CN110264301A - Recommended method, device, electronic equipment and non-volatile memory medium - Google Patents
Recommended method, device, electronic equipment and non-volatile memory medium Download PDFInfo
- Publication number
- CN110264301A CN110264301A CN201910389818.XA CN201910389818A CN110264301A CN 110264301 A CN110264301 A CN 110264301A CN 201910389818 A CN201910389818 A CN 201910389818A CN 110264301 A CN110264301 A CN 110264301A
- Authority
- CN
- China
- Prior art keywords
- resource
- target user
- user
- lower single
- recommended
- 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
- G06Q30/0207—Discounts or incentives, e.g. coupons or rebates
-
- 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]
- G06Q30/0623—Item investigation
- G06Q30/0625—Directed, with specific intent or strategy
-
- 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]
- G06Q30/0631—Item recommendations
-
- 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]
- G06Q30/0641—Shopping interfaces
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present embodiments relate to technical field of information processing, a kind of recommended method, device, electronic equipment and non-volatile memory medium are disclosed.Above-mentioned recommended method includes: that target user was played list and recommended candidate collection is added in the resource for meeting preset price request according to the History Order of target trade company；According to the prediction model for estimating lower single probability, lower single probability that the resource that the recommended candidate is concentrated is placed an order by the target user is estimated；Wherein, the prediction model is obtained previously according to the user of acquisition and the history feature data training of resource；The resource that the recommended candidate is concentrated recommends the target user according to the height for the lower single probability estimated, can satisfy the individual demand of different user.
Description
Technical field
The present invention relates to technical field of information processing, in particular to a kind of recommended method, device, electronic equipment and non-volatile
Property storage medium.
Background technique
With the fast development of Internet technology, application Internet-based is more and more, such as takes out class application, shopping
Class application.Based on these applications, user stays indoors the article that can be obtained needed for oneself, for user life provide it is more next
More convenience.How the article for more meeting user intention, the increasingly concern by major internet recommended for user.
Inventors have found that in taking out class application carrying out that the way of vegetable is recommended to be normally based on the whole network user to user
The combined information bought, it is dynamic in conjunction with current vegetable price and full deactivation, it is arranged according to combined purchase frequency descending.So
And this recommended method is a kind of suggested design of thousand people's one sides, is unable to satisfy the demand of different user personalization.
Summary of the invention
Embodiment of the present invention is designed to provide a kind of recommended method, device, electronic equipment and non-volatile memories
Medium can satisfy the individual demand of different user.
In order to solve the above technical problems, embodiments of the present invention provide a kind of recommended method, comprising: according to target quotient
The History Order at family, played list for target user and recommended candidate collection is added in the resource for meeting preset price request；According to
In the prediction model for estimating lower single probability, estimates the resource that the recommended candidate is concentrated and placed an order generally by what the target user placed an order
Rate；Wherein, the prediction model is obtained previously according to the user of acquisition and the history feature data training of resource；By the recommendation
Resource in Candidate Set recommends the target user according to the height for the lower single probability estimated.
Embodiments of the present invention additionally provide a kind of recommendation apparatus, comprising: processing module, for according to target trade company
History Order, played list for target user and recommended candidate collection is added in the resource for meeting preset price request；Module is estimated, is used
According to the prediction model for estimating lower single probability, estimates the resource that the recommended candidate is concentrated and placed an order by the target user
Lower single probability；Wherein, the prediction model is obtained previously according to the user of acquisition and the history feature data training of resource；It pushes away
Module is recommended, the resource for concentrating the recommended candidate is recommended the target according to the height for the lower single probability estimated and used
Family.
Embodiments of the present invention additionally provide a kind of electronic equipment, including memory and processor, memory storage meter
Calculation machine program, processor execute when running program: according to the History Order of target trade company, descending single by target user and met pre-
If price request resource be added recommended candidate collection；According to the prediction model for estimating lower single probability, the recommendation is estimated
Lower single probability that resource in Candidate Set is placed an order by the target user；Wherein, use of the prediction model previously according to acquisition
The training of the history feature data of family and resource obtains；The resource that the recommended candidate is concentrated, according to the lower single probability estimated
Height recommends the target user.
Embodiments of the present invention additionally provide a kind of non-volatile memory medium, for storing computer-readable program,
The computer-readable program is used to execute recommended method as described above for computer.
In terms of existing technologies, the main distinction and its effect are embodiment of the present invention: according to target trade company
History Order, played list for target user and recommended candidate collection is added in the resource for meeting preset price request.Since target is used
Single resource was descended at family, can characterize lower single wish of target user to a certain extent, while if the resource is also able to satisfy
Preset price request, then the resource is very big a possibility that lower single again by target user.Therefore, target user was descended singly
And the resource that meets preset condition recommended candidate collection is added, be conducive to obtain the recommended candidate collection for target user, thus
Effective personalized recommendation can be carried out to target user.According to the prediction model for estimating lower single probability, estimates recommendation and wait
Lower single probability that resource in selected works is placed an order by target user is conducive to quickly and accurately estimate lower single probability, and
Since prediction model is obtained previously according to the user of acquisition and the history feature data training of resource, that is, train the number of prediction model
According to the real history data for deriving from user and resource, reference value is high, may make the lower single probability estimated it is more accurate, can
It leans on.The resource that recommended candidate is concentrated recommends target user according to the height for the lower single probability estimated, so that target user is most
First being recommended is the highest resource of lower single possibility, improves the efficiency that target user places an order in target trade company, while can also
Promote the usage experience of target user.Also, since History Order of the different target users in same target trade company may
It is not identical, therefore, be suitble to respective recommended candidate collection with having per family for different target, be conducive to different target user into
Row meets the individual demand of different target user when recommending.
In addition, the resource that the recommended candidate is concentrated is divided into: single resource and combination of resources；Estimate the recommended candidate collection
In lower single probability for being placed an order by the target user of combination of resources, comprising: estimate each of described combination of resources respectively
Lower single probability that resource is placed an order by the target user；According to lower single probability of each resource, the resource group is estimated
Close the lower single probability to be placed an order by the target user.Target user descended the resource of price request that is single and meeting target user
It may also may be combination of resources for single resource, for lower single probability of combination of resources, provide a kind of specific side of estimating
Formula is conducive to accurately estimate lower single probability of combination of resources.
