented can be to the red of businessman by prediction model Packet material laying condition is predicted that use manpower and material resources sparingly cost, at the same based on accurate data predict it is available compared with For accurate prediction result, and can be without geographical restrictions with man's activity, realization more comprehensively predicts each setting quotient The red packet material laying condition of family.

Corresponding to above-mentioned model training method embodiment, this specification embodiment also provides a kind of model training apparatus, ginseng As shown in Figure 3, the apparatus may include: first obtain module 310, the first extraction module 320, training module 330.

Wherein, first module 310 is obtained, can be used for obtaining at least one barcode scanning neck red packet data and at least one barcode scanning Payment data；

First extraction module 320 can be used for for any setting businessman at least one setting businessman, according to described The relevant information of businessman is set, feature extraction is carried out to each data got respectively, it is corresponding to obtain the setting businessman Training sample, wherein the training sample is using the characteristic value extracted as input value；

Training module 330 can be used for being trained using unsupervised learning algorithm and training sample obtained, obtain Prediction model, the prediction model using barcode scanning neck red packet data characteristic value and barcode scanning payment data characteristic value as input value, The probability of red packet material is equipped with as output valve to set businessman.

In one embodiment, the relevant information of the setting businessman includes at least: the location information of the setting businessman, institute State the Receiving information of setting businessman；

The barcode scanning neck red packet data include at least: red packet mark, red packet gets position, red packet is got the moment, described red Packet mark includes that red packet code value and red packet are numbered, wherein the red packet red packet code having the same got by same two dimensional code Value is numbered with different red packets, and has different red packet code values by the red packet that different two dimensional codes are got；

The barcode scanning payment data includes at least: Receiving information, location for payment, payment moment；Or,

Receiving information, location for payment, payment moment, red packet mark.

In one embodiment, first extraction module 320 may include (being not shown in Fig. 3):

First matched sub-block, for will it is described setting businessman location information and each barcode scanning neck red packet data in Red packet is got position and is matched, and determines the total number that red packet is got in the specified range of the setting businessman；

Second matched sub-block, for by it is described setting businessman location information and each barcode scanning payment data in branch It pays position to be matched, determines total stroke count of the payment transaction in the specified range；

First determines submodule, for will be between the total number that be got red packet and total stroke count of the payment transaction Ratio, get accounting as the red packet in the specified range；

Second determines submodule, for being believed according to the location information of the setting businessman and the position of other settings businessman Breath, determines businessman's quantity in the specified range, and by the total number for being got red packet and businessman's quantity it Between ratio get closeness as the red packet in the specified range；

Fisrt feature determines submodule, for by total stroke count of the total number for being got red packet, the payment transaction, The red packet gets accounting, businessman's quantity and the red packet and gets closeness as the characteristic value extracted.

Third matched sub-block, for by it is described setting businessman Receiving information and each barcode scanning payment data in receipts Money information is matched, and Receiving information and the consistent barcode scanning payment data of Receiving information of the setting businessman are set as described in Determine the corresponding barcode scanning payment data of businessman；

Third determines submodule, for that in the corresponding barcode scanning payment data of the setting businessman, will include that red packet identifies Barcode scanning payment data be determined as target barcode scanning payment data；

4th determines submodule, for determining that red packet transaction accounting, red packet are checked and write off according to the target barcode scanning payment data Average time interval, the contribution degree of best red packet code value, the transaction accounting of best red packet code value are checked and write off apart from dispersion, red packet；

Second feature determines submodule, for checking and writing off trade accounting, the red packet of the red packet apart from dispersion, described Red packet checks and writes off the transaction accounting conduct of the contribution degree, the best red packet code value of average time interval, the best red packet code value The characteristic value extracted.

In one embodiment, the described 4th determine that submodule can be specifically used for:

By the quantity of the target barcode scanning payment data and the quantity of the corresponding barcode scanning payment data of the setting businessman it Between ratio, as it is described setting businessman red packet trade accounting.

