CN108876436A - A kind of electric business discount coupon based on integrated model uses probability forecasting method - Google Patents
A kind of electric business discount coupon based on integrated model uses probability forecasting method Download PDFInfo
- Publication number
- CN108876436A CN108876436A CN201810517366.4A CN201810517366A CN108876436A CN 108876436 A CN108876436 A CN 108876436A CN 201810517366 A CN201810517366 A CN 201810517366A CN 108876436 A CN108876436 A CN 108876436A
- Authority
- CN
- China
- Prior art keywords
- data
- feature
- discount coupon
- electric business
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to the technical fields of data mining, use probability forecasting method more particularly, to a kind of electric business discount coupon based on integrated model.Electric business discount coupon based on integrated model of the invention uses probability forecasting method, and building includes the integrated model of primary learner and secondary learner, the advantages of in conjunction with specific single model when predicting that discount coupon uses probability, evades the defect of single model prediction.In the case where sample data is very big, prediction accuracy is promoted, the extensive error of prediction model is reduced.Electric business can construct integrated model by this technology, and more accurate prediction user uses the probability value of discount coupon, be decided whether to provide discount coupon to user according to probability value, to reduce the operation cost of trade company, preferably reach the desired promotion effect of trade company.
Description
Technical field
The present invention relates to the technical fields of data mining, preferential more particularly, to a kind of electric business based on integrated model
Certificate uses probability forecasting method.
Background technique
With the improvement and popularization of mobile device, mobile Internet+all trades and professions enter the high speed development stage, among these
It is attracted eyeball the most with electric business consumption.According to incompletely statistics, the more than one hundred million venture company of electric business industry valuation at least 10, also not
The figure of weary 10,000,000,000 giant.The several hundred million consumers of electric business business association, all kinds of APP have recorded daily more than 10,000,000,000 user behaviors and
Position record, thus become one of big data scientific research and the best joint of commercial operation.With discount coupon vitalize old user or
Attracting new client to consume into shop is a kind of important marketing mode of electric business.
The discount coupon launched at random causes meaningless interference to most users.For businessman, the discount coupon distributed indiscriminately can
Brand reputation can be reduced, while being difficult to estimate cost of marketing, it is the important technology for improving discount coupon utilization rate that personalization, which is launched, it
The consumer with certain preference can be allowed to obtain real material benefit, while assigning the stronger marketing ability of businessman.It is existing excellent
Favour certificate uses Predicting Technique, is mostly that the buying behavior data based on user carry out the prediction based on single disaggregated model, however
When data volume is very big, single disaggregated model is easy over-fitting, and extensive error is big, influences to use probability finally for discount coupon
Prediction is not able to satisfy the forecast demand that electric business uses discount coupon probability.
Summary of the invention
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of, and the electric business discount coupon based on integrated model makes
With probability forecasting method, the advantages of in conjunction with single model when predicting that discount coupon uses probability, evades the defect of single model prediction, mention
Prediction accuracy is risen, the extensive error of prediction model is reduced.
In order to solve the above technical problems, the technical solution adopted by the present invention is that:
A kind of electric business discount coupon based on integrated model is provided and uses probability forecasting method, which is characterized in that including following
Step:
S1. it the on-line off-line historical consumption data of counting user and is pre-processed, the pretreatment includes rejecting missing
Rate is more than the data of given threshold and fills up to the data of missing, obtains the first data set;
S2. feature extraction is carried out to the data in the first data set in step S1 and obtains feature to be analyzed, construction feature
System, the feature architecture include trade company's feature, discount coupon feature, user characteristics, user and trade company's assemblage characteristic and other
Feature;
S3. the first data set in step S1 is divided into positive class data and negative class data, the negative class data be get it is excellent
Favour certificate quantity is greater than the data of the discount coupon quantity used, and the positive class data are the number in the first data set in addition to negative class data
According to;All positive class data are chosen, and repeatedly take the negative class data with 1~1.5 times of positive class data bulk at random, form training
Sample;Primary learner is constructed, the characteristic variable of the training sample is inputted, exports the probability value of each sampling samples prediction;
S4. secondary learner is constructed, the probability predicted for learner primary in step S3 based on each training sample is inputted
Value, exports the probability value to predict based on entire training sample.