In addition, estimating the combination of resources according to lower single probability of each resource and being placed an order by the target user
Lower single probability, comprising: according to lower single probability of each resource, calculate lower single probability average；It places an order described generally
Rate average value, lower single probability as the combination of resources estimated.Lower list probability average can be balanced consideration to resource group
A possibility that each resource is by lower list in conjunction, so that the lower single probability for the combination of resources finally estimated is more rationally, accurately.
In addition, the resource that recommended candidate is concentrated is the current resource on sale of the target trade company, be conducive to improve candidate
The validity of the resource of concentration guarantees that the resource for recommending user is resource that is currently true, effective, can placing an order, is conducive to mention
The usage experience of high user.
In addition, the history feature data of the user and resource include: user property feature, Resource Properties feature, user
With the cross attribute feature of resource, the real-time behavioural characteristic of user.In view of the feature of multiple dimensions when training prediction model, especially
It is the cross attribute feature of user and resource, so that better reflecting the user to the reality of the resource for lower single estimating for probability
Border fancy grade, to further increase the accuracy estimated.
In addition, the cross attribute feature of the user and resource, comprising: user is to the history evaluation of resource, user to money
The purchase volume in source.User can accurately measure the possibility that user places an order to current vegetable to the history evaluation and purchase volume of resource
Property, be conducive to improve the accuracy estimated.
In addition, the prediction model is the prediction model according to XGBoost training.XGBoost can increase prediction model
Robustness, compared to traditional machine learning algorithm, speed is fast, effect is good, can handle large-scale data and supports a variety of languages
It makes peace customized loss function.
Detailed description of the invention
Fig. 1 is the flow chart of the recommended method in first embodiment according to the present invention；
Fig. 2 is the flow chart of the realization process of the step S102 in second embodiment according to the present invention；
Fig. 3 is the schematic diagram of the recommendation apparatus in third embodiment according to the present invention；
Fig. 4 is the structural schematic diagram for the electronic equipment that the 4th embodiment provides according to the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with attached drawing to the present invention
Each embodiment be explained in detail.However, it will be understood by those skilled in the art that in each embodiment party of the present invention
In formula, in order to make the reader understand this application better, many technical details are proposed.But even if without these technical details
And various changes and modifications based on the following respective embodiments, the application technical solution claimed also may be implemented.With
Under the division of each embodiment be for convenience, any restriction should not to be constituted to specific implementation of the invention, it is each
Embodiment can be combined with each other mutual reference under the premise of reconcilable.
The first embodiment of the present invention is related to a kind of recommended methods, are applied to electronic equipment, which can be
Server, such as, if the vegetable that the recommended method of present embodiment is applied to take out on platform is recommended, above-mentioned server
The take-away Platform Server of platform can be taken out for management.It is clear that the recommendation of present embodiment predominantly clicks to enter trade company in user
When looking at, recommend the resource for the price request for meeting the user, the commodity that resource can sell for trade company, different type for the user
Trade company sell different types of commodity, such as, taking out the commodity that the trade company of type sells is vegetable.Below to this embodiment party
The realization details of the recommended method of formula is specifically described, and the following contents is only for convenience of the realization details provided is understood, not
Implement the necessary of this programme.
It should be noted that the recommended method in present embodiment is carried out specifically for being applied to take out Platform Server
Illustrate, but is not limited thereto in practical applications.The flow chart of the recommended method of present embodiment can be as shown in Figure 1, packet
It includes:
Target user was played list and met preset price request by step S101 according to the History Order of target trade company
Resource be added recommended candidate collection.
Specifically, target user can recommend the user of resource request for initiation, and target user initiates that resource is recommended to ask
The mode asked can be with are as follows: the interface of target trade company can have the virtual push button for recommending resource for trigger request, press above-mentioned
The user of virtual push button is target user, such as, above-mentioned virtual push button can be shown on the mobile phone interface of target trade company, when
When mobile phone detects that virtual push button is pressed, mobile phone sends the recommendation resource request of target user to Platform Server is taken out.Or
The user for issuing the voice messaging is considered as target user when receiving for requesting to recommend the voice messaging of resource by person.Money
Source can for target trade company sell commodity, such as may include: various dishes, beverage, rice, cake, etc..
Target trade company in present embodiment can provide different price requests, and different price requests can be difference
Expire the condition of subtracting, such as full 30 subtract 15, and full 50 subtract 25, and full 70 subtract 40.What target trade company provided, which expires the condition of subtracting, to show
In the interface homepage of the target trade company, independently selected by target user.Preset price request is target user in target trade company
There is provided different price requests in select it is one or more, target user initiate recommendation resource request can carry selection
Price request.Server can be inquired in the History Order of target trade company according to the price request that user selects, and look into
It askes and meets the selected price request of target user and descended single resource for target user.The resource inquired addition is pushed away
Recommend Candidate Set.That is, the resource that recommended candidate is concentrated is added being in order that target user once descend, and expire
The resource of the selected price request of foot-eye user.
In one example, meet price if target user descended in single resource there are the price of single resource to want
It asks, then recommended candidate collection is added in the single resource for meeting price request.
In one example, it is contemplated that the current resource on sale of target trade company would generally be by factors such as season, weather
Influence, such as, target trade company winter resource on sale spring may undercarriage will not sell, alternatively, target trade company because
The source of goods lacks problem, the last week resource also on sale, in current weekly assembly undercarriage.Therefore, in order to avoid the above problem, it can be ensured that
It is the current resource on sale of target trade company that recommended candidate, which concentrates the resource being added,.That is, the history in target trade company is ordered
It is inquired in list and meets the selected price request of target user and after the resource in the order that descended for target user, judgement is looked into
Whether the resource ask is resource that target trade company currently sells, is determining that the resource inquired is that target trade company is currently on sale
Resource when, recommended candidate is added in the determining resource and is concentrated.
Step S102 estimates the resource of recommended candidate concentration by target according to the prediction model for estimating lower single probability
Lower single probability that user places an order.
Specifically, prediction model is obtained previously according to the user of acquisition and the history feature data training of resource.User
History feature data with resource include: user property feature, Resource Properties feature, user and resource cross attribute feature,
The real-time behavioural characteristic of user.Wherein, user property feature can be with are as follows: gender, age, place city, taste information etc..Resource category
Property feature can be with are as follows: conversion ratio, sales volume, positive rating of resource etc., wherein the sales volume of resource can be understood as should in a period of time
The amount of placing an order of resource, the amount of placing an order that the conversion ratio of resource can be understood as the resource in a period of time account for the percentage of pageview,
The favorable comment quantity that the positive rating of resource can be understood as the resource in a period of time accounts for the percentage of all evaluation quantity.User with
The cross attribute feature of resource can be with are as follows: user is to the history evaluation of resource, user to purchase volume of resource etc..The real-time row of user
Being characterized can be with are as follows: and user is in real-time behavioural characteristic at that time in history, such as, it is different daily in past a period of time
The browsing feature of moment user, taste, the average price of resource etc. browsed including user.It should be noted that above-mentioned
" a period of time " can be configured according to actual needs, in this regard, present embodiment is not specifically limited.