In one embodiment, the described 4th determine that submodule may include (being not shown in Fig. 3):

4th matched sub-block, for being directed to any bar target barcode scanning payment data, by the target barcode scanning payment data In red packet mark and each barcode scanning neck red packet data in red packet mark matched, will with the target barcode scanning pay number According to the barcode scanning neck red packet data identified with identical red packet as the corresponding barcode scanning neck red packet number of the target barcode scanning payment data According to；

5th determines submodule, for the barcode scanning corresponding according to the payment moment in the target barcode scanning payment data Red packet in neck red packet data is got determines that red packet checks and writes off time interval constantly；

6th determines submodule, for the barcode scanning corresponding according to the location for payment in the target barcode scanning payment data Red packet in neck red packet data gets position and determines that red packet checks and writes off distance interval；

First computational submodule checks and writes off the average value of time interval for calculating identified all red packets, will be described Average value checks and writes off average time interval as red packet；

Second computational submodule checks and writes off the variance of distance interval for calculating identified all red packets, by the side Difference is checked and write off as red packet apart from dispersion.

It is grouped submodule, the target barcode scanning payment data is divided for the red packet code value in being identified according to red packet Group, wherein the red packet code value in the target barcode scanning payment data of each grouping is identical, the target barcode scanning payment data of different grouping In red packet code value it is not identical；

Sorting sub-module, for the sequence according to the quantity of target barcode scanning payment data in each grouping from high to low, to each It is grouped corresponding red packet code value to be ranked up, determines the first red packet code value to rank first according to ranking results and position of being number two The second red packet code value；

Quantity determines submodule, for determining there is the of the target barcode scanning payment data of the first red packet code value respectively One quantity, the second quantity with the target barcode scanning payment data with the second red packet code value；

7th determines submodule, for by first quantity and the number for setting the corresponding barcode scanning payment data of businessman Ratio between amount, as the transaction accounting of best red packet code value, and will be between first quantity and second quantity Contribution degree of the ratio as best red packet code value, wherein the best red packet code value is the first red packet code value.

In one embodiment, the unsupervised learning algorithm includes:

Principal component analysis PCA algorithm.

In one embodiment, the output valve of the prediction model further include:

Set the red packet code value for the red packet material that businessman is laid with.

It is understood that first obtains module 310, the first extraction module 320 and training module 330 as three kinds The module of functional independence, both can as shown in Figure 3 simultaneously configuration in a device, can also individually configure in a device, because This structure shown in Fig. 3 should not be construed as the restriction to this specification example scheme.

In addition, the function of modules and the realization process of effect are specifically detailed in the above method corresponding step in above-mentioned apparatus Rapid realization process, details are not described herein.

It is laid with prediction technique embodiment corresponding to the above-mentioned red packet material based on model, this specification embodiment also provides one Red packet material of the kind based on model is laid with prediction meanss, shown in Figure 4, the apparatus may include: the second acquisition module 410, Second extraction module 420, input module 430.

Wherein, second module 410 is obtained, can be used for obtaining at least one barcode scanning neck red packet data and at least one barcode scanning Payment data；

Second extraction module 420 can be used for accessed barcode scanning neck red packet data and barcode scanning payment data difference Feature extraction is carried out, the characteristic value of the barcode scanning neck red packet data and the characteristic value of barcode scanning payment data are obtained；

Input module 430 can be used for characteristic value input prediction model obtained obtaining corresponding output valve, institute Prediction model is stated using the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data as input value, to set businessman's paving Probability equipped with red packet material is output valve.

It is understood that second obtains module 410, the second extraction module 420 and input module 430 as three kinds The module of functional independence, both can as shown in Figure 4 simultaneously configuration in a device, can also individually configure in a device, because This structure shown in Fig. 4 should not be construed as the restriction to this specification example scheme.

This specification embodiment also provides a kind of computer equipment, includes at least memory, processor and is stored in On reservoir and the computer program that can run on a processor, wherein processor realizes model above-mentioned when executing described program Training method, this method include at least: obtaining at least one barcode scanning neck red packet data and at least one barcode scanning payment data；For Any setting businessman at least one setting businessman, it is every to what is got respectively according to the relevant information of the setting businessman One data carries out feature extraction, obtains the corresponding training sample of the setting businessman, wherein the training sample is to extract Characteristic value be input value；It is trained using unsupervised learning algorithm and training sample obtained, obtains prediction model, institute Prediction model is stated using the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data as input value, to set businessman's paving Probability equipped with red packet material is output valve.

This specification embodiment also provides another computer equipment, includes at least memory, processor and is stored in On memory and the computer program that can run on a processor, wherein processor realizes base above-mentioned when executing described program It is laid with prediction technique in the red packet material of model, this method includes at least: obtaining at least one barcode scanning neck red packet data and at least One barcode scanning payment data；Feature extraction is carried out to accessed barcode scanning neck red packet data and barcode scanning payment data respectively, is obtained Obtain the characteristic value of the barcode scanning neck red packet data and the characteristic value of barcode scanning payment data；By characteristic value input prediction mould obtained Type obtains corresponding output valve, and the prediction model is with the feature of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data Value is input value, is equipped with the probability of red packet material as output valve to set businessman.

