CN110428281B - Method and device for jointly determining peer-to-peer resource quantity aiming at multiple associated products - Google Patents
Method and device for jointly determining peer-to-peer resource quantity aiming at multiple associated products Download PDFInfo
- Publication number
- CN110428281B CN110428281B CN201910660160.1A CN201910660160A CN110428281B CN 110428281 B CN110428281 B CN 110428281B CN 201910660160 A CN201910660160 A CN 201910660160A CN 110428281 B CN110428281 B CN 110428281B
- Authority
- CN
- China
- Prior art keywords
- peer
- user
- resource
- associated product
- products
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06Q30/0207—Discounts or incentives, e.g. coupons or rebates
- G06Q30/0208—Trade or exchange of goods or services in exchange for incentives or rewards
-
- 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Entrepreneurship & Innovation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the specification provides a method and a device for jointly determining peer-to-peer resource quantity for a plurality of associated products, wherein the method comprises the following steps: taking the peer-to-peer resource quantity characteristics, the user portrait characteristics and the scene characteristics of the target user under each peer-to-peer resource quantity combination in n peer-to-peer resource quantity combinations of a plurality of associated products as inputs of a pre-trained neural network model, and obtaining the estimated conversion rate of various associated products under each peer-to-peer resource quantity combination through the output of the neural network model; determining estimated resource benefits under each peer-to-peer resource amount combination according to the estimated conversion rate of each associated product under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of each associated product under each peer-to-peer resource amount combination and the cost resource amounts of each associated product; and combining the peer-to-peer resource amount with the maximum estimated resource benefit, and determining the peer-to-peer resource amount determined by the final combination of the multiple associated products aiming at the target user, so that the overall benefit maximization of the multiple associated products can be realized.
Description
Technical Field
One or more embodiments of the present specification relate to the field of computers, and more particularly, to methods and apparatus for jointly determining an amount of peer-to-peer resources for a plurality of associated products.
Background
In many scenarios, product demanders may exchange certain products with some kind of resource, such as "energy" in an ant forest for tree seedlings, and money for merchandise in a merchandise sales. In combination with a commodity sales scenario, multiple associated commodities are often required to be displayed together for sale so as to meet different requirements of different customers, and the purchase conversion rate of the whole customer is increased. However, in the process of simultaneously displaying multiple commodities and selling prices of each commodity, the selling price change of a single commodity affects not only the purchasing conversion rate of the commodity, but also the purchasing conversion rate of other related commodities displayed together, so that the selling prices of the multiple related commodities can affect the respective purchasing conversion rates.
Merchants pursue benefit maximization in a vending scenario, but no good commodity pricing method exists at present, so a method for jointly determining the amount of peer-to-peer resources for multiple associated products is urgently needed to achieve overall benefit maximization of the multiple associated products.
Disclosure of Invention
One or more embodiments of the present specification describe a method and apparatus for jointly determining an amount of peer-to-peer resources for a plurality of associated products, which can maximize the overall benefit of the plurality of associated products.
In a first aspect, a method is provided for jointly determining an amount of peer-to-peer resources for a plurality of associated products, the method comprising:
acquiring user portrait features of a target user and scene features of the target user;
taking the peer-to-peer resource quantity characteristics, the user portrait characteristics and the scene characteristics of each of n peer-to-peer resource quantity combinations of a plurality of associated products as inputs of a pre-trained neural network model, and obtaining the estimated conversion rate of each associated product of each of the n peer-to-peer resource quantity combinations through the output of the neural network model;
determining estimated resource benefits under each peer-to-peer resource amount combination according to the estimated conversion rate of each associated product under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of each associated product under each peer-to-peer resource amount combination and the cost resource amounts of each associated product;
and determining the peer-to-peer resource amount combination with the maximum estimated resource benefit in the n peer-to-peer resource amount combinations as the peer-to-peer resource amount determined by the final combination of the plurality of associated products aiming at the target user.
In one possible implementation, the multiple associated products are cycle cards of multiple different durations for the same service.
In one possible implementation, the n peer-to-peer resource amount combinations are determined by:
sampling n times by adopting a random sampling method according to the peer-to-peer resource amount interval corresponding to each associated product, and obtaining an optional peer-to-peer resource amount corresponding to the associated product each time, wherein the optional peer-to-peer resource amounts of various associated products form a peer-to-peer resource amount combination; wherein, the value of n is a preset value.
In one possible implementation manner, the multiple associated products are simultaneously used as products to be converted, so that the user is allowed to simultaneously convert the multiple associated products, and the neural network model is a multi-label classification model.
In one possible implementation manner, when the multiple associated products are simultaneously used as products to be converted, the user can only convert one associated product, and the neural network model is a multi-classification model.