Electric business discount coupon based on integrated model of the invention uses probability forecasting method, by constructing integrated model, more
Add the user that calculates to a nicety to use the probability value of discount coupon, is decided whether to send discount coupon to user according to probability value, to subtract
The operation cost at Shaoshang family preferably reaches the desired promotion effect of trade company.
Preferably, in step S1, the historical consumption data on-line off-line to the obtained user of statistics is normalized.
When due to statistics, the dimension and dimensional unit disunity of different characteristic can assess feature weight to prediction model and cause larger shadow
It rings, needs that feature is normalized, to solve the comparativity between feature.
Preferably, step S2 includes the following steps:
S21. data cleansing is carried out to the first data set in step S1, removes noise, obtain the second data set;In data set
Inevitably there are noise, noise prevents data from reflecting the actual conditions of problem very accurately, and feature extraction is come
It says, exceptional value often increases the variance of data, so that the change sensitivity of the data in normal range (NR) drops in prediction model
It is low.
S22. the constructing variable of the Data Analysis Data in the second data set described in step S21, the constructing variable
Including distribution characteristics, coverage rate, discrimination and availability;The distribution characteristics of data is detected from multiple angles, can also often be opened
Issue new angle extract, construction feature, the detecting of data can also primarily determine the data cover rate of some features, distinguish
The parameters such as degree, availability.
S23. the second data set described in step S21 establishes multi-dimensional data model, extracts multiple feature architectures, institute
Stating feature architecture includes trade company's feature, discount coupon feature, user characteristics, user and trade company's assemblage characteristic and other features;Gu
Fixed wherein one or two of dimension can count the data of other remaining dimensions, respectively it can be observed that each observation in the dimension
The difference of many index between object.
S24. pass through trade company's feature, discount coupon feature, user characteristics, user and trade company's assemblage characteristic in combination step S23
And the latent structure complex characteristic in other big feature architectures of feature five, the method for the feature combination include ratio method, row
Sequence method, interior extrapolation method and weighted sum method.Complex characteristic is constructed by feature combination, combines two or more features one
It rises, increasingly complex derivative feature can be formed, these features further extract valuable information from essential characteristic.
Preferably, multi-dimensional data described in step S23 include User, Merchant, Coupon, Discount_rate,
The data of seven dimensions of Distance, Date_received, Date.This seven dimensions enumerate discount coupon and use probabilistic forecasting
The influence factor of accuracy.
Preferably, after step S24, further include:By feature architecture described in step S23 by SQL or
Python.Pandas merges sentence and is fused on a table, and the prediction of the multi-dimensional data model is generated by this table
Variable.
Preferably, primary learner described in step S3 is that Logistic is returned, in decision tree and support vector machines
At least two combination.The number of primary learner according to the positive and negative ratio-dependent of sample, in conjunction with specific single model predict it is excellent
The advantages of when favour certificate uses, the defect for avoiding single model from predicting.
Preferably, Logistic recurrence, decision tree and the support equal by quantity of primary learner described in step S3
Vector machine composition,
Preferably, the secondary learner is support vector machines.Support vector machines is capable of handling nonlinear characteristic, can locate
Manage large-scale feature space, without relying on entire data set, not by the advantage that noise sample is interfered be other two kinds of classifiers institutes not
Have, therefore chooses support vector machines as secondary learner.
Compared with prior art, the beneficial effects of the invention are as follows:
Electric business discount coupon based on integrated model of the invention uses probability forecasting method, building include primary learner and
The integrated model of secondary learner, in conjunction with specific single model predict discount coupon use probability when the advantages of, it is pre- to evade single model
The defect of survey.In the case where sample data is very big, prediction accuracy is promoted, the extensive error of prediction model is reduced.Electric business can
To construct integrated model by this technology, more accurate prediction user uses the probability value of discount coupon, is determined according to probability value
It is fixed whether discount coupon to be provided to user, to reduce the operation cost of trade company, preferably reach the desired promotion effect of trade company.
Detailed description of the invention
Fig. 1 is the flow chart that the electric business discount coupon of the invention based on integrated model uses probability forecasting method.
