CN108415970B - Search result sort method, device, electronic equipment and storage medium - Google Patents

Search result sort method, device, electronic equipment and storage medium Download PDF

Info

Publication number
CN108415970B
CN108415970B CN201810128612.7A CN201810128612A CN108415970B CN 108415970 B CN108415970 B CN 108415970B CN 201810128612 A CN201810128612 A CN 201810128612A CN 108415970 B CN108415970 B CN 108415970B
Authority
CN
China
Prior art keywords
search result
user
corresponding goods
probability
product features
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
Application number
CN201810128612.7A
Other languages
Chinese (zh)
Other versions
CN108415970A (en
Inventor
易根良
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Sankuai Online Technology Co Ltd
Original Assignee
Beijing Sankuai Online Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Sankuai Online Technology Co Ltd filed Critical Beijing Sankuai Online Technology Co Ltd
Priority to CN201810128612.7A priority Critical patent/CN108415970B/en
Publication of CN108415970A publication Critical patent/CN108415970A/en
Application granted granted Critical
Publication of CN108415970B publication Critical patent/CN108415970B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/951—Indexing; Web crawling techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • 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
    • G06Q30/0282—Rating or review of business operators or products

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Software Systems (AREA)
  • Finance (AREA)
  • Databases & Information Systems (AREA)
  • Game Theory and Decision Science (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The present invention provides a kind of search result sort method, device, electronic equipment and storage mediums, are related to search technique field.The described method includes: receiving searching request, user identifier and end article information are carried in described search request, multiple search results are obtained according to the end article information search, determine the product features of corresponding goods in each search result, user characteristics are obtained according to the user identifier, according to the product features of corresponding goods in the user characteristics and each search result, determine user for the consumption intention data of corresponding goods in each search result, according to the consumption intention data, the multiple search result is ranked up.The present invention being capable of consumption demand according to user to the different commodity searched, the search result that targetedly the opposite direction user provides is ranked up, the accuracy of the accuracy and provided search result that are ranked up to search result is improved, and then also improves the accuracy for providing a user commodity.