In one possible implementation, the neural network model is trained by:
acquiring a training sample set, wherein each training sample comprises the peer-to-peer resource amount of each associated product facing a potential user, and the user portrait characteristic and scene characteristic of the potential user, and the conversion condition of the potential user on each associated product;
and training the neural network model by taking the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in at least a part of the training samples, the user portrait characteristic and the scene characteristic of the potential user as sample characteristics and the conversion condition of the potential user of each associated product in the training samples as sample labels.
Further, the acquiring a training sample set includes:
in a preset sample acquisition period, for potential users, randomly determining the peer-to-peer resource quantity of each associated product facing the potential users according to the peer-to-peer resource quantity interval corresponding to each associated product;
and generating a training sample of the training sample set according to the corresponding peer-to-peer resource quantity of each associated product faced by the potential user and the conversion condition of the potential user to each associated product under the peer-to-peer resource quantity.
Further, after the training of the neural network model, the method further includes:
and cross-verifying the trained neural network model according to the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in another part of training samples, and the user portrait characteristic and the scene characteristic of the potential user as sample characteristics, and taking the conversion condition of the potential user of each training sample to each associated product as a sample label.
In one possible implementation, the user portrait features include at least one of the following features:
gender, age, academic, consumption.
In one possible implementation, the scene features include at least one of the following features:
the current city, the application stay time.
In a second aspect, there is provided an apparatus for jointly determining an amount of peer-to-peer resources for a plurality of associated products, the apparatus comprising:
the acquisition unit is used for acquiring user portrait features of a target user and scene features of the target user;
the first estimating unit is used for taking the peer-to-peer resource quantity characteristics under each peer-to-peer resource quantity combination in n peer-to-peer resource quantity combinations of a plurality of associated products, the user portrait characteristics and the scene characteristics acquired by the acquiring unit as inputs of a pre-trained neural network model, and obtaining the estimated conversion rate of each associated product under each peer-to-peer resource quantity combination in the n peer-to-peer resource quantity combinations through the output of the neural network model;
the second estimating unit is used for determining estimated resource benefits under each peer-to-peer resource amount combination according to the estimated conversion rate of each associated product under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of each associated product under each peer-to-peer resource amount combination and the cost resource amounts of each associated product obtained by the first estimating unit;
and the determining unit is used for determining the peer-to-peer resource amount combination with the maximum estimated resource benefit in the n peer-to-peer resource amount combinations as the peer-to-peer resource amount determined by the final combination of the plurality of associated products aiming at the target user.
In a third aspect, there is provided a computer readable storage medium having stored thereon a computer program which, when executed in a computer, causes the computer to perform the method of the first aspect.
In a fourth aspect, there is provided a computing device comprising a memory having executable code stored therein and a processor which, when executing the executable code, implements the method of the first aspect.
Through the method and the device provided by the embodiment of the specification, firstly, the user portrait characteristics of the target user and the scene characteristics of the target user are obtained; then taking the peer-to-peer resource quantity characteristics, the user portrait characteristics and the scene characteristics of each peer-to-peer resource quantity combination in n peer-to-peer resource quantity combinations of a plurality of associated products as inputs of a pre-trained neural network model, and obtaining the estimated conversion rate of each associated product in each peer-to-peer resource quantity combination in the n peer-to-peer resource quantity combinations through the output of the neural network model; then, according to the estimated conversion rate of various associated products under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of various associated products under each peer-to-peer resource amount combination and the cost resource amounts of various associated products, determining the estimated resource benefits under each peer-to-peer resource amount combination; and finally, determining the peer-to-peer resource amount combination with the maximum estimated resource benefit in the n peer-to-peer resource amount combinations as the peer-to-peer resource amount determined by the final combination of the plurality of associated products aiming at the target user. From the above, the embodiment of the specification can jointly consider the interaction of the peer-to-peer resource amounts of various associated products on the conversion rate of the products, so that the overall benefit maximization of the various associated products can be realized.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic illustration of an implementation scenario of an embodiment disclosed herein;
FIG. 2 illustrates a method flow diagram for jointly determining peer-to-peer resource amounts for multiple associated products, according to one embodiment;
FIG. 3 illustrates a multi-label classification model structure schematic according to one embodiment;
FIG. 4 illustrates a schematic block diagram of an apparatus for jointly determining peer-to-peer resource amounts for multiple associated products, according to one embodiment.
Detailed Description
The following describes the scheme provided in the present specification with reference to the drawings.
Fig. 1 is a schematic diagram of an implementation scenario of an embodiment disclosed in the present specification. The implementation scenario involves jointly determining an amount of peer-to-peer resources for a plurality of associated products. It will be appreciated that in the embodiment of the present disclosure, the associated product may be a plurality of products displayed simultaneously in a sales scenario, or may be a plurality of products of other types that may be exchanged with other resources. When the resource exchanged with the associated product is currency, the peer-to-peer resource amount may be understood as the commodity price. Taking the commodity sales scenario shown in fig. 1 as an example, dynamic pricing of commodities is often involved, that is, different commodity prices may be available to different users, and each user may be re-priced at different times. It will be appreciated that the price of the good will have an impact on whether the user has purchased the good, that is, will have an impact on the purchase conversion, which represents the ratio of the number of people purchasing the good to the total number of people arriving.