The following are according to the user of acquisition and the history feature data of resource, training obtains the substantially method of prediction model:
Firstly, selection training sample；That is, choosing user's resource browsed but the resource not placed an order work in a period of time
For negative sample, and chooses and browsed in this period and the resource that places an order is as positive sample.
Secondly, selection sample characteristics；I.e., it is possible to choose history feature data relevant to user and resource, such as above-mentioned
User property feature, Resource Properties feature, user and the cross attribute feature of resource, the real-time behavioural characteristic of user etc..
Finally, sample training；Sample training is carried out namely based on training sample and sample characteristics, such as passes through machine learning frame
Frame XGBoost carries out off-line training, obtains prediction model；Wherein, the input of prediction model is that target user and resource are relevant
Characteristic exports the lower single probability to be placed an order for resource by target user.
In one example, after training obtains prediction model, prediction model can be updated at regular intervals.
The estimated result obtained using prediction model can also be compared with actual result, so that the parameter to prediction model carries out
Adjustment, such as, the parameter of prediction model can be adjusted by increasing sample data volume or increasing frequency of training,
So that the estimated result using prediction model is more accurate.
In present embodiment, recommended candidate concentrate resource may have it is multiple, therefore can successively to recommended candidate concentrate
Lower single probability for being placed an order by target user of each resource estimate.Specifically, to estimate the resource A of recommended candidate concentration
It is illustrated for lower single probability that target user places an order:
It is possible, firstly, to first obtain the attributive character of target user, the real-time behavioural characteristic of target user, resource A attribute
The cross attribute feature of feature and target user and resource A builds the real-time forecast sample of resource A.
It then, can be using the real-time forecast sample of the resource A of building as the input of the prediction model of training, by estimating
Lower single probability that the resource A of model output estimation is placed an order by target user.
In addition, other resources concentrated for recommended candidate can be successively complete according to the above-mentioned mode of estimating for resource A
At estimating, target user is obtained to lower single probability of each of recommended candidate collection resource.
Step S103, the resource that recommended candidate is concentrated recommend target user according to the height for the lower single probability estimated.
Specifically, it since each of recommended candidate collection resource has corresponding lower single probability, can will push away
The resource recommended in Candidate Set is ranked up according to the height of lower single probability, and the result of sequence is showed target user.Such as
Can being checked in an interface for target user as the result is shown by sequence, can also be notified by way of voice broadcast to
Target user.
In one example, top n resource shape therein can be selected according to the height of lower single probability of each resource
At list, target user is recommended.N is natural number more than or equal to 1, and the size of N can be specified by target user, can also be with
For according to the pre-set default value of actual needs.
For selling scene in addition, it is to take out the method for recommending vegetable under scene that present embodiment, which recommends the method for resource,.
Assuming that target user A clicks to enter target trade company B, the price request that target trade company B is provided includes: to expire 30 to subtract 15, and full 50 subtract 25,
Full 70 three types that subtract 40.The recommendation vegetable request that target user A is initiated can carry full 30 and subtract 15 this completely subtract condition, take out
Platform Server can search mesh in the History Order of target trade company B after the recommendation vegetable request for receiving target user A
The order that user A was descended is marked, is filtered out in the order then crossed at target user A, satisfaction full 30 subtracts 15, and this completely subtracts condition
Vegetable, be added recommended candidate collection.Assuming that the title and price for the vegetable that recommended candidate is concentrated is added, it is as shown in table 1 below:
Table 1
| Menu name | Pungent steamed rice in clay pot | Chicken with several spices leg steamed rice in clay pot | Full element steamed rice in clay pot | Pork braised in brown sauce steamed rice in clay pot | Steamed dumpling |
| Price (member) | 35 | 36 | 31 | 40 | 30 |
The price of single vegetable in table 1 is all satisfied preset price request i.e. completely 30 and subtracts 15, so the vegetable in table 1 is all
For the single resource for meeting preset price request.Below to estimate " the pungent steamed rice in clay pot " of recommended candidate concentration by target user
It is specifically described for lower single probability that A places an order:
It is used firstly, obtaining the attributive character of target user A, real-time behavioural characteristic, the attributive character of pungent steamed rice in clay pot, target
The cross attribute feature of family A and pungent steamed rice in clay pot, as forecast sample.Wherein, the attributive character of target user A are as follows: gender female,
25 years old age, place city Beijing, the inclined sweet tea of taste etc..The real-time behavioural characteristic of target user A are as follows: the mouth for the vegetable currently seen
The inclined sweet tea of taste, the vegetable currently seen average price between 25 yuan to 30 yuan.The attributive character of pungent steamed rice in clay pot are as follows: two months
Interior conversion ratio is 60%, sales volume 1000, positive rating 65%.The cross attribute feature of target user A and pungent steamed rice in clay pot are as follows:
Target user A is 15 times to the purchase volume of pungent steamed rice in clay pot in two months, and target user A is to the positive rating of pungent steamed rice in clay pot
92%.It should be noted that in this example, conversion ratio, sales volume, positive rating, purchase volume reference time be only with two months
Example, is not limited thereto in practical applications.
Secondly, the forecast sample for pungent steamed rice in clay pot that will acquire exports pungent and stews son as the input of prediction model
Lower single probability that meal is placed an order by target user A.