Description

Search result sort method, device, electronic equipment and storage medium
Technical field
The present invention relates to search technique field, more particularly to a kind of search result sort method, device, electronic equipment and Storage medium.
Background technique
With the development of electronic technology, the application of e-commerce is also more and more extensive.Businessman can be by way of webpage Commodity are provided a user, user can search for required commodity by access webpage.Due to that may be wrapped in search result Multiple commodity are included, and would generally be variant between each commodity, such as quality of commodity etc., so providing the standard of commodity to improve True property, needs a kind of search result ordering method.
It in the prior art, can be by counting the sales volume of corresponding goods in each search result, user for each search result The number of reviews and collection number of middle corresponding goods.It, can be according to the sales volume, evaluation number or collection number when providing commodity Mesh is ranked up corresponding goods in each search result, and the commodity after sequence are supplied to user.
Although the prior art can be ranked up search result, improves to a certain extent and the accurate of commodity is provided Property, but due to also usually there will be difference, such as taste, hobby, purchasing power etc., therefore different user between different users Demand to commodity is also different, and in the prior art, for any user, the sortord of provided commodity is identical , also it is difficult to meet without user the needs of to commodity certainly, the accuracy for providing commodity is fairly limited.
Summary of the invention
In view of the above problems, it proposes on the present invention overcomes the above problem or at least be partially solved in order to provide one kind State search result sort method, device, electronic equipment and the storage medium of problem.
On the one hand, the present invention provides a kind of search result ordering methods, comprising:
Searching request is received, carries user identifier and end article information in described search request；
Multiple search results are obtained according to the end article information search, determine the quotient of corresponding goods in each search result Product feature；
User characteristics are obtained according to the user identifier；
According to the product features of corresponding goods in the user characteristics and each search result, determine that user ties for each search The consumption intention data of corresponding goods in fruit；
According to the consumption intention data, the multiple search result is ranked up.
Optionally, the product features according to corresponding goods in the user characteristics and each search result, determine user Consumption intention data for corresponding goods in each search result includes:
According to the product features of corresponding goods in the user characteristics and each search result, pass through the first machine being pre-created Device learning model determines the user for the payment probability of corresponding goods in each search result, the first machine learning mould Type is used to determine payment probability according to the product features of corresponding goods in user characteristics and each search result；
According to the product features of corresponding goods in the user characteristics and each search result, pass through the second machine being pre-created Device learning model determines the user for the reimbursement probability of corresponding goods in each search result, the second machine learning mould Type is used to determine reimbursement probability according to the product features of corresponding goods in user characteristics and each search result；
According to the payment probability and the reimbursement probability, determine the user for corresponding goods in each search result Probability is consumed, is the consumption intention data by the consumption determine the probability.
Optionally, in the product features according to corresponding goods in the user characteristics and each search result, by pre- The first machine learning model first created, before determining the user for the payment probability of corresponding goods in each search result, Further include:
The history Access Events that multiple users are directed to commodity are obtained, the history Access Events include payment event and unpaid Money event；
User characteristics and product features are extracted from the history Access Events, and calculate each use according to history Access Events Payment probability of the family to commodity；
According to extracted product features, user characteristics and payment probability calculated, first engineering is created Practise model.
Optionally, in the product features according to corresponding goods in the user characteristics and each search result, by pre- The second machine learning model first created, before determining the user for the reimbursement probability of corresponding goods in each search result, Further include:
The history Access Events that multiple users are directed to commodity are obtained, the history Access Events include refund event and do not move back Money event；
User characteristics and product features are extracted from the history Access Events, and calculate each use according to history Access Events Reimbursement probability of the family to commodity；
According to extracted product features, user characteristics and reimbursement probability calculated, second engineering is created Practise model.
Optionally, the product features according to corresponding goods in the user characteristics and each search result, determine user Consumption intention data for corresponding goods in each search result includes:
According to the product features of corresponding goods in the user characteristics and each search result, pass through the third machine being pre-created Device learning model determines the user for the payment probability of corresponding goods in each search result, the third machine learning mould Type is used to determine consumption probability according to the product features of corresponding goods in user characteristics and each search result；
It is the consumption intention data by the consumption determine the probability.
Optionally, in the product features according to corresponding goods in the user characteristics and each search result, by pre- The third machine learning model first created, before determining the user for the payment probability of corresponding goods in each search result, Further include:
The history Access Events that multiple users are directed to commodity are obtained, the history Access Events include payment event, unpaid Money event, refund event and non-refund event；
User characteristics and product features are extracted from the history Access Events, and calculate each use according to history Access Events Consumption probability of the family to commodity；
According to extracted product features, user characteristics and consumption probability calculated, the third engineering is created Practise model.
Optionally, further includes:
Environmental characteristic is obtained, the environmental characteristic includes the user respectively between corresponding goods in each search result Distance；
It is described according to the consumption intention data, the multiple search result be ranked up include:
In conjunction with the environmental characteristic and the consumption intention data, the multiple commodity are ranked up.
Optionally, described that multiple search results are obtained according to the end article information search, it determines in each search result The product features of corresponding goods include:
Determine product features of the corresponding goods in the second preset time period in each search result, second preset time Section is using current time as the period in the third preset duration of starting point；
It is described to include: according to user identifier acquisition user characteristics
The user characteristics in second preset time period are obtained according to the user identifier.
On the other hand, the present invention provides a kind of search results ranking devices, comprising:
Receiving module carries user identifier and end article information in described search request for receiving searching request；
First determining module determines each search for obtaining multiple search results according to the end article information search As a result the product features of middle corresponding goods；
First obtains module, for obtaining user characteristics according to the user identifier；
Second determining module, for the product features according to corresponding goods in the user characteristics and each search result, really User is determined for the consumption intention data of corresponding goods in each search result；
Sorting module, for being ranked up to the multiple search result according to the consumption intention data.
On the other hand, the present invention provides a kind of electronic equipment, including memory, processor and storage are on a memory simultaneously The computer program that can be run on a processor, the processor are realized when executing described program such as any one of aforementioned the method The step of.
On the other hand, the present invention provides a kind of computer readable storage medium, it is stored thereon with computer program, the journey It is realized when sequence is executed by processor such as the step of any one of aforementioned the method.
In embodiments of the present invention, due to may be had differences between different users and different commodity, from And different users has different consumption demands to different commodity, so the user for submitting searching request, Neng Gougen Multiple searching results are obtained according to the end article information search carried in searching request, are marked according to the user carried in searching request Knowledge gets corresponding user characteristics, then according to the product features of corresponding goods in the user characteristics and each search result, really The fixed user is directed to the consumption intention data of corresponding goods in each search result, and then can will be tied according to the user to each search The consumption intention data of corresponding goods is ranked up search result in fruit, and that is to say can be according to user to the difference searched The consumption demand of commodity, the search result that targetedly the opposite direction user provides are ranked up, improve and carry out to search result The accuracy of the accuracy of sequence and provided search result, and then also improve the accuracy for providing a user commodity.
The above description is only an overview of the technical scheme of the present invention, in order to better understand the technical means of the present invention, And it can be implemented in accordance with the contents of the specification, and in order to allow above and other objects of the present invention, feature and advantage can It is clearer and more comprehensible, the followings are specific embodiments of the present invention.
Detailed description of the invention
By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
Fig. 1 is a kind of step flow chart of the search result ordering method provided according to embodiments of the present invention；
Fig. 2 is the step flow chart of another search result ordering method provided according to embodiments of the present invention；
Fig. 3 is a kind of structural block diagram of the search results ranking device provided according to embodiments of the present invention.
Specific embodiment
Exemplary embodiments of the present disclosure are described in more detail below with reference to accompanying drawings.Although showing the disclosure in attached drawing Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here It is limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure It is fully disclosed to those skilled in the art.
Embodiment one
Referring to Fig.1, the step flow chart of one of embodiment of the present invention search result ordering method, the search are shown Result ordering method includes:
Step 101, searching request is received, carries user identifier and end article information in described search request.
Search result ordering method provided by the embodiment of the present invention can be applied in cloud platform, and user can pass through net Page application or client submit searching request to the cloud platform, and user identifier and end article letter are carried in the searching request Breath, so that the cloud platform can be scanned for according to the end article information, and provides search result according to the user identifier To user.
The user identifier is used for one user of unique identification, which can be user's name or User ID (Identification, identity card) etc..
The end article information is the keyword for commodity required for obtaining, for example, the end article information can be with At least one of the title of the affiliated businessman of the type of merchandise, commodity.
It include " User ID 1 " and " express hotel " in searching request, wherein User ID 1 for example, receiving searching request 1 For the user identifier of the user of submission searching request 1, express hotel is the type of merchandise.
Step 102, multiple search results are obtained according to the end article information search, determined corresponding in each search result The product features of commodity.
Due to being had differences between unused commodity, such as quality, price, material etc., so as to meet user demand Program also can be different, therefore, search result is ranked up for the ease of user for submission searching request, Jin Erti Height provides the accuracy of search result and commodity, can in each search result corresponding goods product features.
The commodity with the end article information matches can be obtained from least one businessman according to the end article information, To obtain search result, each search result can correspond to a commodity.
Product features are feature possessed by commodity, may include reimbursement amount, sales volume, evaluation number, in the type of merchandise At least one, certainly, in practical applications, which can also include other features of commodity, such as the affiliated quotient of commodity The title of family, the reimbursement rate in geographical location (address or longitude and latitude etc.), commodity locating for businessman click number, collection number, price Deng.Certainly, in practical applications, in order to improve the accuracy of determining product features, and then improve and search result is ranked up Accuracy, to accurately provide a user commodity, which may include commodity subcharacter, for example, commodity Reimbursement rate may include the client from different clients type (such as Android client and apple client) to the commodity The reimbursement rate of reimbursement, weekend and non-all foot couple commodity are carried out to carry out the reimbursement rate of reimbursement, (for example roll into a ball for different Payment Types Purchase or forward purchasing) reimbursement rate.
Wherein it is possible to which the history obtained for the commodity accesses according to the commodity sign of corresponding goods in each search result Event, the history Access Events may include click event, payment event, arrearage event, refund event, non-refund event, Evaluation event such as does not evaluate event, collection event, does not collect event at the history Access Events, passes through and calculates different history access The number of event is so that it is determined that different product features, that is to say and extract the commodity spy for obtaining the commodity from history Access Events Sign.For example, the number for the click event being calculated is the click number of the commodity, what is be calculated searches the number of hiding event The as collection number of the commodity, the number for the refund event being calculated, the number with refund event and non-refund event The sum of between ratio, the as reimbursement rate of the commodity.Certainly, in embodiments of the present invention, it is preferred that for will not be because of going through History Access Events and the product features that change can also be obtained by obtaining pre-stored product features, for example, the type of merchandise, Geographical location locating for the title of the affiliated businessman of commodity and businessman etc. can be the commodity class by receiving the commodity that businessman submits Geographical location locating for the title of the affiliated businessman of type, commodity and businessman obtains.
Commodity sign is used for one commodity of unique identification, and commodity sign can be title or ID of commodity etc..
History Access Events are to detect when user accesses to the commodity to the user characteristics of the user, the commodity The event that product features record generates.For example, if detecting that user clicks the commodity, can recorde as click event, it should It may include the user characteristics of the user, the type of merchandise of the commodity, the title of the affiliated businessman of commodity or businessman in click event Locating geographical location etc.；If detecting that the user is to click the commodity and successful payment, can recorde as payment event；Such as Fruit detects that the user is to click the commodity but arrearage success, can recorde as arrearage event；If detecting the use Family is the commodity successful payment and reimbursement success, be can recorde as refund event；If detecting the user for commodity payment Success but non-reimbursement success, can recorde as non-refund event.
In embodiments of the present invention, it is preferred that reimbursement refers in the first preset duration after commodity successful payment Reimbursement.
First preset duration can be determined that the first preset duration can be 1 week or 1 month etc. by cloud platform.
In embodiments of the present invention, it is preferred that further include other feature, such as environmental characteristic in history Access Events.It should Environmental characteristic is for illustrating environment locating for corresponding goods in the user or each search result, such as current time and user Respectively with the distance between corresponding goods in each search result, weather etc..
Wherein, the distance between the user and corresponding goods in each search result can be geographical position locating for the user It sets respectively the distance between with geographical location locating for corresponding goods in each search result.
Step 103, user characteristics are obtained according to the user identifier.
Since there is also differences between different users, correspondingly, to the desirability of commodity different in search result It can difference therefore for the ease of being ranked up for the user of searching request is submitted to search result, and then mention For the accuracy of search result and commodity, available user characteristics.
User characteristics are feature possessed by user, which may include user activity, user gradation, user At least one of consuming capacity, user's reimbursement number.Certainly, in practical applications, which can also include other Feature, for example, geographical location locating for the client type of the client of the user, user.
User activity can be by buying number and the last time of commodity according to user in the first preset time period Purchase commodity determine that the user activity may be greater than 0 and be less than or equal to 1 apart from second preset duration at current time Numerical value.Wherein, the first preset time period can be the period of the second preset duration before current time, the first preset time Section or the second preset duration can be determined by the cloud platform, for example, the second preset duration can be nearest 30 days, 60 days or 90 It etc..
For example, user activity can be calculated by formula 1:ACT=a*N1+b*N2.Wherein, ACT indicates that user is living Jerk, N1 buy the number of commodity in the first preset time period, and N2 is the first duration, and a is to purchase in the first preset time period The weight of the number of commodity is bought, b is the weight of the second preset duration, and the value of a and b can be by presetting to obtain.
User gradation can buy commodity according to user and the sum of pay the bill and payment number determination, which can be with It is greater than 0 numerical value for being less than or equal to 10, and commodity number of the sum of paying the bill and pay the bill is bought according to user and determines user etc. The mode of grade, can determine the user with number and the first duration of commodity is bought in the first preset time period according to user The mode of liveness is similar, no longer repeats one by one herein.
Customer consumption ability can be bought at least one of commodity are paid the bill and/or average value by user.
User's reimbursement number can be determined by counting the number of the refund event of the user identifier including the user.
In addition, in practical applications, step 103 can also be first carried out and execute step 102 again, alternatively, being performed simultaneously step 102 or 103, it that is to say, the embodiment of the present invention is not specifically limited the execution order of step 102 and 103.
Step 104, according to the product features of corresponding goods in the user characteristics and each search result, determine that user is directed to The consumption intention data of corresponding goods in each search result.
In order to be accurately ranked up to search result, so that commodity are accurately provided a user, it can be according to this The product features of corresponding goods in the user characteristics of user and each search result determine the user for right in each search result The consumption intention data for answering commodity, that is to say, can be directed to different users, determine the user for corresponding in each search result The consumption intention data of commodity.
Consumption intention data is to carry out consuming related data to commodity with user, so as to illustrate user to commodity Consumption demand or tendency, the consumption intention data can be the payment probability that the user pays the bill to the commodity.
Wherein it is possible to according to the product features of corresponding goods in the user characteristics and each search result, by creating in advance The first machine learning model built determines the user for the payment probability of corresponding goods in each search result, the first machine Learning model is used to determine payment probability according to the product features of corresponding goods in user characteristics and each search result.
For example, the user characteristics of user 1 include user gradation be 3, client type is Android, user activity 0.6, The product features of commodity 1 include that nearest 90 days reimbursement numbers are 10, then by the first machine learning model, according to the use of user 1 The product features of family feature and commodity 1, it is 50% to the payment probability of commodity 1 that user 1, which is calculated,.The user of user 2 is special Levying including user gradation is 1, client type is Android, and user activity 0.1 then passes through the first machine learning model, root According to the user characteristics of user 2 and the product features of commodity 1, it is 10% to the payment probability of commodity 1 that user 2, which is calculated,.
In embodiments of the present invention, it is preferred that described according to corresponding goods in the user characteristics and each search result Product features determine the user for corresponding to quotient in each search result by the first machine learning model being pre-created Before the payment probability of product, the first machine learning model is created.
First machine learning model can be XGBOOST model.