The related goods can be any type of goods, for example, the shared bicycle cycle card is divided into a plurality of related card types such as a month card, a quarter card, a half year card and a year card, taking a bicycle month card as an example, a user purchases a month card, the number of riding times is not limited within one month from the effective moment of the month card, and the user is free 2 hours before riding for one time. The month card, the quarter card, the half year card and the year card can be regarded as a group of related commodities, and the selling price of the month card, the quarter card, the half year card and the year card in the same showing page is determined in a pricing project.
The method for jointly determining the peer-to-peer resource quantity aiming at the multiple associated products provided by the embodiment of the specification can be suitable for an on-line scene for resource exchange and also can be suitable for an off-line scene for resource exchange. Wherein, on-line scene: on the internet, by displaying goods and prices on mobile applications, web pages, etc., the buyer and seller do not have to meet to complete the sales transaction. Offline scene: in the traditional sales mode, goods are sold in shops or booths, and the buyers and sellers meet to complete transactions.
Since the selling prices of multiple associated commodities affect each other to their respective purchase conversions, and the merchant pursues the maximization of interest in the sales scenario, where the interest is determined by the selling price, cost, and conversion of the commodity, the maximization of interest of multiple associated commodities requires joint consideration of the selling price, cost, and the effect of each commodity on the purchase conversion, the embodiment of the present disclosure achieves the maximization of interest by a joint modeling commodity sales pricing method.
In the present embodiment, the commodity sales scenario is merely described as an example, but the present invention is not limited thereto, and the present embodiment may be applied to various resource exchange scenarios.
FIG. 2 illustrates a flow diagram of a method for jointly determining peer-to-peer resource amounts for multiple associated products, which may be based on the application scenario illustrated in FIG. 1, according to one embodiment. As shown in fig. 2, the method for jointly determining the peer-to-peer resource amount for a plurality of associated products in this embodiment includes the steps of: step 21, obtaining user portrait features of a target user and scene features of the target user; step 22, taking the peer-to-peer resource quantity characteristics, the user portrait characteristics and the scene characteristics of each peer-to-peer resource quantity combination in n peer-to-peer resource quantity combinations of a plurality of associated products as inputs of a pre-trained neural network model, and obtaining estimated conversion rates of various associated products under each peer-to-peer resource quantity combination in the n peer-to-peer resource quantity combinations through the output of the neural network model; step 23, determining estimated resource benefits under each peer-to-peer resource amount combination according to the estimated conversion rate of each associated product under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of each associated product under each peer-to-peer resource amount combination and the cost resource amounts of each associated product; and step 24, determining the peer-to-peer resource amount combination with the maximum estimated resource benefit in the n peer-to-peer resource amount combinations as the peer-to-peer resource amount determined by the final combination of the plurality of associated products aiming at the target user. Specific implementations of the above steps are described below.
First, in step 21, user portrait features of a target user and scene features of the target user are acquired. It will be appreciated that different users may be affected differently by the price of the commodity, and thus different prices of the commodity may be determined for different users to achieve maximum purchase conversion.
In one example, the user portrait features include at least one of the following features:
gender, age, academic, consumption.
In one example, the scene features include at least one of the following:
the current city, the application stay time.
And then, in step 22, taking the peer-to-peer resource quantity characteristics, the user portrait characteristics and the scene characteristics of each of n peer-to-peer resource quantity combinations of the plurality of associated products as inputs of a pre-trained neural network model, and obtaining the estimated conversion rate of each associated product of each of the n peer-to-peer resource quantity combinations through the output of the neural network model. It will be appreciated that multiple products may be manually set as a set of associated products for which the peer-to-peer resource amount for the resource conversion is jointly determined.
In one example, the plurality of associated products are cycle cards of a plurality of different durations of a same service. For example, a month card, a quarter card, a half year card, a year card of a shared bicycle.
In one example, the n peer-to-peer resource amount combinations are determined by:
sampling n times by adopting a random sampling method according to the peer-to-peer resource amount interval corresponding to each associated product, and obtaining an optional peer-to-peer resource amount corresponding to the associated product each time, wherein the optional peer-to-peer resource amounts of various associated products form a peer-to-peer resource amount combination; wherein, the value of n is a preset value.
Taking the related products as related commodities, the equivalent resource amount as commodity price, 5 prices can be usually obtained for each commodity, and 5 price combinations corresponding to 4 commodities are obtained 4 =625 combinations, 7 goods corresponding to 78125 combinations.
In the embodiment of the present disclosure, the number of price combinations is preset, and a random sampling method is adopted for each price combination, so that n price combinations are obtained, where n is a limited number of price combinations, for example, n is 200, where n is related to the performance required to predict in real time, and the larger n is the larger the calculation amount for each user, the longer the time is, so that the number of price combinations can be manually set according to the real-time requirement.