Then, the forecast sample for successively obtaining other vegetables concentrated for recommended candidate, as the input of prediction model,
It is sequentially output lower single probability that other vegetables are placed an order by target user A.Placing an order for the different vegetables estimated in present embodiment is general
Rate can be as shown in table 2 below:
Table 2
| Menu name | Pungent steamed rice in clay pot | Chicken with several spices leg steamed rice in clay pot | Full element steamed rice in clay pot | Pork braised in brown sauce steamed rice in clay pot | Steamed dumpling |
| Lower list probability | 60% | 62% | 50% | 70% | 40% |
According to the lower single probability for each vegetable that the recommended candidate estimated is concentrated, ranking is carried out to each vegetable, is formed and is recommended
The recommendation list of formation is recommended target user A by list, and the recommendation list of formation can be as shown in table 3 below:
Table 3
| Menu name | Ranking |
| Pork braised in brown sauce steamed rice in clay pot (70%) | 1 |
| Chicken with several spices leg steamed rice in clay pot (62%) | 2 |
| Pungent steamed rice in clay pot (60%) | 3 |
| Full element steamed rice in clay pot (50%) | 4 |
| Steamed dumpling (40%) | 5 |
Compared with prior art, in present embodiment, according to the History Order of target trade company, by target user descended it is single and
Recommended candidate collection is added in the resource for meeting preset price request.Since target user descended single resource, can be in certain journey
Lower single wish of target user is characterized on degree, while if the resource is also able to satisfy the price request of target user, the money
A possibility that source is lower single again by target user is very big.Therefore, target user was descended into resource that is single and meeting preset condition
Recommended candidate collection is added, is conducive to obtain the recommended candidate collection for target user, it is effective so as to be carried out to target user
Personalized recommendation.According to the prediction model for estimating lower single probability, the resource of recommended candidate concentration is estimated by target user
The lower single probability to place an order is conducive to accurately estimate lower single probability, and since prediction model is previously according to acquisition
The training of the history feature data of user and resource obtains, that is, the data source of prediction model is trained to go through in user and the true of resource
History data, reference value is high, may make the lower single probability estimated more accurate, reliable.The resource that recommended candidate is concentrated, according to
The height for the lower single probability estimated recommends target user, so that it is that lower single possibility is highest that target user is recommended at first
Resource improves the efficiency that target user places an order in target trade company, while can also promote the usage experience of target user.Also,
Since History Order of the different target users in same target trade company may be not identical, it is used for different target
Have per family and be suitble to respective recommended candidate collection, is conducive to for meeting different target user when recommending different target user
Property demand.
Second embodiment of the present invention is related to a kind of recommended method, present embodiment and first embodiment substantially phase
Together, the difference is that, the resource in resource Candidate Set is distinguished in present embodiment, is concentrated when recommended candidate
When resource is single resource, in the method and first embodiment of estimating lower single probability that the single resource is placed an order by target user
Roughly the same, this is no longer going to repeat them.It is mainly introduced in present embodiment, when the resource that recommended candidate is concentrated is combination of resources
When, the lower single probability how to be placed an order to the combination of resources by target user is estimated.
In present embodiment, if target user descended the price of single single resource to be unsatisfactory for price request, basis
The price and attribute of each single resource for being unsatisfactory for price request generate combination of resources, wherein the price of combination of resources meets
Recommended candidate collection is added in combination of resources by price request.If resource, by taking food product as an example, combination of resources is different list
The food product combination of a food product composition, food product combination can be according to preset price requests, the unit price and category of each single food product
Property determine, the attribute of food product can be vegetables, staple food, drink class etc., i.e. different type belonging to the single food product.It protects as far as possible
It preferably can also include drink class such as soup or drink both including vegetables or including staple food in each food product combination that card generates
Material.The price of each single food product, which adds up, in the food product combination of generation meets preset price request.
Such as preset price request subtracts 15 for full 30, it includes Sichuan-style pork (unit price in single food product that target user, which descended,
16 yuan), cooking shredded potato (12 yuan of unit price), rice (2 yuan of unit price), fish-flavoured shredded pork (15 yuan of unit price), deep fried chicken cube with chili (14 yuan), egg flower soup (5
Member), Hot and sour cabbage (12 yuan of unit price).It can be seen that single food product is not satisfied preset price and wants in the order that user descended
It asks to expire and 30 subtracts 15, therefore 30 can be met to descending single each single food product to be combined and subtract 15 this completely subtracts item
Part.The food product combination for descending single food product to obtain according to above-mentioned target user can be with are as follows: combination 1: Sichuan-style pork+cooking shredded potato+rice
(10+8+2=30 member)；Combination 2: fish-flavoured shredded pork+deep fried chicken cube with chili+rice (15+14+2=31)；Combination 3: cooking shredded potato+Hot and sour cabbage+
Egg flower soup+rice (12+12+5+2=31).
In present embodiment, estimate to lower single probability that combination of resources is placed an order by target user that can be used as first real
Apply in mode that " step S102 estimates the resource of recommended candidate concentration by target according to the prediction model for estimating lower single probability
The sub-step of lower single probability that user places an order ", flow chart is as shown in Fig. 2, specifically include:
Step S201 estimates lower single probability that each of combination of resources resource is placed an order by target user respectively.
Specifically, resource single for each of combination of resources can carry out placing an order according to prediction model general
Rate is estimated.Such as the probability that above-mentioned " combination 1: Sichuan-style pork+cooking shredded potato+rice " is placed an order by target user is estimated, it can be with
Estimate lower single probability that Sichuan-style pork is placed an order by target user in combination 1 respectively, lower single probability that cooking shredded potato is placed an order by target user,
Lower single probability that rice is placed an order by target user.The estimating for lower single probability of each single food product can be implemented with reference to first
The method introduced in mode, to avoid repeating, this is no longer going to repeat them.
Step S202 estimates combination of resources and is placed an order generally by what target user placed an order according to lower single probability of each resource
Rate.
Specifically, lower single probability average can be calculated according to lower single probability of each resource；Lower single probability is put down
Mean value, lower single probability as the combination of resources estimated.The balanced consideration of lower list probability average energy is each into combination of resources
A possibility that a resource is by lower list, so that the lower single probability for the combination of resources finally estimated is more rationally, accurately.
It in one example, can also be using lower single probability of the single resource of list maximum probability lower in combination of resources as pre-
Lower single probability that the combination of resources estimated is placed an order by target user, or the smallest single resource of list probability will be played in combination of resources
Lower single probability that lower list probability is placed an order as the combination of resources estimated by target user.
For selling scene other than same, with reference to the example in first embodiment, mesh is still clicked to enter with target user A
For marking trade company B, the recommendation vegetable request that target user A is initiated carries full 30 and subtracts 15 this completely subtracts condition, takes out platform clothes
Business device can search target user in the History Order of target trade company B after the recommendation vegetable request for receiving target user A
The order that A was descended filters out satisfaction full 30 in the order then crossed at target user A and subtracts 15 this completely subtracts the food product of condition,
Recommended candidate collection is added.Assuming that be added recommended candidate concentrate food product other than single vegetable as listed in Table 1, further include as
The combination of food product shown in the following table 4, that is to say, that the resource that the recommended candidate in this example is concentrated both had included list as shown in Table 1
A vegetable also includes that food product as shown in table 4 combines, and the single food product in food product combination is unsatisfactory for full 30 and subtracts 15 this completely subtracts item
Part is combined to obtain combination 1 in table 4, combination 2 and combination 3 to the single food product that this completely subtracts condition is unsatisfactory for, obtains
The price of three kinds of combinations is all satisfied full 30 conditions for subtracting 15.