Fig. 5 shows one kind provided by this specification embodiment and more specifically calculates device hardware structural schematic diagram, The equipment may include: processor 510, memory 520, input/output interface 530, communication interface 540 and bus 550.Wherein Processor 510, memory 520, input/output interface 530 and communication interface 540 between the realization of bus 550 by setting Standby internal communication connection.

Processor 510 can use general CPU (Central Processing Unit, central processing unit), micro process Device, application specific integrated circuit (Application Specific Integrated Circuit, ASIC) or one or The modes such as multiple integrated circuits are realized, for executing relative program, to realize technical solution provided by this specification embodiment.

Memory 520 can use ROM (Read Only Memory, read-only memory), RAM (Random Access Memory, random access memory), static storage device, the forms such as dynamic memory realize.Memory 520 can store Operating system and other applications are realizing technical solution provided by this specification embodiment by software or firmware When, relevant program code is stored in memory 520, and execution is called by processor 510.

Input/output interface 530 is for connecting input/output module, to realize information input and output.Input and output/ Module can be used as component Configuration and (be not shown in Fig. 5) in a device, can also be external in equipment to provide corresponding function.Wherein Input equipment may include keyboard, mouse, touch screen, microphone, various kinds of sensors etc., output equipment may include display, Loudspeaker, vibrator, indicator light etc..

Communication interface 540 is used for connection communication module (being not shown in Fig. 5), to realize the communication of this equipment and other equipment Interaction.Wherein communication module can be realized by wired mode (such as USB, cable etc.) and be communicated, can also be wirelessly (such as mobile network, WIFI, bluetooth etc.) realizes communication.

Bus 550 includes an access, in various components (such as the processor 510, memory 520, input/output of equipment Interface 530 and communication interface 540) between transmit information.

It should be noted that although above equipment illustrates only processor 510, memory 520, input/output interface 530, communication interface 540 and bus 550, but in the specific implementation process, which can also include realizing to operate normally Necessary other assemblies.In addition, it will be appreciated by those skilled in the art that, it can also be only comprising realizing in above equipment Component necessary to this specification example scheme, without including all components shown in figure.

This specification embodiment also provides a kind of computer readable storage medium, is stored thereon with computer program, the journey Model training method above-mentioned is realized when sequence is executed by processor, this method includes at least: obtaining at least one barcode scanning neck red packet Data and at least one barcode scanning payment data；For any setting businessman at least one setting businessman, according to the setting The relevant information of businessman carries out feature extraction to each data got respectively, obtains the corresponding instruction of the setting businessman Practice sample, wherein the training sample is using the characteristic value extracted as input value；Using unsupervised learning algorithm with it is obtained Training sample is trained, and obtains prediction model, and the prediction model is paid with the characteristic value and barcode scanning of barcode scanning neck red packet data The characteristic value of data is input value, is equipped with the probability of red packet material as output valve to set businessman.

This specification embodiment also provides another computer readable storage medium, is stored thereon with computer program, should Realize that the red packet material above-mentioned based on model is laid with prediction technique when program is executed by processor, this method includes at least: obtaining Take at least one barcode scanning neck red packet data and at least one barcode scanning payment data；To accessed barcode scanning neck red packet data and sweep Code payment data carries out feature extraction respectively, obtains the characteristic value of the barcode scanning neck red packet data and the feature of barcode scanning payment data Value；By characteristic value input prediction model obtained, corresponding output valve is obtained, the prediction model leads red packet data with barcode scanning Characteristic value and barcode scanning payment data characteristic value be input value, with set businessman be equipped with red packet material probability be export Value.

Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data. The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM), Digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape magnetic disk storage or other magnetic storage devices Or any other non-transmission medium, can be used for storage can be accessed by a computing device information.As defined in this article, it calculates Machine readable medium does not include temporary computer readable media (transitory media), such as the data-signal and carrier wave of modulation.

As seen through the above description of the embodiments, those skilled in the art can be understood that this specification Embodiment can be realized by means of software and necessary general hardware platform.Based on this understanding, this specification is implemented Substantially the part that contributes to existing technology can be embodied in the form of software products the technical solution of example in other words, The computer software product can store in storage medium, such as ROM/RAM, magnetic disk, CD, including some instructions are to make It is each to obtain computer equipment (can be personal computer, server or the network equipment etc.) execution this specification embodiment Method described in certain parts of a embodiment or embodiment.