Furthermore, the n peer-to-peer resource amount combinations described in the embodiments of the present specification may also be determined by means of hierarchical random sampling. Taking the commodity sales scenario as an example, price combination sampling can also use layered random sampling besides pure random sampling, if the price interval is 5 to 9 yuan, firstly, one interval is randomly selected from four intervals of 5 to 6,6 to 7,7 to 8 and 8 to 9 yuan, and then random sampling is carried out in the interval, so that the sampling is more uniform.
In one example, the multiple associated products are simultaneously used as products to be converted, so that a user is allowed to simultaneously convert the multiple associated products, and the neural network model is a multi-label classification model.
In another example, the user can only convert one associated product when the plurality of associated products are simultaneously used as products to be converted, and the neural network model is a multi-classification model.
In one example, the neural network model is trained by:
acquiring a training sample set, wherein each training sample comprises the peer-to-peer resource amount of each associated product facing a potential user, and the user portrait characteristic and scene characteristic of the potential user, and the conversion condition of the potential user on each associated product;
and training the neural network model by taking the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in at least a part of the training samples, the user portrait characteristic and the scene characteristic of the potential user as sample characteristics and the conversion condition of the potential user of each associated product in the training samples as sample labels.
Further, the acquiring a training sample set includes:
in a preset sample acquisition period, for potential users, randomly determining the peer-to-peer resource quantity of each associated product facing the potential users according to the peer-to-peer resource quantity interval corresponding to each associated product;
and generating a training sample of the training sample set according to the corresponding peer-to-peer resource quantity of each associated product faced by the potential user and the conversion condition of the potential user to each associated product under the peer-to-peer resource quantity.
Further, after the training of the neural network model, the method further includes:
and cross-verifying the trained neural network model according to the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in another part of training samples, and the user portrait characteristic and the scene characteristic of the potential user as sample characteristics, and taking the conversion condition of the potential user of each training sample to each associated product as a sample label.
FIG. 3 illustrates a multi-label classification model structure schematic according to one embodiment. Referring to fig. 3, the multi-label classification model is specifically a multi-label deep neural network (deep neural network, DNN) model, taking a commodity sales scenario as an example, in a cold start stage, for each commodity, for each user, bidding randomly within a respective specified pricing interval; in the model training phase, data are first prepared: for each incoming user, recording the respective bids of various associated commodities and what kind of commodities the user finally buys; thus one record per user can be used as one training sample: user portrait features (gender, age, academic, consumption condition, etc) +scene features (current city, application residence time) +commodity prices (monthly card selling price, quaternary card selling price, half-year card selling price, annual card collection) respectively associated with the user, and the label is whether the user purchases the commodity (whether the monthly card, quaternary card, half-year card, annual card is purchased or not); the model is then trained: the collected sample data is used for training a multi-label DNN model, and the model effect is evaluated and optimized in a cross-validation mode. In the embodiment of the present specification, the price of a plurality of commodities are jointly put together as a feature, and the prediction of the model is regarded as a multi-label classification problem, wherein each label classification refers to whether a single commodity is purchased or not, for example, 0 is not purchased, and 1 is purchased.
Next, in step 23, the estimated resource benefit for each peer-to-peer resource amount combination is determined based on the estimated conversion rates of the various associated products for each peer-to-peer resource amount combination, the peer-to-peer resource amounts of the various associated products for each peer-to-peer resource amount combination, and the cost resource amounts of the various associated products. It is understood that the estimated resource benefit is the overall resource benefit of the plurality of associated products.
In one example, a given user predicts the revenue situation for that user at n price combinations:
for the trained model, the conversion rate of each commodity under each price combination can be estimated, so that the profit under the ith price combination can be calculatedi is the ith price combination, m is the total number of related commodities, k is the kth commodity, P i，k For the price of the kth commodity in the ith price combination, C k For the cost of the kth commodity, U is the characteristic of the known user, S is the characteristic of the known scene, R k (P i，k U, S) is the price P of the kth commodity in the ith price combination under the scene characteristics of the known user i，k And estimating the conversion rate.
Finally, in step 24, the peer-to-peer resource amount combination with the largest estimated resource benefit in the n peer-to-peer resource amount combinations is determined as the peer-to-peer resource amount finally combined and determined by the multiple associated products for the target user. It is understood that the peer-to-peer resource amount combination includes peer-to-peer resource amounts respectively corresponding to the plurality of associated products.
In one example, for a commodity sales scenario, the price combination with the largest return of the n price combinations is taken as the optimal price combination, and then the optimal price for each commodity can be obtained as a result of the final dynamic pricing for the user.