Table 4
It is specifically described for estimating lower single probability that combination 1 is placed an order by target user A below:
Firstly, estimating lower single probability that Sichuan-style pork is placed an order by target user in combination 1 respectively, cooking shredded potato is by under target user
Single lower single probability, lower single probability that rice is placed an order by target user.Assuming that each single food product is by target user in combination 1
The lower single probability to place an order is as shown in table 5 below:
Table 5
| Single food product | Sichuan-style pork | Cooking shredded potato | Rice |
| Lower list probability | 61% | 72% | 80% |
Then, using the average value on lower single probability of three in table 5 single food products as lower single probability of combination 1, i.e. group
The lower single probability for closing 1 is (61%+72%+80%)/3=71%.Mode point is estimated according to above-mentioned lower single probability to combination 1
Other lower single probability to combination 2 and combination 3 is estimated, and the result estimated can be as shown in table 6 below:
Table 6
Then, after learning lower single probability of the single vegetable that recommended candidate is concentrated and food product combination, to each single vegetable
It is ranked up with lower single probability of each food product combination, forms recommendation list as shown in table 7, looked into recommending user for user
It sees.
Table 7
| Menu name | Ranking |
| Combine 1 (71%) | 1 |
| Pork braised in brown sauce steamed rice in clay pot (70%) | 2 |
| Combine 2 (65%) | 3 |
| Chicken with several spices leg steamed rice in clay pot (62%) | 4 |
| Pungent steamed rice in clay pot (60%) | 5 |
| Full element steamed rice in clay pot (50%) | 6 |
| Combine 3 (45%) | 7 |
| Steamed dumpling (40%) | 8 |
In one example, the vegetable that directly according to the ranking in table 7, can also choose before ranking 5 recommends target use
Family A in this example only for before to choose ranking 5, but is not limited thereto in practical applications.
Compared with prior art, in present embodiment, the resource concentrated to recommended candidate is distinguished, and provides one
The specific implementation for lower single probability that kind is placed an order to combination of resources by target user, so that for single resource and combination of resources
Lower single probability estimate, have and respectively targetedly estimate mode, be conducive to further to the single resource of recommended candidate concentration
With lower single probability of combination of resources carry out it is more accurate estimate, to preferably recommend user.
Third embodiment of the present invention is related to a kind of recommendation apparatus, as shown in figure 3, the device includes: processing module
301, for the History Order according to target trade company, target user was descended into list and the resource for meeting preset price request is added
Recommended candidate collection；Module 302 is estimated, for estimating the recommended candidate collection according to the prediction model for estimating lower single probability
In lower single probability for being placed an order by the target user of resource；Wherein, user and money of the prediction model previously according to acquisition
The history feature data training in source obtains；Recommending module 303, the resource for concentrating the recommended candidate, according to what is estimated
The height of lower list probability recommends the target user.
In one example, the resource that the recommended candidate is concentrated is divided into: single resource and combination of resources；Estimate module
302 estimate lower single probability that the combination of resources that the recommended candidate is concentrated is placed an order by the target user, comprising: estimate institute respectively
State lower single probability that each of combination of resources resource is placed an order by the target user；
According to lower single probability of each resource, estimates the combination of resources and placed an order by what the target user placed an order
Probability.
In one example, module 302 is estimated according to lower single probability of each resource, estimates the combination of resources
The lower single probability to be placed an order by the target user, comprising: according to lower single probability of each resource, it is flat to calculate lower single probability
Mean value；Lower single probability by lower single probability average, as the combination of resources estimated.
In one example, processing module 301, if descending the valence of single single resource specifically for the target user
Lattice meet the price request, then recommended candidate collection are added in the single resource for meeting the price request；If the target
User descended the price of single single resource to be unsatisfactory for the price request, then according to each list for being unsatisfactory for the price request
The price and attribute of a resource generate combination of resources, wherein the price of the combination of resources meets the price request, will be described
The recommended candidate collection is added in combination of resources.
In one example, the resource that the recommended candidate is concentrated is the current resource on sale of the target trade company.
In one example, the history feature data of the user and resource include: user property feature, Resource Properties spy
Sign, user and the cross attribute feature of resource, the real-time behavioural characteristic of user.The cross attribute feature of the user and resource, packet
Include: user is to the history evaluation of resource, user to the purchase volume of resource.
In one example, the prediction model is the prediction model according to XGBoost training.
It is not difficult to find that present embodiment is Installation practice corresponding with the first and second embodiment, present embodiment can
It works in coordination implementation with the first and second embodiment.The relevant technical details mentioned in first and second embodiment are in present embodiment
In still effectively, in order to reduce repetition, which is not described herein again.Correspondingly, the relevant technical details mentioned in present embodiment
It can be applicable in the first and second embodiment.
It is noted that each module involved in present embodiment is logic module, and in practical applications, one
A logic unit can be a physical unit, be also possible to a part of a physical unit, can also be with multiple physics lists
The combination of member is realized.In addition, in order to protrude innovative part of the invention, it will not be with solution institute of the present invention in present embodiment
The technical issues of proposition, the less close unit of relationship introduced, but this does not indicate that there is no other single in present embodiment
Member.
Compared with prior art, in present embodiment, processing module is according to the History Order of target trade company, by target user
It played list and met the resource addition recommended candidate collection of preset price request.Since target user descended single resource, can
Lower single wish of target user is characterized to a certain extent, while if the resource is also able to satisfy the price request of target user,
A possibility that so resource is lower single again by target user is very big.Therefore, target user was descended single and met default item
Recommended candidate collection is added in the resource of part, is conducive to obtain the recommended candidate collection for target user, so as to target user
Carry out effective personalized recommendation.Module is estimated according to the prediction model for estimating lower single probability, estimates recommended candidate concentration
Lower single probability for being placed an order by target user of resource, be conducive to quickly and accurately estimate lower single probability, and due to pre-
Estimate model to obtain previously according to the user of acquisition and the history feature data training of resource, that is, trains the data source of prediction model
In the real history data of user and resource, reference value is high, may make the lower single probability estimated more accurate, reliable.Recommend
The resource that module concentrates recommended candidate recommends target user according to the height for the lower single probability estimated, so that target user
Being recommended at first is the highest resource of lower single possibility, improves the efficiency that target user places an order in target trade company, simultaneously also
The usage experience of target user can be promoted.Also, since History Order of the different target users in same target trade company can
Can be not identical, therefore, it is suitble to respective recommended candidate collection with having per family for different target, is conducive to different target user
Meet the individual demand of different target user when being recommended.