Fig. 2 is the structural schematic diagram of integrated model of the invention.
Fig. 3 is the building flow chart of feature architecture of the invention.
Specific embodiment
The present invention is further illustrated With reference to embodiment.Wherein, attached drawing only for illustration,
What is indicated is only schematic diagram, rather than pictorial diagram, should not be understood as the limitation to this patent;Reality in order to better illustrate the present invention
Example is applied, the certain components of attached drawing have omission, zoom in or out, and do not represent the size of actual product;To those skilled in the art
For, the omitting of some known structures and their instructions in the attached drawings are understandable.
The same or similar label correspond to the same or similar components in the attached drawing of the embodiment of the present invention;It is retouched in of the invention
In stating, it is to be understood that if the orientation or positional relationship for having the instructions such as term " on ", "lower", "left", "right" is based on attached drawing
Shown in orientation or positional relationship, be merely for convenience of description of the present invention and simplification of the description, rather than indication or suggestion is signified
Device or element must have a particular orientation, be constructed and operated in a specific orientation, therefore positional relationship is described in attached drawing
Term only for illustration, should not be understood as the limitation to this patent, for the ordinary skill in the art, can
To understand the concrete meaning of above-mentioned term as the case may be.
Embodiment 1
The flow chart of probability forecasting method is used for the electric business discount coupon of the invention based on integrated model as shown in Figure 1,
Include the following steps:
S1. it the on-line off-line historical consumption data of counting user and is pre-processed, the pretreatment includes rejecting missing
Rate is more than the data of given threshold and fills up to the data of missing, obtains the first data set;
S2. feature extraction is carried out to the data in the first data set in step S1 and obtains feature to be analyzed, construction feature
System, the feature architecture include trade company's feature, discount coupon feature, user characteristics, user and trade company's assemblage characteristic and other
Feature;The feature that trade company's feature, discount coupon feature, user characteristics, user and trade company's assemblage characteristic and other features include point
It is not listed in table one, table two, table three, table four, table five;
One trade company's feature of table includes feature list
Two discount coupon feature of table includes feature list
day_of_week | The date for obtaining discount coupon is that week is several |
day_of_month | Which of middle of the month the date for obtaining discount coupon be |
discount_man | Some discount coupon completely subtract in minimum charge |
discount_jian | Some discount coupon completely subtract in preferential amount |
is_man_jian | Whether some discount coupon is completely to subtract certificate |
discount_rate | The discount rate of some discount coupon |
coupon_count | The total quantity of some discount coupon |
Three user characteristics of table include feature list
Four user of table and trade company's assemblage characteristic include feature list
Other features of table five include feature list
S3. the first data set in step S1 is divided into positive class data and negative class data, the negative class data be get it is excellent
Favour certificate quantity is greater than the data of the discount coupon quantity used, and the positive class data are the number in the first data set in addition to negative class data
According to;All positive class data are chosen, and repeatedly take the negative class data with 1~1.5 times of positive class data bulk at random, form training
Sample;Primary learner is constructed, the characteristic variable of the training sample is inputted, exports the probability value of each sampling samples prediction;
The method of multiple repairing weld avoids the defect that mass data is not used, key message is easy to be lost.
S4. secondary learner is constructed, the probability predicted for learner primary in step S3 based on each training sample is inputted
Value, exports the probability value to predict based on entire training sample.
Electric business discount coupon based on integrated model of the invention uses probability forecasting method, includes primary study by building
The integrated model of device and secondary learner predicts that user uses the probability value of discount coupon more accurately, is determined according to probability value
It is fixed whether discount coupon to be sent to user, to reduce the operation cost of trade company, preferably reach the desired promotion effect of trade company.
S21. data cleansing is carried out to the first data set in step S1, removes noise, obtain the second data set;In data set
Inevitably there are noise, noise prevents data from reflecting the actual conditions of problem very accurately, and feature extraction is come
It says, exceptional value often increases the variance of data, so that the change sensitivity of the data in normal range (NR) drops in prediction model
It is low.