Firstly, acquiring user portrait characteristics of a target user and scene characteristics of the target user by the method provided by the embodiment of the specification; then taking the peer-to-peer resource quantity characteristics, the user portrait characteristics and the scene characteristics of each peer-to-peer resource quantity combination in n peer-to-peer resource quantity combinations of a plurality of associated products as inputs of a pre-trained neural network model, and obtaining the estimated conversion rate of each associated product in each peer-to-peer resource quantity combination in the n peer-to-peer resource quantity combinations through the output of the neural network model; then, according to the estimated conversion rate of various associated products under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of various associated products under each peer-to-peer resource amount combination and the cost resource amounts of various associated products, determining the estimated resource benefits under each peer-to-peer resource amount combination; and finally, determining the peer-to-peer resource amount combination with the maximum estimated resource benefit in the n peer-to-peer resource amount combinations as the peer-to-peer resource amount determined by the final combination of the plurality of associated products aiming at the target user. From the above, the embodiment of the specification can jointly consider the interaction of the peer-to-peer resource amounts of various associated products on the conversion rate of the products, so that the overall benefit maximization of the various associated products can be realized.
In addition, the method provided by the embodiment of the specification does not depend on subjective analysis, and is less in dependence on manual experience and operation rules; the information is mined from the historical behavior data of the user, and under the condition that the training data is sufficient, a better model can be obtained through training and optimization, and as the accumulated data is increased, the explored price combination is increased, and the accuracy is also improved continuously; meanwhile, the accuracy of model modeling can be measured through mature indexes, and the indexes are gradually improved through optimization iteration of the model, so that the pricing accuracy is further improved, and the modeling scheme can be optimized and iterated.
According to another aspect, there is further provided an apparatus for jointly determining an amount of peer-to-peer resources for a plurality of associated products, the apparatus being configured to perform the method for jointly determining an amount of peer-to-peer resources for a plurality of associated products provided in the embodiments of the present specification. FIG. 4 illustrates a schematic block diagram of an apparatus for jointly determining peer-to-peer resource amounts for multiple associated products, according to one embodiment. As shown in fig. 4, the apparatus 400 includes:
an acquisition unit 41 for acquiring a user portrait feature of a target user and a scene feature of the target user;
a first estimating unit 42, configured to take the peer-to-peer resource amount feature under each of n peer-to-peer resource amount combinations of multiple associated products, the user portrait feature and the scene feature acquired by the acquiring unit 41 as input of a pre-trained neural network model, and obtain, through output of the neural network model, estimated conversion rates of the various associated products under each of the n peer-to-peer resource amount combinations;
a second estimating unit 43, configured to determine estimated resource benefits under each peer-to-peer resource amount combination according to the estimated conversion rates of the various associated products under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of the various associated products under each peer-to-peer resource amount combination, and the cost resource amounts of the various associated products obtained by the first estimating unit 42;
a determining unit 44, configured to determine, from among the n peer-to-peer resource amount combinations, a peer-to-peer resource amount combination with the largest estimated resource benefit, as a peer-to-peer resource amount determined by the final union of the multiple association products for the target user.
Optionally, as an embodiment, the multiple associated products are cycle cards with multiple different durations of the same service.
Optionally, as an embodiment, the n peer-to-peer resource amount combinations are determined by:
sampling n times by adopting a random sampling method according to the peer-to-peer resource amount interval corresponding to each associated product, and obtaining an optional peer-to-peer resource amount corresponding to the associated product each time, wherein the optional peer-to-peer resource amounts of various associated products form a peer-to-peer resource amount combination; wherein, the value of n is a preset value.
Optionally, as an embodiment, the multiple associated products are simultaneously used as products to be converted, so that the user is allowed to convert the multiple associated products simultaneously, and the neural network model is a multi-label classification model.
Optionally, as an embodiment, the user can only convert one associated product when the multiple associated products are simultaneously used as products to be converted, and the neural network model is a multi-classification model.
Optionally, as an embodiment, the neural network model is trained by:
acquiring a training sample set, wherein each training sample comprises the peer-to-peer resource amount of each associated product facing a potential user, and the user portrait characteristic and scene characteristic of the potential user, and the conversion condition of the potential user on each associated product;
and training the neural network model by taking the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in at least a part of the training samples, the user portrait characteristic and the scene characteristic of the potential user as sample characteristics and the conversion condition of the potential user of each associated product in the training samples as sample labels.
Further, the acquiring a training sample set includes:
in a preset sample acquisition period, for potential users, randomly determining the peer-to-peer resource quantity of each associated product facing the potential users according to the peer-to-peer resource quantity interval corresponding to each associated product;
and generating a training sample of the training sample set according to the corresponding peer-to-peer resource quantity of each associated product faced by the potential user and the conversion condition of the potential user to each associated product under the peer-to-peer resource quantity.
Further, the apparatus further comprises:
and the verification unit is used for carrying out cross verification on the trained neural network model according to the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in another part of training samples, the user portrait characteristic and the scene characteristic of the potential user serving as sample characteristics and the conversion condition of the potential user of each associated product in the training sample serving as a sample label after training the neural network model.
Optionally, as an embodiment, the user portrait features include at least one of the following features:
gender, age, academic, consumption.
Optionally, as an embodiment, the scene feature includes at least one of the following features:
the current city, the application stay time.