4th embodiment of the invention is related to a kind of electronic equipment, as shown in figure 4, the electronic equipment includes: at least one
A processor 401；And the memory 402 with the communication connection of at least one processor 401；And with scanning means communication link
The communication component 403 connect, communication component 403 send and receive data under the control of processor 401；Wherein, memory 402 is deposited
The instruction that can be executed by least one processor 401 is contained, instruction is executed by least one processor 401 to realize:
According to the History Order of target trade company, target user was descended into list and the resource for meeting preset price request is added
Recommended candidate collection；According to the prediction model for estimating lower single probability, the resource of the recommended candidate concentration is estimated by the mesh
Lower single probability that mark user places an order；Wherein, the prediction model is previously according to the user of acquisition and the history feature data of resource
Training obtains；The resource that the recommended candidate is concentrated recommends the target user according to the height for the lower single probability estimated.
Specifically, which includes: one or more processors 401 and memory 402, at one in Fig. 4
For reason device 401.Processor 401, memory 402 can be connected by bus or other modes, to be connected by bus in Fig. 4
It is connected in example.Memory 402 is used as a kind of non-volatile computer readable storage medium storing program for executing, can be used for storing non-volatile software journey
Sequence, non-volatile computer executable program and module.Processor 401 is stored in non-easy in memory 402 by operation
The property lost software program, instruction and module realize above-mentioned push away thereby executing the various function application and data processing of equipment
Recommend method.
Memory 402 may include storing program area and storage data area, wherein storing program area can store operation system
Application program required for system, at least one function；It storage data area can the Save option list etc..In addition, memory 402 can be with
It can also include nonvolatile memory, for example, at least disk memory, a flash memory including high-speed random access memory
Device or other non-volatile solid state memory parts.In some embodiments, it includes relative to processing that memory 402 is optional
The remotely located memory 402 of device 401, these remote memories 402 can pass through network connection to external equipment.Above-mentioned network
Example include but is not limited to internet, intranet, local area network, mobile radio communication and combinations thereof.
One or more module is stored in memory 402, when being executed by one or more processor 401, is held
Recommended method in the above-mentioned any means embodiment of row.
The said goods can be performed the application embodiment provided by method, have the corresponding functional module of execution method and
Beneficial effect, the not technical detail of detailed description in the present embodiment, reference can be made to method provided by the application embodiment.
5th embodiment of the invention is related to a kind of non-volatile memory medium, for storing computer-readable program,
The computer-readable program is used to execute above-mentioned all or part of embodiment of the method for computer.
That is, it will be understood by those skilled in the art that implement the method for the above embodiments be can be with
Relevant hardware is instructed to complete by program, which is stored in a storage medium, including some instructions are to make
It obtains an equipment (can be single-chip microcontroller, chip etc.) or processor (processor) executes side described in each embodiment of the application
The all or part of the steps of method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic or disk etc. are various can store journey
The medium of sequence code.
It will be understood by those skilled in the art that the respective embodiments described above are to realize specific embodiments of the present invention,
And in practical applications, can to it, various changes can be made in the form and details, without departing from the spirit and scope of the present invention.
The embodiment of the present application provides a kind of recommended method of A1., comprising:
According to the History Order of target trade company, target user was descended into list and the resource for meeting preset price request is added
Recommended candidate collection；
According to the prediction model for estimating lower single probability, estimates the resource that the recommended candidate is concentrated and used by the target
Lower single probability that family places an order；Wherein, the prediction model is previously according to the user of acquisition and the history feature data training of resource
It obtains；
The resource that the recommended candidate is concentrated recommends the target user according to the height for the lower single probability estimated.
A2. recommended method according to a1, the resource that the recommended candidate is concentrated are divided into: single resource and resource group
It closes；Estimate lower single probability that the combination of resources that the recommended candidate is concentrated is placed an order by the target user, comprising:
Lower single probability that each of combination of resources resource is placed an order by the target user is estimated respectively；
According to lower single probability of each resource, estimates the combination of resources and placed an order by what the target user placed an order
Probability.
A3. the recommended method according to A2, lower single probability of each resource according to, estimates the resource
Lower single probability that combination is placed an order by the target user, comprising:
According to lower single probability of each resource, lower single probability average is calculated；
Lower single probability by lower single probability average, as the combination of resources estimated.
A4. recommended method according to a1, it is described that target user was descended into list and meets the money of preset price request
Recommended candidate collection is added in source, comprising:
If the target user descended the price of single single resource to meet the price request, the valence will be met
Recommended candidate collection is added in the single resource that lattice require；
If the target user descended the price of single single resource to be unsatisfactory for the price request, according to it is each not
The price and attribute for meeting the single resource of the price request generate combination of resources, wherein the price of the combination of resources is full
The recommended candidate collection is added in the combination of resources by the foot price request.
A5. recommended method according to a1, the resource that the recommended candidate is concentrated is that the target trade company currently exists
The resource sold.
A6. the history feature data of recommended method according to a1, the user and resource include: user property spy
Sign, Resource Properties feature, user and the cross attribute feature of resource, the real-time behavioural characteristic of user.
A7. the cross attribute feature of the recommended method according to A6, the user and resource, comprising: user is to resource
History evaluation, user is to the purchase volume of resource.
A8. the recommended method according to any one of A1 to A7, the prediction model are according to the pre- of XGBoost training
Estimate model.
The embodiment of the present application provides a kind of recommendation apparatus of B1., comprising:
Target user was played list and met preset price by processing module for the History Order according to target trade company
It is required that resource be added recommended candidate collection；
Module is estimated, for estimating the money that the recommended candidate is concentrated according to the prediction model for estimating lower single probability
Lower single probability that source is placed an order by the target user；Wherein, the prediction model is previously according to the user of acquisition and going through for resource
The training of history characteristic obtains；
Recommending module, the resource for concentrating the recommended candidate are recommended according to the height for the lower single probability estimated
The target user.