S22. the constructing variable of the Data Analysis Data in the second data set described in step S21, the constructing variable
Including distribution characteristics, coverage rate, discrimination and availability;The distribution characteristics of data is detected from multiple angles, can also often be opened
Issue new angle extract, construction feature, the detecting of data can also primarily determine the data cover rate of some features, distinguish
The parameters such as degree, availability.
S23. the second data set described in step S21 establishes multi-dimensional data model, and the multi-dimensional data includes
The data of seven dimensions of User, Merchant, Coupon, Discount_rate, Distance, Date_received, Date;
According to the understanding to business, multiple feature architectures are extracted, the feature architecture includes trade company's feature, discount coupon feature, user
Feature, user and trade company's assemblage characteristic and other features;Fixed wherein one or two of dimension, can be to the number of other remaining dimensions
According to counting respectively, it can be observed that the difference of many index in the dimension between each observation object.
S24. pass through trade company's feature, discount coupon feature, user characteristics, user and trade company's assemblage characteristic in combination step S23
And the latent structure complex characteristic in other big feature architectures of feature five, the method for the feature combination include ratio method, row
Sequence method, interior extrapolation method and weighted sum method.Complex characteristic is constructed by feature combination, combines two or more features one
It rises, increasingly complex derivative feature can be formed, these features further extract valuable information from essential characteristic.
Ratio method:If class of subscriber is in the discount coupon utilization rate trend feature of trade company's classification, by user in trade company with preferential
The quantity of certificate consumption is divided by with the user in the commodity sum of this merchant purchasing, reflects class of subscriber in trade company by ratio
The discount coupon utilization rate of classification is distributed.
Ranking method:A certain range of characteristic value is ranked up, new feature is formed.If the utilization rate of discount coupon is at this
Sequence in trade company's classification, this Characterizations burning hot degree of the discount coupon in trade company's classification.
Interior extrapolation method:The data set of different characteristic covering is compared, is counted again after taking intersection.Some user is such as asked to use preferential
When certificate shopping rate, need first to find out the set of discount coupon set and user purchase commodity that the user receives, then at the two
Intersection is taken in set, show that the user uses the set of discount coupon purchase commodity.This Characterizations user is in purchase quotient
Product be using discount coupon consume intention size, to predict the user whether will use discount coupon purchase commodity have biggish meaning
Justice.
Weighted sum method:The value of different characteristic is summed it up by different weights, obtains a new feature.For example, right
It is done shopping in a user using discount coupon, get discount coupon but does not use discount coupon shopping, do not got discount coupon shopping, do not lead
It takes discount coupon also not do shopping in these four behaviors, if only distinguishing statistics number by four kinds of behaviors, has been ignored as four
Connection between kind behavior obtains a comprehensive interbehavior score, by the weighted sum to four kinds of behavior numbers with this
As a new feature.
The present embodiment in construction feature system, first from User, Merchant, Coupon, Discount_rate,
Seven dimensions of Distance, Date_received, Date count essential characteristic, then carry out feature respectively on this basis and turn
It changes, the construction of complex characteristic is completed by the combination between functional transformation and different characteristic.Then five big feature architectures are passed through
The merging sentence of SQL or Python.Pandas is fused on a table, is become by the prediction that this table can generate model
Amount.When due to statistics, the dimension and dimensional unit disunity of different characteristic can be caused prediction model assessment feature weight larger
It influences, needs that feature is normalized, to solve the comparativity between feature.Finally, the missing values to feature carry out
Filling, in this link it is noted that carrying out Missing Data Filling according to the concrete condition of different characteristic.Implementation step such as Fig. 3 institute
Show.
Primary learner in the present embodiment chooses that related coefficient is lower and the Logistic of better performances is returned, decision
Tree, these three common classifiers of support vector machines are as primary learner.The number of primary learner is according to the positive and negative of sample
Ratio-dependent, if the positive and negative class ratio of data set is 1:N, our stochastical sampling n times train n primary learner, and wherein n/3
Logistic returns primary learner, n/3 decision tree primary learner, and n/3 support vector machines primary learner is kept away with this
Exempt from information loss.Its input of primary learner is the characteristic variable of training sample, exports and is based on each sampling for primary learner
The probability value of sample predictions.The positive and negative class ratio data of the data set that the present embodiment uses is about 1:9, stochastical sampling 9 times is instructed
Practice 9 primary learners, wherein 3 Logistic return primary learner, 3 decision tree primary learners, 3 supporting vectors
Machine primary learner.