With the apparatus provided in the embodiment of the present specification, first, the acquisition unit 41 acquires the user portrait characteristics of the target user and the scene characteristics of the target user; then, the first estimating unit 42 uses the peer-to-peer resource amount characteristics, the user portrait characteristics and the scene characteristics of each peer-to-peer resource amount combination in n peer-to-peer resource amount combinations of the plurality of associated products as inputs of a pre-trained neural network model, and obtains estimated conversion rates of the associated products of each peer-to-peer resource amount combination in the n peer-to-peer resource amount combinations through outputs of the neural network model; the second estimating unit 43 then determines estimated resource benefits under each peer-to-peer resource amount combination according to the estimated conversion rate of each associated product under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of each associated product under each peer-to-peer resource amount combination, and the cost resource amounts of each associated product; finally, the determining unit 44 determines a peer-to-peer resource amount combination with the largest estimated resource benefit from the n peer-to-peer resource amount combinations as a peer-to-peer resource amount determined by the final combination of the multiple associated products for the target user. From the above, the embodiment of the specification can jointly consider the interaction of the peer-to-peer resource amounts of various associated products on the conversion rate of the products, so that the overall benefit maximization of the various associated products can be realized.
According to an embodiment of another aspect, there is also provided a computer-readable storage medium having stored thereon a computer program which, when executed in a computer, causes the computer to perform the method described in connection with fig. 2.
According to an embodiment of yet another aspect, there is also provided a computing device including a memory having executable code stored therein and a processor that, when executing the executable code, implements the method described in connection with fig. 2.
Those skilled in the art will appreciate that in one or more of the examples described above, the functions described in the present invention may be implemented in hardware, software, firmware, or any combination thereof. When implemented in software, these functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium.
The foregoing embodiments have been provided for the purpose of illustrating the general principles of the present invention in further detail, and are not to be construed as limiting the scope of the invention, but are merely intended to cover any modifications, equivalents, improvements, etc. based on the teachings of the invention.
Claims (22)
1. A method of jointly determining an amount of peer-to-peer resources for a plurality of associated products, the method comprising:
acquiring user portrait features of a target user and scene features of the target user;
taking the peer-to-peer resource quantity characteristics, the user portrait characteristics and the scene characteristics of each of n peer-to-peer resource quantity combinations of a plurality of associated products as inputs of a pre-trained neural network model, and obtaining the estimated conversion rate of each associated product of each of the n peer-to-peer resource quantity combinations through the output of the neural network model;
determining estimated resource benefits under each peer-to-peer resource amount combination according to the estimated conversion rate of each associated product under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of each associated product under each peer-to-peer resource amount combination and the cost resource amounts of each associated product;
and determining the peer-to-peer resource amount combination with the maximum estimated resource benefit in the n peer-to-peer resource amount combinations as the peer-to-peer resource amount determined by the final combination of the plurality of associated products aiming at the target user.
2. The method of claim 1, wherein the plurality of associated products are cycle cards of a same service for a plurality of different durations.
3. The method of claim 1, wherein the n peer-to-peer resource amount combinations are determined by:
sampling n times by adopting a random sampling method according to the peer-to-peer resource amount interval corresponding to each associated product, and obtaining an optional peer-to-peer resource amount corresponding to the associated product each time, wherein the optional peer-to-peer resource amounts of various associated products form a peer-to-peer resource amount combination; wherein, the value of n is a preset value.
4. The method of claim 1, wherein the plurality of associated products are simultaneously used as products to be converted, the user is allowed to simultaneously convert the plurality of associated products therein, and the neural network model is a multi-label classification model.
5. The method of claim 1, wherein the user can only convert one associated product when the plurality of associated products are simultaneously used as products to be converted, and the neural network model is a multi-classification model.
6. The method of claim 1, wherein the neural network model is trained by:
acquiring a training sample set, wherein each training sample comprises the peer-to-peer resource amount of each associated product facing a potential user, and the user portrait characteristic and scene characteristic of the potential user, and the conversion condition of the potential user on each associated product;
and training the neural network model by taking the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in at least a part of the training samples, the user portrait characteristic and the scene characteristic of the potential user as sample characteristics and the conversion condition of the potential user of each associated product in the training samples as sample labels.
7. The method of claim 6, wherein the acquiring a training sample set comprises:
in a preset sample acquisition period, for potential users, randomly determining the peer-to-peer resource quantity of each associated product facing the potential users according to the peer-to-peer resource quantity interval corresponding to each associated product;
and generating a training sample of the training sample set according to the corresponding peer-to-peer resource quantity of each associated product faced by the potential user and the conversion condition of the potential user to each associated product under the peer-to-peer resource quantity.
8. The method of claim 6, wherein after the training of the neural network model, the method further comprises:
and cross-verifying the trained neural network model according to the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in another part of training samples, and the user portrait characteristic and the scene characteristic of the potential user as sample characteristics, and taking the conversion condition of the potential user of each training sample to each associated product as a sample label.