The embodiment of the present application provides C1. a kind of electronic equipment, including memory and processor, and memory stores computer
Program, processor execute when running program:
According to the History Order of target trade company, target user was descended into list and the resource for meeting preset price request is added
Recommended candidate collection；
According to the prediction model for estimating lower single probability, estimates the resource that the recommended candidate is concentrated and used by the target
Lower single probability that family places an order；Wherein, the prediction model is previously according to the user of acquisition and the history feature data training of resource
It obtains；
The resource that the recommended candidate is concentrated recommends the target user according to the height for the lower single probability estimated.
C2. the electronic equipment according to C1, the resource that the recommended candidate is concentrated are divided into: single resource and resource group
It closes；Estimate lower single probability that the combination of resources that the recommended candidate is concentrated is placed an order by the target user, comprising:
Lower single probability that each of combination of resources resource is placed an order by the target user is estimated respectively；
According to lower single probability of each resource, estimates the combination of resources and placed an order by what the target user placed an order
Probability.
C3. the electronic equipment according to C2, lower single probability of each resource according to, estimates the resource
Lower single probability that combination is placed an order by the target user, comprising:
According to lower single probability of each resource, lower single probability average is calculated；
Lower single probability by lower single probability average, as the combination of resources estimated.
C4. the electronic equipment according to C1, it is described that target user was descended into list and meets the money of preset price request
Recommended candidate collection is added in source, comprising:
If the target user descended the price of single single resource to meet the price request, the valence will be met
Recommended candidate collection is added in the single resource that lattice require；
If the target user descended the price of single single resource to be unsatisfactory for the price request, according to it is each not
The price and attribute for meeting the single resource of the price request generate combination of resources, wherein the price of the combination of resources is full
The recommended candidate collection is added in the combination of resources by the foot price request.
C5. the electronic equipment according to C1, the resource that the recommended candidate is concentrated is that the target trade company currently exists
The resource sold.
C6. the history feature data of the electronic equipment according to C1, the user and resource include: user property spy
Sign, Resource Properties feature, user and the cross attribute feature of resource, the real-time behavioural characteristic of user.
C7. the cross attribute feature of the electronic equipment according to C6, the user and resource, comprising: user is to resource
History evaluation, user is to the purchase volume of resource.
C8. the electronic equipment according to any one of C1 to C7, the prediction model are according to the pre- of XGBoost training
Estimate model.
The embodiment of the present application provides a kind of non-volatile memory medium of D1., described for storing computer-readable program
Computer-readable program is used to execute the recommended method as described in any one of A1 to A8 for computer.
Claims (10)
1. a kind of recommended method characterized by comprising
According to the History Order of target trade company, target user was descended into list and recommendation is added in the resource for meeting preset price request
Candidate Set；
According to the prediction model for estimating lower single probability, the resource of the recommended candidate concentration is estimated by under the target user
Single lower single probability；Wherein, the prediction model is obtained previously according to the user of acquisition and the history feature data training of resource；
The resource that the recommended candidate is concentrated recommends the target user according to the height for the lower single probability estimated.
2. recommended method according to claim 1, which is characterized in that the resource that the recommended candidate is concentrated is divided into: single
Resource and combination of resources；Estimate lower single probability that the combination of resources that the recommended candidate is concentrated is placed an order by the target user, packet
It includes:
Lower single probability that each of combination of resources resource is placed an order by the target user is estimated respectively；
According to lower single probability of each resource, estimates the combination of resources and placed an order generally by what the target user placed an order
Rate.
3. recommended method according to claim 2, which is characterized in that placing an order for each resource according to is general
Rate estimates lower single probability that the combination of resources is placed an order by the target user, comprising:
According to lower single probability of each resource, lower single probability average is calculated；
Lower single probability by lower single probability average, as the combination of resources estimated.
4. recommended method according to claim 1, which is characterized in that described that target user was played list and is met preset
Recommended candidate collection is added in the resource of price request, comprising:
If the target user descended the price of single single resource to meet the price request, the price will be met and wanted
Recommended candidate collection is added in the single resource asked；
If the target user descended the price of single single resource to be unsatisfactory for the price request, it is unsatisfactory for according to each
The price and attribute of the single resource of the price request generate combination of resources, wherein the price of the combination of resources meets institute
Price request is stated, the recommended candidate collection is added in the combination of resources.
5. recommended method according to claim 1, which is characterized in that the resource that the recommended candidate is concentrated is the mesh
Mark the current resource on sale of trade company.
6. recommended method according to claim 1, which is characterized in that the history feature data packet of the user and resource
It includes: user property feature, Resource Properties feature, user and the cross attribute feature of resource, the real-time behavioural characteristic of user.
7. recommended method according to claim 6, which is characterized in that the cross attribute feature of the user and resource, packet
Include: user is to the history evaluation of resource, user to the purchase volume of resource.
8. a kind of recommendation apparatus characterized by comprising
Target user was played list and met preset price request by processing module for the History Order according to target trade company
Resource be added recommended candidate collection；
Module is estimated, for estimating the resource quilt that the recommended candidate is concentrated according to the prediction model for estimating lower single probability
Lower single probability that the target user places an order；Wherein, the prediction model is special previously according to the user of acquisition and the history of resource
Sign data training obtains；
Recommending module, the resource for concentrating the recommended candidate are recommended described according to the height for the lower single probability estimated
Target user.
9. a kind of electronic equipment, including memory and processor, memory stores computer program, and processor is held when running program
Row:
According to the History Order of target trade company, target user was descended into list and recommendation is added in the resource for meeting preset price request
Candidate Set；
According to the prediction model for estimating lower single probability, the resource of the recommended candidate concentration is estimated by under the target user
Single lower single probability；Wherein, the prediction model is obtained previously according to the user of acquisition and the history feature data training of resource；
The resource that the recommended candidate is concentrated recommends the target user according to the height for the lower single probability estimated.