The selection of secondary learner is returned based on Logistic, decision tree, these three disaggregated models of support vector machines are excellent to be lacked
Point compares, and support vector machines is capable of handling nonlinear characteristic, is capable of handling large-scale feature space, without relying on entire data set,
It by the advantage that noise sample is interfered is learned so choosing support vector machines as secondary not available for other two kinds of classifiers
Practise device.It inputs the probability value predicted for primary learner based on each sampling samples, exports and is based on entirely for secondary learner
The probability value of sample predictions.To guarantee the accuracy calculated, probability value by secondary learner based on entire sample predictions and just
The probability value that grade learner is predicted based on each sampling samples compares, and casts out and secondary learner prediction probability value is compared to deviation
Big primary learner.
The primary learner of the present embodiment and secondary learner are set up to form integrated model, and process such as Fig. 2 is embodied
Shown, by the feature x of each discount coupon sample, including trade company's feature, discount coupon feature, user characteristics, user are combined with trade company
The input variable of feature and other features as primary learner, exports and uses prediction probability for each discount coupon sample.9
Primary learner is to raw 9 prediction probabilities of each discount coupon sample common property.Then, with the prediction probability of 9 primary learners
As the input of secondary learner, the probability value for secondary learner based on entire sample predictions is exported.Then, secondary is learnt
The probability value that probability value of the device based on entire sample predictions is predicted with primary learner based on each sampling samples compares, and casts out
And the secondary learner prediction probability value primary learner big compared to deviation carries out the final pre- of secondary learner again
It surveys.
Obviously, the above embodiment of the present invention be only to clearly illustrate example of the present invention, and not be pair
The restriction of embodiments of the present invention.For those of ordinary skill in the art, may be used also on the basis of the above description
To make other variations or changes in different ways.There is no necessity and possibility to exhaust all the enbodiments.It is all this
Made any modifications, equivalent replacements, and improvements etc., should be included in the claims in the present invention within the spirit and principle of invention
Protection scope within.
Claims (8)
1. a kind of electric business discount coupon based on integrated model uses probability forecasting method, which is characterized in that include the following steps:
S1. it the on-line off-line historical consumption data of counting user and is pre-processed, the pretreatment includes that reject miss rate super
It crosses the data of given threshold and the data of missing is filled up, obtain the first data set;
S2. feature extraction is carried out to the data in the first data set in step S1 and obtains feature to be analyzed, construction feature system,
The feature architecture includes trade company's feature, discount coupon feature, user characteristics, user and trade company's assemblage characteristic and other features;
S3. the first data set in step S1 is divided into positive class data and negative class data, the negative class data are the discount coupon got
Quantity is greater than the data of the discount coupon quantity used, and the positive class data are the data in the first data set in addition to negative class data;
All positive class data are chosen, and repeatedly take the negative class data with 1~1.5 times of positive class data bulk at random, form training sample
This;Primary learner is constructed, the characteristic variable of the training sample is inputted, exports the probability value of each sampling samples prediction;
S4. secondary learner is constructed, the probability value predicted for learner primary in step S3 based on each training sample is inputted, it is defeated
Probability value to be predicted based on entire training sample out.
2. the electric business discount coupon according to claim 1 based on integrated model uses probability forecasting method, which is characterized in that
In step S1, the historical consumption data on-line off-line to the obtained user of statistics is normalized.
3. the electric business discount coupon according to claim 1 based on integrated model uses probability forecasting method, which is characterized in that
Step S2 includes the following steps:
S21. data cleansing is carried out to the first data set in step S1, removes noise, obtain the second data set;
S22. the constructing variable of the Data Analysis Data in the second data set described in step S21, the constructing variable include
Distribution characteristics, coverage rate, discrimination and availability;
S23. the second data set described in step S21 establishes multi-dimensional data model, extracts multiple feature architectures, the spy
Sign system includes trade company's feature, discount coupon feature, user characteristics, user and trade company's assemblage characteristic and other features;
S24. by trade company's feature, discount coupon feature, user characteristics, user and trade company's assemblage characteristic in combination step S23 and
Latent structure complex characteristic in other big feature architectures of feature five, the method for feature combination include ratio method, ranking method,
Interior extrapolation method and weighted sum method.