In embodiments of the present invention, it is preferred that from the foregoing it will be appreciated that including record payment thing in the history Access Events of commodity Part and arrearage event, therefore in order to improve the accuracy of the first machine learning model, more accurately to consumption intention number According to being calculated, to improve the subsequent accuracy being ranked up to search result, available multiple users are for commodity History Access Events, the history Access Events include payment event and arrearage event, are mentioned from the history Access Events It takes user characteristics and product features, and calculates each user to the payment probability of commodity according to history Access Events, according to being extracted Product features, user characteristics and payment probability calculated, create first machine learning model.
Wherein, the user characteristics and product features extracted from history Access Events can be used as sample, pass through the first machine Device learning model calculates each user to the payment probability of commodity according to the sample extracted, can be to the first machine learning mould Type is trained.
Step 105, according to the consumption intention data, the multiple search result is ranked up.
Since different users is different for the consumption intention data of different commodity, i.e., different user can to different commodity There can be different consumption propensity, therefore, for the ease of being directed to any user, targetedly provide more accurate search knot Fruit will be ranked up search result according to consumption intention data of the user to corresponding goods in each search result, that is to say Consumption demand according to user to the different commodity searched, the search result that targetedly the opposite direction user provides are arranged Sequence improves the accuracy of accuracy and provided search result being ranked up to search result, so also improve to Family provides the accuracy of commodity, can be according to the user to the consumption intention data of corresponding goods in each search result, to multiple Search result is ranked up.
Wherein, according to the difference of consumption intention data, the mode being ranked up to multiple commodity be can also be different, than Such as, when the consumption intention data is payment probability, can according to payment probability sequence from big to small, to multiple commodity into Row sequence, to guarantee to pay the bill, the biggish commodity of probability are preferentially supplied to user.If the consumption intention data is unpaid When money probability, multiple commodity can be ranked up according to arrearage probability from small arrival volume sequence.
Arrearage probability is the probability that user does not pay the bill to commodity, determines that the mode of arrearage probability can be with determination The mode of payment probability is identical, no longer repeats one by one herein.
For example, user 1 is 50% to the payment probability of commodity 1, the payment probability to commodity 2 is 65%, to paying for commodity 3 Money probability is 20%, is 80% to the payment probability of commodity 4, therefore, according to the sequence of payment probability from big to small, by commodity 1, Commodity 2, commodity 3 and commodity 4 are ranked up, and obtain commodity 4, commodity 2, commodity 1, commodity 3.
Furthermore it is possible to the search result after sequence is supplied to user, so that user be made rapidly and accurately to tie according to search Fruit selects suitable commodity, consumes convenient for user, while ensuring the interests of user and businessman.
The search result after sequence can be shown, by webpage or client so that search result be provided Give the user.
In embodiments of the present invention, due to may be had differences between different users and different commodity, from And different users has different consumption demands to different commodity, so the user for submitting searching request, Neng Gougen Multiple searching results are obtained according to the end article information search carried in searching request, are marked according to the user carried in searching request Knowledge gets corresponding user characteristics, then according to the product features of corresponding goods in the user characteristics and each search result, really The fixed user is directed to the consumption intention data of corresponding goods in each search result, and then can will be tied according to the user to each search The consumption intention data of corresponding goods is ranked up search result in fruit, and that is to say can be according to user to the difference searched The consumption demand of commodity, the search result that targetedly the opposite direction user provides are ranked up, improve and carry out to search result The accuracy of the accuracy of sequence and provided search result, and then also improve the accuracy for providing a user commodity.
Embodiment two
Referring to Fig. 2, the step flow chart of one of embodiment of the present invention search result ordering method, the search are shown Result ordering method includes:
Step 201, cloud platform receives the searching request submitted, and carries user identifier and target quotient in described search request Product information.
In embodiments of the present invention, it is preferred that in order to improve the subsequent reliability for providing a user search result and flexibly Property, user identifier can not be carried in the searching request, but Portable device identifies.When receiving the device identification, according to User identifier corresponding with the device identification is searched in the device identification.
The device identification is used for one equipment of unique identification, which can be MAC (the Media Access of equipment Control, media access control) address, IP (Internet Protocol, network protocol) address etc..The equipment can be with The equipment that the equipment of user identifier binding or the corresponding user of the user identifier log in.
For example, Portable device is identified as device id 1 in searching request, the cloud platform is when receiving the searching request, root Corresponding user is found from the corresponding relationship between device identification and user identifier as shown in table 1 below according to device id 1 It is identified as User ID 1.
Table 1
Device identification User identifier
Device id 1 User ID 1
Device id 3 Device id 2
Device id 4 Device id 3
…… ……
The embodiment of the present invention only by taking above-mentioned table 1 as an example, is said the corresponding relationship between device identification and user identifier Bright, above-mentioned table 1 does not constitute the corresponding relationship between device identification and user identifier and limits.
In embodiments of the present invention, it is preferred that the equipment for being logged in the user identifier can be interacted with the cloud platform, The cloud platform is set to get the device identification of the equipment, for example, the equipment is to the cloud platform sending device bind request, the binding The device identification and the user identifier of the equipment are carried in request.The cloud platform is getting the device identification and the user identifier When, the device identification and the user identifier can be stored to pair as shown in 1 on table between device identification and user identifier In should being related to.
Step 202, the cloud platform obtains multiple search results according to the end article information search, determines each search As a result the product features of middle corresponding goods.
In embodiments of the present invention, it is preferred that since product features may change over time, such as the valence of commodity Lattice etc., therefore, in order to improve the accuracy for getting product features, so improve it is subsequent search result is ranked up it is accurate Property, it can determine product features of the corresponding goods in the second preset time period in each search result, second preset time Section is using current time as the period in the third preset duration of starting point.
Wherein, the second preset time period or third preset duration can be determined by the cloud platform, and third preset duration can be with It is 1 hour, 1 day, 1 week, 1 month etc..And when third preset duration more in short-term, the second preset time period it is also shorter, obtained The real-time of the product features or the subsequent user characteristics got got is also stronger, so ensure that get user characteristics or The accuracy of product features.
The product features of corresponding goods in each search result got can be carried out according to the second preset time period Filter, to obtain product features of the corresponding goods in the second preset time period in each search result.
In embodiments of the present invention, it is preferred that in order to further increase the accuracy being ranked up to search result, the cloud Platform can also obtain other feature, such as environmental characteristic.
The environmental characteristic for illustrating environment locating for corresponding goods in the user or each search result, such as it is current when It carves and user is respectively with the distance between corresponding goods in each search result etc..
Wherein, the distance between the user and corresponding goods in each search result can be geographical position locating for the user It sets respectively the distance between with geographical location locating for corresponding goods in each search result.
Step 203, the cloud platform obtains user characteristics according to the user identifier.
In embodiments of the present invention, it is preferred that since user characteristics may change over time, in order to mention Height gets the accuracy of user characteristics, and then improves the subsequent accuracy being ranked up to search result, can be according to described User identifier obtains the user characteristics in second preset time period, and second preset time period is to be with current time Period in the third preset duration of point.
It can be filtered to user characteristics are obtained, according to the second preset time period to obtain for user identifier User characteristics in second preset time period.
Step 204, the cloud platform is according to the product features of corresponding goods in the user characteristics and each search result, really User is determined for the consumption intention data of corresponding goods in each search result.
In embodiments of the present invention, it is preferred that since user to commodity after paying the bill to buy commodity, it is also possible to by In commercial quality difference or other reasons, to the commodity reimbursement, both it is unfavorable for improving the consumption experience of user, it also can be to businessman Interests damage, therefore, which can be consumption probability, the consumption probability refer to user to commodity payment and Nonrefundable probability.
In embodiments of the present invention, it is preferred that in order to improve the accuracy for calculating consumption probability, and then improve and search is tied The accuracy that fruit is ranked up can pass through according to the product features of corresponding goods in the user characteristics and each search result The first machine learning model being pre-created determines the user for the payment probability of corresponding goods in each search result, institute It states the first machine learning model and determines that payment is general for the product features according to corresponding goods in user characteristics and each search result Rate；According to the product features of corresponding goods in the user characteristics and each search result, pass through the second engineering being pre-created Model is practised, determines the user for the reimbursement probability of corresponding goods in each search result, second machine learning model is used In determining reimbursement probability according to the product features of corresponding goods in user characteristics and each search result, according to the payment probability and The reimbursement probability determines that the user is directed to the consumption probability of corresponding goods in each search result, and the consumption probability is true It is set to the consumption intention data.
It can be by formula 2:P1=P2* (1-P3) come according to payment probability and the reimbursement determine the probability consumption probability.Its In, P1 mark consumption probability, P2 indicates that payment probability, P3 indicate reimbursement probability.