10. a kind of non-volatile memory medium, for storing computer-readable program, the computer-readable program is by for based on
Calculation machine executes the recommended method as described in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910389818.XA CN110264301A (en) | 2019-05-10 | 2019-05-10 | Recommended method, device, electronic equipment and non-volatile memory medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910389818.XA CN110264301A (en) | 2019-05-10 | 2019-05-10 | Recommended method, device, electronic equipment and non-volatile memory medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110264301A true CN110264301A (en) | 2019-09-20 |
Family
ID=67914580
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910389818.XA Pending CN110264301A (en) | 2019-05-10 | 2019-05-10 | Recommended method, device, electronic equipment and non-volatile memory medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110264301A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310020A (en) * | 2019-09-20 | 2020-06-19 | 拉扎斯网络科技（上海）有限公司 | Dish recommendation method, recommended content uploading method and device and electronic equipment |
| CN111476632A (en) * | 2020-03-31 | 2020-07-31 | 北京三快在线科技有限公司 | Method and device for displaying resources, electronic equipment and readable storage medium |
| CN111597447A (en) * | 2020-05-14 | 2020-08-28 | 拉扎斯网络科技（上海）有限公司 | Object pushing method and device, electronic equipment and storage medium |
| CN112685627A (en) * | 2019-10-17 | 2021-04-20 | 北京星选科技有限公司 | Object pushing method and device, electronic equipment and storage medium |
| CN112734462A (en) * | 2020-12-30 | 2021-04-30 | 北京字跳网络技术有限公司 | Information recommendation method, device, equipment and medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150129808A1 (en) * | 2013-11-13 | 2015-05-14 | U.S. Army Research Laboratory Attn: Rdrl-Loc-I | Deformable polymer composites with controlled electrical performance during deformation through tailored strain-dependent conductive filler contact |
| US20150363850A1 (en) * | 2013-03-06 | 2015-12-17 | Tencent Technology (Shenzhen) Company Limited | Method for processing order, transaction server and computer readable storage medium |
| CN106327227A (en) * | 2015-06-19 | 2017-01-11 | 北京航天在线网络科技有限公司 | Information recommendation system and information recommendation method |
| CN106485562A (en) * | 2015-09-01 | 2017-03-08 | 苏宁云商集团股份有限公司 | A kind of commodity information recommendation method based on user's history behavior and system |
| CN106649774A (en) * | 2016-12-27 | 2017-05-10 | 北京百度网讯科技有限公司 | Artificial intelligence-based object pushing method and apparatus |
| CN109165350A (en) * | 2018-08-23 | 2019-01-08 | 成都品果科技有限公司 | A kind of information recommendation method and system based on deep knowledge perception |
| CN109685631A (en) * | 2019-01-10 | 2019-04-26 | 博拉网络股份有限公司 | A kind of personalized recommendation method based on big data user behavior analysis |
-
2019
- 2019-05-10 CN CN201910389818.XA patent/CN110264301A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150363850A1 (en) * | 2013-03-06 | 2015-12-17 | Tencent Technology (Shenzhen) Company Limited | Method for processing order, transaction server and computer readable storage medium |
| US20150129808A1 (en) * | 2013-11-13 | 2015-05-14 | U.S. Army Research Laboratory Attn: Rdrl-Loc-I | Deformable polymer composites with controlled electrical performance during deformation through tailored strain-dependent conductive filler contact |
| CN106327227A (en) * | 2015-06-19 | 2017-01-11 | 北京航天在线网络科技有限公司 | Information recommendation system and information recommendation method |
| CN106485562A (en) * | 2015-09-01 | 2017-03-08 | 苏宁云商集团股份有限公司 | A kind of commodity information recommendation method based on user's history behavior and system |
| CN106649774A (en) * | 2016-12-27 | 2017-05-10 | 北京百度网讯科技有限公司 | Artificial intelligence-based object pushing method and apparatus |
| CN109165350A (en) * | 2018-08-23 | 2019-01-08 | 成都品果科技有限公司 | A kind of information recommendation method and system based on deep knowledge perception |
| CN109685631A (en) * | 2019-01-10 | 2019-04-26 | 博拉网络股份有限公司 | A kind of personalized recommendation method based on big data user behavior analysis |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310020A (en) * | 2019-09-20 | 2020-06-19 | 拉扎斯网络科技（上海）有限公司 | Dish recommendation method, recommended content uploading method and device and electronic equipment |
| CN112685627A (en) * | 2019-10-17 | 2021-04-20 | 北京星选科技有限公司 | Object pushing method and device, electronic equipment and storage medium |
| CN111476632A (en) * | 2020-03-31 | 2020-07-31 | 北京三快在线科技有限公司 | Method and device for displaying resources, electronic equipment and readable storage medium |
| CN111597447A (en) * | 2020-05-14 | 2020-08-28 | 拉扎斯网络科技（上海）有限公司 | Object pushing method and device, electronic equipment and storage medium |
| CN112734462A (en) * | 2020-12-30 | 2021-04-30 | 北京字跳网络技术有限公司 | Information recommendation method, device, equipment and medium |
| CN112734462B (en) * | 2020-12-30 | 2024-04-05 | 北京字跳网络技术有限公司 | Information recommendation method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110264301A (en) | Recommended method, device, electronic equipment and non-volatile memory medium | |
| CN106447545B (en) | Catering order form booking method and booking terminal | |
| CN101697162B (en) | Method and system for intelligently recommending ordering dishes | |
| US7647309B1 (en) | Browse relevance | |
| US20170316488A1 (en) | Systems and Methods of Food Management | |
| CN107862542A (en) | A kind of vegetable recommends method and apparatus | |
| CN108288219A (en) | The method for pushing and system of food product information | |
| CN109493254A (en) | A kind of vegetable production sequence with urge dish method, apparatus and vegetable processing system | |
| CN109214955B (en) | Method and device for generating meal package | |
| CN108228579A (en) | Network interaction system | |
| JP2012003319A (en) | Menu determination support device, menu determination support program, and memory medium for storing menu determination support program | |
| KR20120029027A (en) | Service method and apparatus for providing diet menu | |
| CN104112217A (en) | Taste selection information marking method and application method | |
| CN109214956B (en) | Meal pushing method and device | |
| CN109993623A (en) | Item Information methods of exhibiting, system and electronic equipment based on image recognition | |
| CN110349648A (en) | A kind of generation recommending recipes method based on recipe library | |
| CN109064296A (en) | A kind of methods, devices and systems assisting snack | |
| JP2002056180A (en) | Cooking recipe providing system and computer-readable recording medium with recorded cooking recipe providing program | |
| CN106886921A (en) | Personalized recommendation method based on user interest | |
| US20140295382A1 (en) | System and a method for generating a meal plan | |
| Wang et al. | Combining fuzzy integral and GRA method for evaluating the service quality of fast-food restaurants | |
| CN110362739A (en) | A kind of recommending recipes method based on big data | |
| CN109493253A (en) | A kind of method of ordering and device, computer storage medium and electronic equipment | |
| CN115221420A (en) | Diet recommendation method and system based on user portrait | |
| JP2002207869A (en) | Provision of recipe |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190920 |