4. the electric business discount coupon according to claim 3 based on integrated model uses probability forecasting method, which is characterized in that
Multi-dimensional data described in step S23 includes User, Merchant, Coupon, Discount_rate, Distance, Date_
The data of seven dimensions of received, Date.
5. the electric business discount coupon according to claim 4 based on integrated model uses probability forecasting method, which is characterized in that
After step S24, further include:Feature architecture described in step S23 is merged sentence by SQL or Python.Pandas to melt
It closes onto a table, and generates the predictive variable of the multi-dimensional data model by this table.
6. the electric business discount coupon according to claim 1 based on integrated model uses probability forecasting method, which is characterized in that
Primary learner described in step S3 is at least two combination in Logistic recurrence, decision tree and support vector machines.
7. the electric business discount coupon according to claim 6 based on integrated model uses probability forecasting method, which is characterized in that
Primary learner described in step S3 is returned by the equal Logistic of quantity, decision tree and support vector machines form.
8. the electric business discount coupon according to claim 7 based on integrated model uses probability forecasting method, which is characterized in that
The secondary learner is support vector machines.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810517366.4A CN108876436A (en) | 2018-05-25 | 2018-05-25 | A kind of electric business discount coupon based on integrated model uses probability forecasting method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810517366.4A CN108876436A (en) | 2018-05-25 | 2018-05-25 | A kind of electric business discount coupon based on integrated model uses probability forecasting method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108876436A true CN108876436A (en) | 2018-11-23 |
Family
ID=64334246
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810517366.4A Pending CN108876436A (en) | 2018-05-25 | 2018-05-25 | A kind of electric business discount coupon based on integrated model uses probability forecasting method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108876436A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109711906A (en) * | 2019-01-10 | 2019-05-03 | 哈步数据科技(上海)有限公司 | A kind of distribution method and system of favor information |
CN109961191A (en) * | 2019-04-03 | 2019-07-02 | 北京奇艺世纪科技有限公司 | A kind of discount coupon distribution method and device |
CN110782277A (en) * | 2019-10-12 | 2020-02-11 | 上海陆家嘴国际金融资产交易市场股份有限公司 | Resource processing method, resource processing device, computer equipment and storage medium |
CN111275470A (en) * | 2018-12-04 | 2020-06-12 | 北京嘀嘀无限科技发展有限公司 | Service initiation probability prediction method and training method and device of model thereof |
CN111310913A (en) * | 2020-01-19 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Data processing method, asset allocation method, model, device and equipment |
WO2020125106A1 (en) * | 2018-12-21 | 2020-06-25 | 苏宁易购集团股份有限公司 | Similarity model-based data processing method and system |
CN113824974A (en) * | 2021-08-27 | 2021-12-21 | 北京达佳互联信息技术有限公司 | Virtual asset sending method and device, electronic equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107301562A (en) * | 2017-05-16 | 2017-10-27 | 重庆邮电大学 | A kind of O2O reward vouchers use big data Forecasting Methodology |
CN107330741A (en) * | 2017-07-07 | 2017-11-07 | 北京京东尚科信息技术有限公司 | Graded electron-like certificate uses Forecasting Methodology, device and electronic equipment |
CN107578281A (en) * | 2017-08-31 | 2018-01-12 | 湖南大学 | User preferential certificate behavior prediction method and model building method under e-commerce environment |
CN107578332A (en) * | 2017-09-22 | 2018-01-12 | 深圳乐信软件技术有限公司 | A kind of method, apparatus, equipment and storage medium for recommending cash commodity |
-
2018
- 2018-05-25 CN CN201810517366.4A patent/CN108876436A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107301562A (en) * | 2017-05-16 | 2017-10-27 | 重庆邮电大学 | A kind of O2O reward vouchers use big data Forecasting Methodology |
CN107330741A (en) * | 2017-07-07 | 2017-11-07 | 北京京东尚科信息技术有限公司 | Graded electron-like certificate uses Forecasting Methodology, device and electronic equipment |