Wherein, the mode for creating the first machine learning model, it is not another herein with the associated description that may refer in aforementioned One repeats.
It can be in the product features according to corresponding goods in the user characteristics and each search result, by creating in advance The second machine learning model built, determine the user for the reimbursement probability of corresponding goods in each search result, described second Before machine learning model is used to determine reimbursement probability according to the product features of corresponding goods in user characteristics and each search result, Create second machine learning model.
Second machine learning model can be XGBOOST model.
In embodiments of the present invention, it is preferred that in order to improve the accuracy of the second machine learning model of creation, with accurately Reimbursement probability is calculated, to improve the accuracy for determining consumption probability and being ranked up to search result, can be obtained Multiple users are taken to be directed to the history Access Events of commodity, the history Access Events include refund event and non-refund event, from User characteristics and product features are extracted in the history Access Events, and calculate each user to commodity according to history Access Events Reimbursement probability creates second engineering according to extracted product features, user characteristics and reimbursement probability calculated Practise model.
Wherein, the mode of the second machine learning model is created, can be similar to the creation mode of the first machine learning model, It no longer repeats one by one herein.
In embodiments of the present invention, it is preferred that in order to reduce needs machine mould quantity, to save creation and instruction Time cost needed for practicing machine mould improves the efficiency for determining consumption intention data, and then improves and arrange search result The efficiency of sequence and the efficiency for providing a user search result can correspond to quotient according in the user characteristics and each search result The product features of product determine the user for corresponding in each search result by the third machine learning model being pre-created The payment probability of commodity, the third machine learning model are used for the quotient according to corresponding goods in user characteristics and each search result Product feature determines consumption probability, is the consumption intention data by the consumption determine the probability.
Wherein it is possible to pass through in the product features according to corresponding goods in the user characteristics and each search result The third machine learning model being pre-created, determine the user for corresponding goods in each search result payment probability it Before, create third machine learning model.
Third machine learning model can be XGBOOST model.
In embodiments of the present invention, it is preferred that in order to improve the accuracy of third machine learning model, and then improve and calculate Consumption intention data and the accuracy being ranked up to search result, available multiple users access for the history of commodity Event, the history Access Events include payment event, arrearage event, refund event and non-refund event, from the history User characteristics and product features are extracted in Access Events, and it is general to the consumption of commodity according to history Access Events to calculate each user Rate creates the third machine learning mould according to extracted product features, user characteristics and consumption probability calculated Type.
Wherein, the mode of third machine learning model is created, can be similar to the creation mode of the first machine learning model, It no longer repeats one by one herein.
In embodiments of the present invention, it is preferred that since the cloud platform can also obtain the other features such as environmental characteristic, because This, which can be other according to the product features and environmental characteristic of corresponding goods in user characteristics, each search result etc. Feature determines the consumption intention data.And according to the product features and environment of corresponding goods in user characteristics, each search result The other features such as feature determine the mode of the consumption intention data, can correspond to quotient with according in user characteristics, each search result The product features of product determine that the mode of the consumption intention data is similar, no longer repeat one by one herein.
For example, the other feature further includes environmental characteristic, quotient can be corresponded to according in the user characteristics, each search result The product features and environmental characteristic of product determine the user for each search by the 4th machine learning model being pre-created As a result the payment probability of middle corresponding goods, the 4th machine learning model are used for according to right in user characteristics, each search result The product features of commodity and environmental characteristic are answered to determine payment probability, according to corresponding goods in the user characteristics and each search result Product features determine the user for corresponding to quotient in each search result by the 5th machine learning model being pre-created The reimbursement probability of product, the 5th machine learning model are used for the commodity according to corresponding goods in user characteristics, each search result Feature and environmental characteristic determine reimbursement probability, according to the payment probability and the reimbursement probability, determine the user for each The consumption determine the probability is the consumption intention data by the consumption probability of corresponding goods in search result.
Wherein, create the mode of the 4th machine learning model or the 5th machine learning model, can with create the first machine The mode of learning model is similar, no longer repeats one by one herein.
Step 205, the cloud platform is ranked up the multiple search result according to the consumption intention data, and Search result after sequence is supplied to user.
In embodiments of the present invention, it is preferred that since environmental characteristic can illustrate environment locating for user or commodity, than Such as, user is respectively at a distance from corresponding goods in each search result etc., therefore, in order to make the sequence to commodity and environmental characteristic phase It coincide, further increases the accuracy being ranked up to search result, it can be in conjunction with the environmental characteristic and the consumption intention Data are ranked up the multiple search result, and the search result after sequence is supplied to user, thus rapidly and accurately Commodity are provided for user.
Wherein it is possible to be preferentially ranked up according to environmental characteristic to corresponding goods in each search result, for environmental characteristic Identical commodity are ranked up according still further to the consumption intention data.It is of course also possible to preferentially according to the consumption intention data to each Corresponding goods are ranked up in search result, and commodity identical for consumption intention data are arranged according still further to the environmental characteristic Sequence.
For example, user 2 is 50% to the consumption probability of commodity 1, the distance between user 2 and commodity 1 are 1 kilometer, and user 2 Consumption probability to commodity 2 is 20%, and the distance between user 2 and commodity 2 are 7 kilometers, consumption probability of the user 2 to commodity 3 It is 20%, the distance between user 2 and commodity 3 are 3 kilometers, and user 2 is 80% to the consumption probability of commodity 4, user 2 and commodity The distance between 4 be 4 kilometers.Therefore, it is preferentially ranked up according to consumption probability to by commodity 1, commodity 2, commodity 3 and commodity 4, Commodity 4, commodity 1, commodity 2, commodity 3 are obtained, since the consumption probability to commodity 2 with commodity 3 is identical, so according to user and quotient The distance between product are from being closely ranked up commodity 2 and commodity 3 to remote sequence to obtain commodity 4, commodity 1, commodity 3, commodity 4.
Certainly, if in the foregoing according to the product features of corresponding goods in user characteristics, each search result and Environmental characteristic determines that the user to the consumption probability of corresponding goods in each search result, can directly disappear according to this in this step Expense intention data is ranked up multiple search result, and the search result after sequence is supplied to the user.
In embodiments of the present invention, firstly, since may exist between different users and different commodity it is poor It is different, so that different users has different consumption demands to different commodity, so the user for submitting searching request, energy It is enough that multiple searching results are obtained according to the end article information search carried in searching request, according to the use carried in searching request Family mark gets corresponding user characteristics, then special according to the commodity of corresponding goods in the user characteristics and each search result Sign determines that the user, and then can will be according to the user to each for the consumption intention data of corresponding goods in each search result The consumption intention data of corresponding goods is ranked up search result in search result, and that is to say can be according to user to searching Different commodity consumption demand, targetedly opposite direction the user provide search result be ranked up, improve to search tie The accuracy of accuracy and provided search result that fruit is ranked up, and then also improve the standard for providing a user commodity True property.
Secondly, it is general to the payment of corresponding goods in each search result to calculate the user according to the first machine learning model Rate calculates the user to the reimbursement probability of corresponding goods in each search result according to the second machine learning model, and according to the use Family determines the user to corresponding goods in each search result the payment probability and reimbursement probability of corresponding goods in each search result Consumption probability.Using the consumption probability as the consumption intention data, the accuracy for determining consumption intention data really is improved And to the accuracy that search result is ranked up, also ensure that can preferentially to provide a user the customer consumption probability higher Commodity, to ensure the interests of businessman and user simultaneously.
Furthermore it is possible to determine that the user is general to the consumption of corresponding goods in each search result according to third machine learning model Rate, reduces the quantity of the machine mould of needs, so that time cost needed for saving creation and training machine model, improves It determines the efficiency of consumption intention data, and then improves the efficiency being ranked up to search result and provide a user search knot The efficiency of fruit.
It should be noted that for the aforementioned method embodiment, for simple description, therefore, it is stated as a series of Combination of actions, but those skilled in the art should understand that, the present invention is not limited by the sequence of acts described, because according to According to the present invention, some steps may be performed in other sequences or simultaneously.Secondly, those skilled in the art should also know that, The embodiments described in the specification are all preferred embodiments, and related movement is not necessarily essential to the invention.
Embodiment three
Referring to Fig. 3, the structural block diagram of one of embodiment of the present invention search results ranking device is shown.Described search Sort result device includes:
Receiving module 301 carries user identifier and end article letter for receiving searching request in described search request Breath；