9. The method of claim 1, wherein the user representation features comprise at least one of the following:
gender, age, academic, consumption.
10. The method of claim 1, wherein the scene features comprise at least one of the following:
the current city, the application stay time.
11. An apparatus for jointly determining an amount of peer-to-peer resources for a plurality of associated products, the apparatus comprising:
the acquisition unit is used for acquiring user portrait features of a target user and scene features of the target user;
the first estimating unit is used for taking the peer-to-peer resource quantity characteristics under each peer-to-peer resource quantity combination in n peer-to-peer resource quantity combinations of a plurality of associated products, the user portrait characteristics and the scene characteristics acquired by the acquiring unit as inputs of a pre-trained neural network model, and obtaining the estimated conversion rate of each associated product under each peer-to-peer resource quantity combination in the n peer-to-peer resource quantity combinations through the output of the neural network model;
the second estimating unit is used for determining estimated resource benefits under each peer-to-peer resource amount combination according to the estimated conversion rate of each associated product under each peer-to-peer resource amount combination, the peer-to-peer resource amounts of each associated product under each peer-to-peer resource amount combination and the cost resource amounts of each associated product obtained by the first estimating unit;
and the determining unit is used for determining the peer-to-peer resource amount combination with the maximum estimated resource benefit in the n peer-to-peer resource amount combinations as the peer-to-peer resource amount determined by the final combination of the plurality of associated products aiming at the target user.
12. The apparatus of claim 11, wherein the plurality of associated products are cycle cards of a same service for a plurality of different durations.
13. The apparatus of claim 11, wherein the n peer-to-peer resource amount combinations are determined by:
sampling n times by adopting a random sampling method according to the peer-to-peer resource amount interval corresponding to each associated product, and obtaining an optional peer-to-peer resource amount corresponding to the associated product each time, wherein the optional peer-to-peer resource amounts of various associated products form a peer-to-peer resource amount combination; wherein, the value of n is a preset value.
14. The apparatus of claim 11, wherein the plurality of associated products are simultaneously as products to be converted, allowing a user to simultaneously convert the plurality of associated products therein, the neural network model being a multi-label classification model.
15. The apparatus of claim 11, wherein the user can only transform one associated product when the plurality of associated products are simultaneously used as products to be transformed, and the neural network model is a multi-classification model.
16. The apparatus of claim 11, wherein the neural network model is trained by:
acquiring a training sample set, wherein each training sample comprises the peer-to-peer resource amount of each associated product facing a potential user, and the user portrait characteristic and scene characteristic of the potential user, and the conversion condition of the potential user on each associated product;
and training the neural network model by taking the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in at least a part of the training samples, the user portrait characteristic and the scene characteristic of the potential user as sample characteristics and the conversion condition of the potential user of each associated product in the training samples as sample labels.
17. The apparatus of claim 16, wherein the obtaining a training sample set comprises:
in a preset sample acquisition period, for potential users, randomly determining the peer-to-peer resource quantity of each associated product facing the potential users according to the peer-to-peer resource quantity interval corresponding to each associated product;
and generating a training sample of the training sample set according to the corresponding peer-to-peer resource quantity of each associated product faced by the potential user and the conversion condition of the potential user to each associated product under the peer-to-peer resource quantity.
18. The apparatus of claim 16, wherein the apparatus further comprises:
and the verification unit is used for carrying out cross verification on the trained neural network model according to the peer-to-peer resource quantity of each associated product faced by the potential user of each training sample in another part of training samples, the user portrait characteristic and the scene characteristic of the potential user serving as sample characteristics and the conversion condition of the potential user of each associated product in the training sample serving as a sample label after training the neural network model.
19. The apparatus of claim 11, wherein the user representation features comprise at least one of the following:
gender, age, academic, consumption.
20. The apparatus of claim 11, wherein the scene features comprise at least one of the following:
the current city, the application stay time.
21. A computer readable storage medium having stored thereon a computer program which, when executed in a computer, causes the computer to perform the method of any of claims 1-10.