System, device, module or the unit that above-described embodiment illustrates can specifically realize by computer chip or entity, Or it is realized by the product with certain function.A kind of typically to realize that equipment is computer, the concrete form of computer can To be personal computer, laptop computer, cellular phone, camera phone, smart phone, personal digital assistant, media play In device, navigation equipment, E-mail receiver/send equipment, game console, tablet computer, wearable device or these equipment The combination of any several equipment.

All the embodiments in this specification are described in a progressive manner, same and similar portion between each embodiment Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.Especially for device reality For applying example, since it is substantially similar to the method embodiment, so describing fairly simple, related place is referring to embodiment of the method Part explanation.The apparatus embodiments described above are merely exemplary, wherein described be used as separate part description Module may or may not be physically separated, can be each module when implementing this specification example scheme Function realize in the same or multiple software and or hardware.Can also select according to the actual needs part therein or Person's whole module achieves the purpose of the solution of this embodiment.Those of ordinary skill in the art are not the case where making the creative labor Under, it can it understands and implements.

The above is only the specific embodiment of this specification embodiment, it is noted that for the general of the art For logical technical staff, under the premise of not departing from this specification embodiment principle, several improvements and modifications can also be made, this A little improvements and modifications also should be regarded as the protection scope of this specification embodiment.

Claims

1. a kind of model training method, which comprises

For any setting businessman at least one setting businessman, according to the relevant information of the setting businessman, respectively to obtaining The each data got carries out feature extraction, obtains the corresponding training sample of the setting businessman, wherein the training sample Using the characteristic value extracted as input value；

Be trained using unsupervised learning algorithm and training sample obtained, obtain prediction model, the prediction model with It is input value that barcode scanning, which leads the characteristic value of red packet data and the characteristic value of barcode scanning payment data, is equipped with red packet material to set businessman Probability be output valve.

2. according to the method described in claim 1, the relevant information of the setting businessman includes at least: the position of the setting businessman Confidence breath, the Receiving information for setting businessman；

The barcode scanning neck red packet data include at least: red packet mark, red packet gets position, red packet is got the moment, the red packet mark Know includes that red packet code value and red packet are numbered, wherein the red packet red packet code value having the same got by same two dimensional code, tool There is different red packet numbers, and there is different red packet code values by the red packet that different two dimensional codes are got；

Receiving information, location for payment, payment moment, red packet mark.

3. according to the method described in claim 2, it is described according to it is described setting businessman relevant information, respectively to each data Carry out feature extraction, comprising:

Red packet in the location information of the setting businessman and each barcode scanning neck red packet data is got position to match, really It is scheduled on the total number that red packet is got in the specified range of the setting businessman；

Location for payment in the location information and each barcode scanning payment data of the setting businessman is matched, is determined in institute State total stroke count of payment transaction in specified range；

By the ratio between the total number for being got red packet and total stroke count of the payment transaction, as the specified range Interior red packet gets accounting；

According to the location information of the location information of the setting businessman and other settings businessman, determine in the specified range Businessman's quantity, and using the ratio between the total number for being got red packet and businessman's quantity as in the specified range Red packet get closeness；

The total number for being got red packet, total stroke count of the payment transaction, the red packet are got into accounting, businessman's number Amount and the red packet get closeness as the characteristic value extracted.

4. according to the method described in claim 2, it is described according to it is described setting businessman relevant information, respectively to each data Carry out feature extraction, comprising:

Receiving information in the Receiving information and each barcode scanning payment data of the setting businessman is matched, gathering is believed Breath is with the consistent barcode scanning payment data of Receiving information of the setting businessman as the setting businessman corresponding barcode scanning payment number According to；

In the corresponding barcode scanning payment data of the setting businessman, the barcode scanning payment data including red packet mark is determined as target Barcode scanning payment data；

Determine that red packet transaction accounting, red packet are checked and write off and checked and write off averagely apart from dispersion, red packet according to the target barcode scanning payment data Time interval, the contribution degree of best red packet code value, the transaction accounting of best red packet code value；

By the red packet trade accounting, the red packet check and write off apart from dispersion, the red packet check and write off average time interval, it is described most The contribution degree of good red packet code value, the transaction accounting of the best red packet code value are as the characteristic value extracted.