First determining module 302, for obtaining multiple search results according to the end article information search, determination is respectively searched The product features of corresponding goods in hitch fruit；
First obtains module 303, for obtaining user characteristics according to the user identifier；
Second determining module 304, for the product features according to corresponding goods in the user characteristics and each search result, Determine user for the consumption intention data of corresponding goods in each search result；
Sorting module 305, for being ranked up to the multiple search result according to the consumption intention data.
Optionally, second determining module includes:
First determines submodule, for the product features according to corresponding goods in the user characteristics and each search result, By the first machine learning model being pre-created, determine that the user is general for the payment of corresponding goods in each search result Rate, first machine learning model are used to be determined according to the product features of corresponding goods in user characteristics and each search result and pay Money probability；
Second determines submodule, for the product features according to corresponding goods in the user characteristics and each search result, By the second machine learning model being pre-created, determine that the user is general for the reimbursement of corresponding goods in each search result Rate, second machine learning model are used to be moved back according to the product features determination of corresponding goods in user characteristics and each search result Money probability；
Third determines submodule, for determining the user for each according to the payment probability and the reimbursement probability The consumption determine the probability is the consumption intention data by the consumption probability of corresponding goods in search result.
Optionally, second determining module further include:
First acquisition submodule, the history Access Events for being directed to commodity for obtaining multiple users, the history access thing Part includes payment event and arrearage event；
First computational submodule, for extraction user characteristics and product features from the history Access Events, and according to History Access Events calculate each user to the payment probability of commodity；
First creation submodule, is used for according to extracted product features, user characteristics and payment probability calculated, Create first machine learning model.
Optionally, second determining module further include:
Second acquisition submodule, the history Access Events for being directed to commodity for obtaining multiple users, the history access thing Part includes refund event and non-refund event；
Second computational submodule, for extraction user characteristics and product features from the history Access Events, and according to History Access Events calculate each user to the reimbursement probability of commodity；
Second creation submodule, is used for according to extracted product features, user characteristics and reimbursement probability calculated, Create second machine learning model.
Optionally, second determining module includes:
4th determines submodule, for the product features according to corresponding goods in the user characteristics and each search result, By the third machine learning model being pre-created, determine that the user is general for the payment of corresponding goods in each search result Rate, the third machine learning model are used to be disappeared according to the product features determination of corresponding goods in user characteristics and each search result Take probability；
5th determines submodule, for being the consumption intention data by the consumption determine the probability.
Optionally, the second determining module further include:
Third acquisition submodule, the history Access Events for being directed to commodity for obtaining multiple users, the history access thing Part includes payment event, arrearage event, refund event and non-refund event；
Third computational submodule, for extraction user characteristics and product features from the history Access Events, and according to History Access Events calculate each user to the consumption probability of commodity；
Third creates submodule, is used for according to extracted product features, user characteristics and consumption probability calculated, Create the third machine learning model.
Optionally, described device further include:
Second obtains module, and for obtaining environmental characteristic, the environmental characteristic includes that the user ties with each search respectively The distance between corresponding goods in fruit；
The sorting module includes:
Sorting sub-module, for being carried out to the multiple commodity in conjunction with the environmental characteristic and the consumption intention data Sequence.
Optionally, first determining module includes:
6th determines submodule, and for determining, commodity of the corresponding goods in the second preset time period are special in each search result Sign, second preset time period are using current time as the period in the third preset duration of starting point；
Described first, which obtains module, includes:
4th acquisition submodule, for obtaining the spy of the user in second preset time period according to the user identifier Sign.
In embodiments of the present invention, due to may be had differences between different users and different commodity, from And different users has different consumption demands to different commodity, so the user for submitting searching request, Neng Gougen Multiple searching results are obtained according to the end article information search carried in searching request, are marked according to the user carried in searching request Knowledge gets corresponding user characteristics, then according to the product features of corresponding goods in the user characteristics and each search result, really The fixed user is directed to the consumption intention data of corresponding goods in each search result, and then can will be tied according to the user to each search The consumption intention data of corresponding goods is ranked up search result in fruit, and that is to say can be according to user to the difference searched The consumption demand of commodity, the search result that targetedly the opposite direction user provides are ranked up, improve and carry out to search result The accuracy of the accuracy of sequence and provided search result, and then also improve the accuracy for providing a user commodity.
For above-mentioned apparatus embodiment, since it is basically similar to the method embodiment, so be described relatively simple, The relevent part can refer to the partial explaination of embodiments of method.
Example IV
The present invention also provides a kind of memory, processor and store the meter that can be run on a memory and on a processor Calculation machine program, the processor realize each process such as the search result ordering method in aforementioned when executing described program, and Identical technical effect can be reached, to avoid repeating, no longer repeated one by one herein.
Embodiment five
The embodiment of the present invention also provides a kind of computer readable storage medium, and meter is stored on computer readable storage medium Calculation machine program realizes each process such as the search result ordering method in aforementioned when the computer program is executed by processor, And identical technical effect can be reached, to avoid repeating, no longer repeat one by one herein.Wherein, described computer-readable to deposit Storage media, such as read-only memory (Read-Only Memory, abbreviation ROM), random access memory (Random Access Memory, abbreviation RAM), magnetic or disk etc..
All the embodiments in this specification are described in a progressive manner, the highlights of each of the examples are with The difference of other embodiments, the same or similar parts between the embodiments can be referred to each other.
It would have readily occurred to a person skilled in the art that: any combination application of above-mentioned each embodiment is all feasible, therefore Any combination between above-mentioned each embodiment is all embodiment of the present invention, but this specification exists as space is limited, This is not just detailed one by one.
Provided herein based on mobile terminal phone report scheme not with any certain computer, virtual system or its Its equipment is inherently related.Various general-purpose systems can also be used together with teachings based herein.As described above, it constructs Structure required by system with the present invention program is obvious.In addition, the present invention is not also directed to any certain programmed Language.It should be understood that can use various programming languages realizes summary of the invention described herein, and above to specific language Say that done description is in order to disclose the best mode of carrying out the invention.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail And technology, so as not to obscure the understanding of this specification.
Similarly, it should be understood that in order to simplify the disclosure and help to understand one or more of the various inventive aspects, Above in the description of exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes In example, figure or descriptions thereof.However, the disclosed method should not be interpreted as reflecting the following intention: i.e. required to protect Shield the present invention claims features more more than feature expressly recited in each claim.More precisely, such as right As claim reflects, inventive aspect is all features less than single embodiment disclosed above.Therefore, it then follows tool Thus claims of body embodiment are expressly incorporated in the specific embodiment, wherein each claim conduct itself Separate embodiments of the invention.
Those skilled in the art will understand that can be carried out adaptively to the module in the equipment in embodiment Change and they are arranged in one or more devices different from this embodiment.It can be the module or list in embodiment Member or component are combined into a module or unit or component, and furthermore they can be divided into multiple submodule or subelement or Sub-component.Other than such feature and/or at least some of process or unit exclude each other, it can use any Combination is to all features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed All process or units of what method or apparatus are combined.Unless expressly stated otherwise, this specification is (including adjoint power Benefit require, abstract and attached drawing) disclosed in each feature can carry out generation with an alternative feature that provides the same, equivalent, or similar purpose It replaces.
In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments In included certain features rather than other feature, but the combination of the feature of different embodiments mean it is of the invention Within the scope of and form different embodiments.For example, in detail in the claims, embodiment claimed it is one of any Can in any combination mode come using.
Various component embodiments of the invention can be implemented in hardware, or to run on one or more processors Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice Microprocessor or digital signal processor (DSP) come realize some in goods providers case according to an embodiment of the present invention or The some or all functions of person's whole component.The present invention is also implemented as one for executing method as described herein Point or whole device or device programs (for example, computer program and computer program product).Such this hair of realization Bright program can store on a computer-readable medium, or may be in the form of one or more signals.It is such Signal can be downloaded from an internet website to obtain, and is perhaps provided on the carrier signal or is provided in any other form.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims, Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch To embody.The use of word first, second, and third does not indicate any sequence.These words can be construed to title.

Claims (10)

1. a kind of search result ordering method characterized by comprising
Searching request is received, carries user identifier and end article information in described search request；
Multiple search results are obtained according to the end article information search, determine that the commodity of corresponding goods in each search result are special Sign；
User characteristics are obtained according to the user identifier；
According to the product features of corresponding goods in the user characteristics and each search result, determine user in each search result The consumption intention data of corresponding goods；
According to the consumption intention data, the multiple search result is ranked up；
The product features according to corresponding goods in the user characteristics and each search result determine that user ties for each search The consumption intention data of corresponding goods includes: in fruit
According to the product features of corresponding goods in the user characteristics and each search result, pass through the first engineering being pre-created Model is practised, determines the user for the payment probability of corresponding goods in each search result, first machine learning model is used Payment probability is determined in the product features according to corresponding goods in user characteristics and each search result；
According to the product features of corresponding goods in the user characteristics and each search result, pass through the second engineering being pre-created Model is practised, determines the user for the reimbursement probability of corresponding goods in each search result, second machine learning model is used In determining reimbursement probability according to the product features of corresponding goods in user characteristics and each search result；
According to the payment probability and the reimbursement probability, determine the user for the consumption of corresponding goods in each search result The consumption determine the probability is the consumption intention data by probability.
2. search result ordering method according to claim 1, which is characterized in that it is described according to the user characteristics and The product features of corresponding goods in each search result determine user's needle by the first machine learning model being pre-created Before the payment probability of corresponding goods in each search result, further includes:
The history Access Events that multiple users are directed to commodity are obtained, the history Access Events include payment event and arrearage thing Part；
User characteristics and product features are extracted from the history Access Events, and calculate each user couple according to history Access Events The payment probability of commodity；
According to extracted product features, user characteristics and payment probability calculated, the first machine learning mould is created Type.
3. search result ordering method according to claim 1, which is characterized in that it is described according to the user characteristics and The product features of corresponding goods in each search result determine user's needle by the second machine learning model being pre-created Before the reimbursement probability of corresponding goods in each search result, further includes:
The history Access Events that multiple users are directed to commodity are obtained, the history Access Events include refund event and non-reimbursement thing Part；
User characteristics and product features are extracted from the history Access Events, and calculate each user couple according to history Access Events The reimbursement probability of commodity；
According to extracted product features, user characteristics and reimbursement probability calculated, the second machine learning mould is created Type.
4. search result ordering method according to claim 1, which is characterized in that described according to the user characteristics and each The product features of corresponding goods in search result determine user for the consumption intention data packet of corresponding goods in each search result It includes:
According to the product features of corresponding goods in the user characteristics and each search result, pass through the third engineering being pre-created Model is practised, determines the user for the payment probability of corresponding goods in each search result, the third machine learning model is used Consumption probability is determined in the product features according to corresponding goods in user characteristics and each search result；
It is the consumption intention data by the consumption determine the probability.
5. search result ordering method according to claim 4, which is characterized in that it is described according to the user characteristics and The product features of corresponding goods in each search result determine user's needle by the third machine learning model being pre-created Before the payment probability of corresponding goods in each search result, further includes:
The history Access Events that multiple users are directed to commodity are obtained, the history Access Events include payment event, arrearage thing Part, refund event and non-refund event；
User characteristics and product features are extracted from the history Access Events, and calculate each user couple according to history Access Events The consumption probability of commodity；
According to extracted product features, user characteristics and consumption probability calculated, the third machine learning mould is created Type.
6. search result ordering method according to claim 1, which is characterized in that further include:
Obtain environmental characteristic, the environmental characteristic include the user respectively between corresponding goods in each search result away from From；
It is described according to the consumption intention data, the multiple search result be ranked up include:
In conjunction with the environmental characteristic and the consumption intention data, the multiple commodity are ranked up.
7. search result ordering method according to claim 1, which is characterized in that described according to the end article information Search obtains multiple search results, determines that the product features of corresponding goods in each search result include:
Determine product features of the corresponding goods in the second preset time period in each search result, second preset time period is Using current time as the period in the third preset duration of starting point；
It is described to include: according to user identifier acquisition user characteristics
The user characteristics in second preset time period are obtained according to the user identifier.
8. a kind of search results ranking device characterized by comprising
Receiving module carries user identifier and end article information in described search request for receiving searching request；
First determining module determines each search result for obtaining multiple search results according to the end article information search The product features of middle corresponding goods；
First obtains module, for obtaining user characteristics according to the user identifier；
Second determining module is determined and is used for the product features according to corresponding goods in the user characteristics and each search result Consumption intention data of the family for corresponding goods in each search result；
Sorting module, for being ranked up to the multiple search result according to the consumption intention data；
Second determining module is also used to be passed through according to the product features of corresponding goods in the user characteristics and each search result The first machine learning model being pre-created determines the user for the payment probability of corresponding goods in each search result, institute It states the first machine learning model and determines that payment is general for the product features according to corresponding goods in user characteristics and each search result Rate；
According to the product features of corresponding goods in the user characteristics and each search result, pass through the second engineering being pre-created Model is practised, determines the user for the reimbursement probability of corresponding goods in each search result, second machine learning model is used In determining reimbursement probability according to the product features of corresponding goods in user characteristics and each search result；
According to the payment probability and the reimbursement probability, determine the user for the consumption of corresponding goods in each search result The consumption determine the probability is the consumption intention data by probability.
9. a kind of electronic equipment including memory, processor and stores the calculating that can be run on a memory and on a processor Machine program, which is characterized in that the processor realizes any one the method for claim 1 to 7 when executing described program Step.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor The step of any one the method for claim 1 to 7 is realized when execution.
CN201810128612.7A 2018-02-08 2018-02-08 Search result sort method, device, electronic equipment and storage medium Active CN108415970B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810128612.7A CN108415970B (en) 2018-02-08 2018-02-08 Search result sort method, device, electronic equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810128612.7A CN108415970B (en) 2018-02-08 2018-02-08 Search result sort method, device, electronic equipment and storage medium

Publications (2)

Publication Number Publication Date
CN108415970A CN108415970A (en) 2018-08-17
CN108415970B true CN108415970B (en) 2019-07-30

Family

ID=63127018

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810128612.7A Active CN108415970B (en) 2018-02-08 2018-02-08 Search result sort method, device, electronic equipment and storage medium

Country Status (1)

Country Link
CN (1) CN108415970B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109242644A (en) * 2018-10-10 2019-01-18 深圳市心版图科技有限公司 Method of Commodity Recommendation, terminal and computer readable storage medium
CN110569439A (en) * 2019-09-10 2019-12-13 拉扎斯网络科技（上海）有限公司 Entity display method, entity display device, storage medium and electronic equipment
CN111091442A (en) * 2019-11-25 2020-05-01 北京一七网络技术有限公司 Demand information synchronous collection method based on user query operation
CN114154046B (en) * 2022-02-07 2022-08-12 深圳华强电子交易网络有限公司 Website search ranking method and system

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100479613B1 (en) * 1999-07-05 2005-03-30 엘지전자 주식회사 Method of controlling image feature weight using auto relevance feedback in content based image retrieval
CN105208113A (en) * 2015-08-31 2015-12-30 北京百度网讯科技有限公司 Information pushing method and device
CN105426528B (en) * 2015-12-15 2018-04-06 中南大学 A kind of retrieval ordering method and system of commodity data
CN106776860A (en) * 2016-11-28 2017-05-31 北京三快在线科技有限公司 One kind search abstraction generating method and device
CN107424043B (en) * 2017-06-15 2021-06-25 北京三快在线科技有限公司 Product recommendation method and device and electronic equipment

Also Published As

Publication number Publication date
CN108415970A (en) 2018-08-17

Similar Documents

Publication Publication Date Title
CN108415970B (en) Search result sort method, device, electronic equipment and storage medium
JP6313334B2 (en) Method and apparatus for generating web page content
CN103473230B (en) Service area determines that method, logistics service provider recommend method and related device
TWI603273B (en) Method and device for placing information search
US20120054113A1 (en) Systems and methods to determine item disposal options
CN110322277A (en) Product information methods of exhibiting and device, electronic equipment
CN104834641A (en) Processing method of network medium information and relevant system
CN106991108A (en) The method for pushing and device of a kind of information
CN109299981A (en) A kind of advertisement recommended method and device
CN107507037B (en) Server, gift pushing method of gift box and storage medium
CN106056408A (en) Method and device for tracking promoter
CN106445954A (en) Business object display method and apparatus
CN109241455B (en) Recommended object display method and device
CN110222285A (en) The methods of exhibiting of reading page calculates equipment and computer storage medium
AU2020245252A1 (en) Computer-implemented method for arranging hyperlinks on a graphical user-interface
CN108182587A (en) A kind of electric business platform brush single act detection method and system
WO2013173194A1 (en) A user recommendation method and device
JP5249415B2 (en) Method and apparatus for providing data statistics
CN109598526A (en) The analysis method and device of media contribution
CN111680213A (en) Information recommendation method, data processing method and device
CN109214583A (en) The prediction technique and device of shop popularity value
US7587340B2 (en) Method and apparatus for selling with short-bidding on goods
KR101312406B1 (en) Online donation method and system
CN109688178A (en) Recommended method, device and equipment
CN108256880A (en) A kind of ad traffic response method, device and equipment

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
GR01 Patent grant
GR01 Patent grant