CN107578281A (en) * | 2017-08-31 | 2018-01-12 | 湖南大学 | User preferential certificate behavior prediction method and model building method under e-commerce environment |
CN107578332A (en) * | 2017-09-22 | 2018-01-12 | 深圳乐信软件技术有限公司 | A kind of method, apparatus, equipment and storage medium for recommending cash commodity |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111275470A (en) * | 2018-12-04 | 2020-06-12 | 北京嘀嘀无限科技发展有限公司 | Service initiation probability prediction method and training method and device of model thereof |
CN111275470B (en) * | 2018-12-04 | 2023-12-01 | 北京嘀嘀无限科技发展有限公司 | Service initiation probability prediction method and training method and device of model thereof |
WO2020125106A1 (en) * | 2018-12-21 | 2020-06-25 | 苏宁易购集团股份有限公司 | Similarity model-based data processing method and system |
CN109711906A (en) * | 2019-01-10 | 2019-05-03 | 哈步数据科技(上海)有限公司 | A kind of distribution method and system of favor information |
CN109961191A (en) * | 2019-04-03 | 2019-07-02 | 北京奇艺世纪科技有限公司 | A kind of discount coupon distribution method and device |
CN110782277A (en) * | 2019-10-12 | 2020-02-11 | 上海陆家嘴国际金融资产交易市场股份有限公司 | Resource processing method, resource processing device, computer equipment and storage medium |
CN111310913A (en) * | 2020-01-19 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Data processing method, asset allocation method, model, device and equipment |
CN113824974A (en) * | 2021-08-27 | 2021-12-21 | 北京达佳互联信息技术有限公司 | Virtual asset sending method and device, electronic equipment and storage medium |
CN113824974B (en) * | 2021-08-27 | 2022-10-04 | 北京达佳互联信息技术有限公司 | Virtual asset sending method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108876436A (en) | A kind of electric business discount coupon based on integrated model uses probability forecasting method | |
Li et al. | Spam review detection with graph convolutional networks | |
Hadden et al. | Computer assisted customer churn management: State-of-the-art and future trends | |
Chen et al. | Predicting default risk on peer-to-peer lending imbalanced datasets | |
CN106407999A (en) | Rule combined machine learning method and system | |
CN105931068A (en) | Cardholder consumption figure generation method and device | |
Muslim et al. | New model combination meta-learner to improve accuracy prediction P2P lending with stacking ensemble learning | |
CN109003091A (en) | A kind of risk prevention system processing method, device and equipment | |
Alareeni | A review of auditors’ GCOs, statistical prediction models and artificial intelligence technology | |
US20220172258A1 (en) | Artificial intelligence-based product design | |
CN110930038A (en) | Loan demand identification method, loan demand identification device, loan demand identification terminal and loan demand identification storage medium | |
Guliyev et al. | Customer churn analysis in banking sector: Evidence from explainable machine learning model | |
Kurinjimalar Ramu et al. | Green supply chain management; with dematel MCDM analysis | |
Giri et al. | Exploitation of social network data for forecasting garment sales | |
Machine | Predicting loan approval of bank direct marketing data using ensemble machine learning algorithms | |
Mei et al. | Research on e-commerce coupon user behavior prediction technology based on decision tree algorithm | |
CN111091409A (en) | Client tag determination method and device and server | |
Wang et al. | Multiview Graph Learning for Small‐and Medium‐Sized Enterprises’ Credit Risk Assessment in Supply Chain Finance | |
CN111931069B (en) | User interest determination method and device and computer equipment | |
Park et al. | Spec guidance for engineering design based on data mining and neural networks | |
Alquhtani et al. | Analytics in support of e-commerce systems using machine learning | |
Jeong et al. | An algorithm for supporting decision making in stock investment through opinion mining and machine learning | |
Arundina et al. | Artificial intelligence for Islamic sukuk rating predictions | |
Deng et al. | Sampling method based on improved C4. 5 decision tree and its application in prediction of telecom customer churn | |
CN115829683A (en) | Power integration commodity recommendation method and system based on inverse reward learning optimization |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181123 |
|
RJ01 | Rejection of invention patent application after publication |