5. according to the method described in claim 4, described determine red packet transaction accounting, packet according to the target barcode scanning payment data It includes:

It will be between the quantity of the quantity of target barcode scanning payment data barcode scanning payment data corresponding with the setting businessman Ratio, the red packet transaction accounting as the setting businessman.

6. according to the method described in claim 4, it is described according to the target barcode scanning payment data determine red packet check and write off distance from Divergence, red packet check and write off average time interval, comprising:

Following steps are executed for any bar target barcode scanning payment data:

Red packet in red packet mark and each barcode scanning neck red packet data in the target barcode scanning payment data is identified and is carried out Matching will have the barcode scanning neck red packet data of identical red packet mark as the target barcode scanning with the target barcode scanning payment data The corresponding barcode scanning of payment data leads red packet data；

According to the payment moment in the target barcode scanning payment data, corresponding barcode scanning leads the red packet in red packet data to get Moment determines that red packet checks and writes off time interval；

It is got according to the red packet in the corresponding barcode scanning neck red packet data of the location for payment in the target barcode scanning payment data Position determines that red packet checks and writes off distance interval；

All red packets determined by calculating check and write off the average value of time interval, check and write off mean time using the average value as red packet Between be spaced；

All red packets determined by calculating check and write off the variance of distance interval, check and write off using the variance as red packet apart from discrete Degree.

7. according to the method described in claim 4, described determine best red packet code value according to the target barcode scanning payment data The transaction accounting of contribution degree, best red packet code value, comprising:

According to red packet identify in red packet code value the target barcode scanning payment data is grouped, wherein the mesh of each grouping The red packet code value marked in barcode scanning payment data is identical, and the red packet code value in the target barcode scanning payment data of different grouping is not identical；

According to the sequence of the quantity of target barcode scanning payment data in each grouping from high to low, to the corresponding red packet code value of each grouping into Row sequence, the second red packet code value of the first red packet code value to rank first and position of being number two is determined according to ranking results；

First quantity with the target barcode scanning payment data of the first red packet code value is determined respectively, and it is red with described second Second quantity of the target barcode scanning payment data of packet code value；

By the ratio between the quantity of first quantity barcode scanning payment data corresponding with the setting businessman, as best red The transaction accounting of packet code value, and using the ratio between first quantity and second quantity as the tribute of best red packet code value Degree of offering, wherein the best red packet code value is the first red packet code value.

8. according to the method described in claim 1, the unsupervised learning algorithm includes:

Principal component analysis PCA algorithm.

9. according to the method described in claim 1, the output valve of the prediction model further include:

10. a kind of red packet material based on model is laid with prediction technique, which comprises

Feature extraction is carried out to accessed barcode scanning neck red packet data and barcode scanning payment data respectively, it is red to obtain the barcode scanning neck The characteristic value of bag data and the characteristic value of barcode scanning payment data；

By characteristic value input prediction model obtained, corresponding output valve is obtained, the prediction model leads red packet number with barcode scanning According to characteristic value and barcode scanning payment data characteristic value be input value, with set businessman be equipped with red packet material probability be export Value.

11. a kind of model training apparatus, described device include:

First extraction module, for setting businessman's according to described for any setting businessman at least one setting businessman Relevant information carries out feature extraction to each data got respectively, obtains the corresponding training sample of the setting businessman, Wherein, the training sample is using the characteristic value extracted as input value；

Training module obtains prediction model, institute for being trained using unsupervised learning algorithm and training sample obtained Prediction model is stated using the characteristic value of the characteristic value of barcode scanning neck red packet data and barcode scanning payment data as input value, to set businessman's paving Probability equipped with red packet material is output valve.

12. the relevant information of device according to claim 11, the setting businessman includes at least: the setting businessman's Location information, the Receiving information for setting businessman；

Receiving information, location for payment, payment moment, red packet mark.

13. device according to claim 12, first extraction module include:

First matched sub-block, for the location information of the setting businessman and each barcode scanning to be led the red packet in red packet data It gets position to be matched, determines the total number for being got red packet in the specified range of the setting businessman；

Second matched sub-block, for by it is described setting businessman location information and each barcode scanning payment data in payment position It sets and is matched, determine total stroke count of the payment transaction in the specified range；

First determines submodule, for by the ratio between the total number for being got red packet and total stroke count of the payment transaction Value, gets accounting as the red packet in the specified range；

Second determines submodule, for the location information and the location information of other settings businessman according to the setting businessman, really The businessman's quantity being scheduled in the specified range, and by the ratio between the total number for being got red packet and businessman's quantity Value gets closeness as the red packet in the specified range；

Fisrt feature determines submodule, for by total stroke count of the total number for being got red packet, the payment transaction, described Red packet gets accounting, businessman's quantity and the red packet and gets closeness as the characteristic value extracted.

14. device according to claim 12, first extraction module include:

Third matched sub-block, for believing the gathering in the Receiving information and each barcode scanning payment data of the setting businessman Breath is matched, using the consistent barcode scanning payment data of Receiving information of Receiving information and the setting businessman as the setting quotient The corresponding barcode scanning payment data of family；

Third determines submodule, in the corresponding barcode scanning payment data of the setting businessman, will include sweeping for red packet mark Code payment data is determined as target barcode scanning payment data；

4th determines submodule, for determining that red packet transaction accounting, red packet check and write off distance according to the target barcode scanning payment data Dispersion, red packet check and write off average time interval, the contribution degree of best red packet code value, the transaction accounting of best red packet code value；

Second feature determines submodule, for checking and writing off red packet transaction accounting, the red packet apart from dispersion, the red packet Average time interval, the contribution degree of the best red packet code value, the transaction accounting of the best red packet code value are checked and write off as extraction The characteristic value arrived.

15. device according to claim 14, the described 4th determines that submodule is specifically used for:

16. device according to claim 14, the described 4th determines that submodule includes:

4th matched sub-block will be in the target barcode scanning payment data for being directed to any bar target barcode scanning payment data Red packet mark in red packet mark and each barcode scanning neck red packet data is matched, and will be had with the target barcode scanning payment data There are the barcode scanning neck red packet data of identical red packet mark as the corresponding barcode scanning neck red packet data of the target barcode scanning payment data；

5th determines submodule, for red according to the corresponding barcode scanning neck of the payment moment in the target barcode scanning payment data Red packet in bag data is got determines that red packet checks and writes off time interval constantly；

6th determines submodule, for red according to the corresponding barcode scanning neck of the location for payment in the target barcode scanning payment data Red packet in bag data gets position and determines that red packet checks and writes off distance interval；

Second computational submodule checks and writes off the variance of distance interval for calculating identified all red packets, the variance is made It checks and writes off for red packet apart from dispersion.

17. device according to claim 14, the described 4th determines that submodule includes:

It is grouped submodule, the target barcode scanning payment data is grouped for the red packet code value in being identified according to red packet, In, the red packet code value in the target barcode scanning payment data of each grouping is identical, in the target barcode scanning payment data of different grouping Red packet code value is not identical；

Sorting sub-module, for the sequence according to the quantity of target barcode scanning payment data in each grouping from high to low, to each grouping Corresponding red packet code value is ranked up, and determines the of the first red packet code value for ranking first and position of being number two according to ranking results Two red packet code values；

Quantity determines submodule, for determining the first number of the target barcode scanning payment data with the first red packet code value respectively Amount, the second quantity with the target barcode scanning payment data with the second red packet code value；

7th determines submodule, for by first quantity and the quantity for setting the corresponding barcode scanning payment data of businessman it Between ratio, as the transaction accounting of best red packet code value, and by the ratio between first quantity and second quantity Contribution degree as best red packet code value, wherein the best red packet code value is the first red packet code value.

18. device according to claim 11, the unsupervised learning algorithm include:

Principal component analysis PCA algorithm.

19. according to the method for claim 11, the output valve of the prediction model further include:

20. a kind of red packet material based on model is laid with prediction meanss, described device includes:

Second extraction module, for accessed barcode scanning neck red packet data and barcode scanning payment data to be carried out feature respectively and mentioned It takes, obtains the characteristic value of the barcode scanning neck red packet data and the characteristic value of barcode scanning payment data；

Input module, for by characteristic value input prediction model obtained, obtaining corresponding output valve, the prediction model with It is input value that barcode scanning, which leads the characteristic value of red packet data and the characteristic value of barcode scanning payment data, is equipped with red packet material to set businessman Probability be output valve.

21. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor Calculation machine program, wherein the processor realizes method as described in any one of claim 1 to 9 when executing described program.

22. a kind of computer equipment including memory, processor and stores the meter that can be run on a memory and on a processor Calculation machine program, wherein the processor realizes method as claimed in claim 10 when executing described program.