22. A computing device comprising a memory having executable code stored therein and a processor, which when executing the executable code, implements the method of any of claims 1-10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910660160.1A CN110428281B (en) | 2019-07-22 | 2019-07-22 | Method and device for jointly determining peer-to-peer resource quantity aiming at multiple associated products |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910660160.1A CN110428281B (en) | 2019-07-22 | 2019-07-22 | Method and device for jointly determining peer-to-peer resource quantity aiming at multiple associated products |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110428281A CN110428281A (en) | 2019-11-08 |
| CN110428281B true CN110428281B (en) | 2023-08-08 |
Family
ID=68411508
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910660160.1A Active CN110428281B (en) | 2019-07-22 | 2019-07-22 | Method and device for jointly determining peer-to-peer resource quantity aiming at multiple associated products |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110428281B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111339408B (en) * | 2020-02-19 | 2023-08-08 | 北京百度网讯科技有限公司 | Information recommendation method and device |
| CN111526058A (en) * | 2020-04-30 | 2020-08-11 | 北京三快在线科技有限公司 | Network resource allocation method, device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105205701A (en) * | 2015-09-22 | 2015-12-30 | 创点客（北京）科技有限公司 | Network dynamic pricing method and system |
| CN106127531A (en) * | 2016-07-14 | 2016-11-16 | 北京物思创想科技有限公司 | The method and system of differentiation price are performed based on machine learning |
| CN107818478A (en) * | 2017-10-31 | 2018-03-20 | 携程计算机技术（上海）有限公司 | Overseas reward voucher distribution method and system based on provisional profit |
| JP2018073389A (en) * | 2016-10-26 | 2018-05-10 | 株式会社デンソー | Data processing device and data processing method |
| CN108460618A (en) * | 2018-01-09 | 2018-08-28 | 北京三快在线科技有限公司 | A kind of resource allocation method and device, electronic equipment |
| CN109993564A (en) * | 2018-01-02 | 2019-07-09 | 北京奇虎科技有限公司 | Methods of exhibiting, device and the computer readable storage medium of commodity price |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7970713B1 (en) * | 2000-05-10 | 2011-06-28 | OIP Technologies, Inc. | Method and apparatus for automatic pricing in electronic commerce |
-
2019
- 2019-07-22 CN CN201910660160.1A patent/CN110428281B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105205701A (en) * | 2015-09-22 | 2015-12-30 | 创点客（北京）科技有限公司 | Network dynamic pricing method and system |
| CN106127531A (en) * | 2016-07-14 | 2016-11-16 | 北京物思创想科技有限公司 | The method and system of differentiation price are performed based on machine learning |
| JP2018073389A (en) * | 2016-10-26 | 2018-05-10 | 株式会社デンソー | Data processing device and data processing method |
| CN107818478A (en) * | 2017-10-31 | 2018-03-20 | 携程计算机技术（上海）有限公司 | Overseas reward voucher distribution method and system based on provisional profit |
| CN109993564A (en) * | 2018-01-02 | 2019-07-09 | 北京奇虎科技有限公司 | Methods of exhibiting, device and the computer readable storage medium of commodity price |
| CN108460618A (en) * | 2018-01-09 | 2018-08-28 | 北京三快在线科技有限公司 | A kind of resource allocation method and device, electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110428281A (en) | 2019-11-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Greenstein-Messica et al. | Personal price aware multi-seller recommender system: Evidence from eBay | |
| Kumar et al. | Customer relationship management: A databased approach | |
| US10181138B2 (en) | System and method for determining retail-business-rule coefficients from current prices | |
| US20140164220A1 (en) | Payment instrument selection | |
| Chevalier et al. | Best prices | |
| KR101807965B1 (en) | Method, apparatus and system for additional random discount after payment in e-commerce in the open market | |
| Fuchs et al. | Successfully selling accommodation packages at online auctions–The case of eBay Austria | |
| US11321724B1 (en) | Product evaluation system and method of use | |
| CN110347924A (en) | Fruits and vegetables market management system and fruit-vegetable information method for pushing | |
| CN110428281B (en) | Method and device for jointly determining peer-to-peer resource quantity aiming at multiple associated products | |
| CN115099878B (en) | Marketing method based on big data analysis | |
| US11900404B2 (en) | Guaranteed offer generator | |
| CN110135994B (en) | Transaction processing method for time sensitive data | |
| CN110866220A (en) | Selection recommendation system in e-commerce activity process | |
| CN111178947A (en) | Advertisement space recommendation method and device, computer-readable storage medium and electronic equipment | |
| KR101865521B1 (en) | Pricing method of online sale product | |
| US20200104892A1 (en) | Group reverse auction shopping method and implementation on a computer network | |
| Soni et al. | Optimal pricing and replenishment policy for non-instantaneous deteriorating items with varying rate of demand and partial backlogging | |
| US20210264461A1 (en) | System and method for managing business loyalty and reward programs | |
| JP2002329097A (en) | Price calculation method and its device for interpersonal transaction system | |
| Latha et al. | A STUDY ON CUSTOMER SATISFACTION TOWARDS ONLINE SHOPPING. | |
| CN110751492A (en) | High-value crowd identification method and device | |
| Joseph et al. | A study on consumer's perception towards online shopping among college students | |
| CN110288379A (en) | Conversion price expectation method, apparatus, storage medium and the computer equipment of advertisement | |
| CN116883121B (en) | POS machine user recommendation method based on big data analysis |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200925 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Advanced innovation technology Co.,Ltd. Address before: A four-storey 847 mailbox in Grand Cayman Capital Building, British Cayman Islands Applicant before: Alibaba Group Holding Ltd. Effective date of registration: 20200925 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant after: Innovative advanced technology Co.,Ltd. Address before: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman Islands Applicant before: Advanced innovation technology Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |