CN117290392A - Product recommendation method and device, electronic equipment and storage medium - Google Patents
Product recommendation method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN117290392A CN117290392A CN202311227793.6A CN202311227793A CN117290392A CN 117290392 A CN117290392 A CN 117290392A CN 202311227793 A CN202311227793 A CN 202311227793A CN 117290392 A CN117290392 A CN 117290392A
- Authority
- CN
- China
- Prior art keywords
- product recommendation
- product
- cache
- strategy
- policy
- 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
- 238000000034 method Methods 0.000 title claims abstract description 57
- 230000004044 response Effects 0.000 claims description 16
- 238000013507 mapping Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 10
- 238000005516 engineering process Methods 0.000 abstract description 3
- 239000000047 product Substances 0.000 description 447
- 241000508269 Psidium Species 0.000 description 23
- 230000008030 elimination Effects 0.000 description 10
- 238000003379 elimination reaction Methods 0.000 description 10
- 239000012467 final product Substances 0.000 description 8
- 238000005457 optimization Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000006399 behavior Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- RYYVLZVUVIJVGH-UHFFFAOYSA-N caffeine Chemical compound CN1C(=O)N(C)C(=O)C2=C1N=CN2C RYYVLZVUVIJVGH-UHFFFAOYSA-N 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 210000001503 joint Anatomy 0.000 description 2
- LPHGQDQBBGAPDZ-UHFFFAOYSA-N Isocaffeine Natural products CN1C(=O)N(C)C(=O)C2=C1N(C)C=N2 LPHGQDQBBGAPDZ-UHFFFAOYSA-N 0.000 description 1
- 229960001948 caffeine Drugs 0.000 description 1
- VJEONQKOZGKCAK-UHFFFAOYSA-N caffeine Natural products CN1C(=O)N(C)C(=O)C2=C1C=CN2C VJEONQKOZGKCAK-UHFFFAOYSA-N 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2455—Query execution
- G06F16/24552—Database cache management
-
- 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
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Finance (AREA)
- Data Mining & Analysis (AREA)
- Accounting & Taxation (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a product recommendation method, a product recommendation device, electronic equipment and a computer readable storage medium, and relates to the field of data processing. The recommendation method comprises the following steps: responding to the product recommendation service call, and acquiring product recommendation strategy information from the multi-level cache according to the user characteristics; if the product recommendation strategy information is not recorded in the multi-level cache, acquiring the product recommendation strategy information from a database; acquiring recommended product information according to the product recommendation strategy information for product recommendation, wherein the multi-level cache comprises but is not limited to: a first cache and a second cache. At least solves the problem that the high concurrent access of the product recommendation service in the related technology causes excessive processing pressure on background service and database. The method is suitable for scenes such as product marketing recommendation and the like.
Description
Technical Field
The present invention relates to the field of data processing, and in particular, to a product recommendation method, apparatus, electronic device, and computer readable storage medium.
Background
In the prior art, a single service deployment is used for a product recommendation service interface, a traditional chimney type system architecture is adopted, a plurality of functional modules are all vertical system architectures, service resources are shared among different modules, and service performance can be reduced in a high concurrency environment. Therefore, in the prior art, different service capacities are decoupled, the product recommendation service is independently deployed by a distributed service framework, the concurrent processing capacity is improved, and other service modules are not influenced.
However, when the concurrency capability of the product recommendation service is larger, the processing capability of the background database becomes a bottleneck of access performance, and as the pressure of the database increases, more slow SQL (Structured Query Language ) appears, and the processing time delay becomes longer. In addition, no cache condition occurs when the product recommendation service is redeployed or restarted, so that a large number of requests penetrate into the database, and the query speed is low and even the system stability is influenced. Therefore, how to optimize the access performance of the product recommendation service interface and prevent the high concurrent access from causing excessive processing pressure on the background service and the database becomes a key problem to be solved.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a product recommending method, a device, electronic equipment and a computer readable storage medium aiming at the defects in the prior art, wherein the method can realize the optimization of the access performance of a product recommending service interface, lighten the overlarge processing pressure of high concurrency access on background service and a database and improve the data processing capacity of the product recommending service in a high concurrency environment.
In a first aspect, the present invention provides a method for recommending a product, comprising: responding to the product recommendation service call, and acquiring product recommendation strategy information from the multi-level cache according to the user characteristics; if the product recommendation strategy information is not recorded in the multi-level cache, acquiring the product recommendation strategy information from a database; acquiring recommended product information according to the product recommendation strategy information for product recommendation, wherein the multi-level cache comprises but is not limited to: a first cache and a second cache.
Preferably, the responding to the product recommendation service call obtains product recommendation policy information from the multi-level cache according to the user characteristics, and specifically includes: responding to the product recommendation service call, and inquiring a product recommendation strategy list from the multi-level cache according to the user characteristics; respectively acquiring product recommendation policy information corresponding to each policy in the product recommendation policy list from the first cache according to the product recommendation policy list; if the product recommendation policy information corresponding to each policy in the product recommendation policy list is not recorded in the first cache, the product recommendation policy information corresponding to each policy in the product recommendation policy list is obtained from the second cache.
Preferably, after the product recommendation service call is responded, the product recommendation method further comprises the following steps of: if the product recommendation policy list is not queried in the multi-level cache, querying the product recommendation policy list from the database according to the user characteristics.
Preferably, after the product recommendation policy information corresponding to each policy in the product recommendation policy list is obtained from the second cache if the product recommendation policy information corresponding to each policy in the product recommendation policy list is not recorded in the first cache, the product recommendation method further includes: if the product recommendation strategy information is obtained from the second cache, cluster broadcasting is carried out through the message queue, and the product recommendation strategy information is written into the first cache; if the product recommendation strategy information is obtained from the database, the product recommendation strategy information is written into the second cache, cluster broadcasting is carried out through the message queue, and the product recommendation strategy information is written into the first cache.
Preferably, if the product recommendation policy information is not recorded in the multi-level cache, the product recommendation policy information is obtained from the database, which specifically includes: and respectively acquiring product recommendation strategy information corresponding to each strategy in the product recommendation strategy list from the database according to the product recommendation strategy list.
Preferably, after the recommended product information is obtained according to the product recommendation policy information, the product recommendation method further includes: recording a product recommendation strategy identification in the product recommendation strategy information to obtain a recommendation record; counting the number of recommendation strategy identifiers of each product in the recommendation record; if the number of the product recommendation strategy identifiers exceeds the preset value, summarizing the product recommendation strategies corresponding to the product recommendation strategy identifiers to obtain a recent frequent query strategy list.
Preferably, before the recommended product information is obtained from the multi-level cache according to the user characteristics in response to the product recommendation service call, the product recommendation method further includes: creating a product recommendation strategy according to the recommended product; defining target user group characteristics according to a product recommendation strategy; inquiring user characteristics according to the target user group characteristics; building a mapping relation between a product recommendation strategy and user characteristics, and storing the mapping relation into a database; summarizing product recommendation strategies according to the same user characteristics to obtain a product recommendation strategy list; and inquiring recent hot spot data, and writing the hot spot data into the multi-level cache, wherein the hot spot data comprises a recent new product recommendation strategy and a recent frequent inquiry strategy list.
In a second aspect, the present invention also provides a recommendation device for a product, including: the system comprises a first acquisition module, a second acquisition module and a recommendation module, wherein the first acquisition module is used for responding to product recommendation service call and acquiring product recommendation strategy information from a multi-level cache according to user characteristics, the first acquisition module is connected with the first acquisition module and used for acquiring the product recommendation strategy information from a database if the product recommendation strategy information is not recorded in the multi-level cache, and the recommendation module is connected with the first acquisition module and the second acquisition module and used for acquiring the recommendation product information according to the product recommendation strategy information so as to be used for product recommendation, wherein the multi-level cache comprises but is not limited to: a first cache and a second cache.
In a third aspect, the present invention also provides an electronic device comprising a memory having a computer program stored therein and a processor arranged to run the computer program to implement the method of recommending a product provided in the first aspect described above.
In a fourth aspect, the present invention also provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the recommendation method for a product provided in the first aspect.
According to the recommending method, the recommending device, the electronic equipment and the computer readable storage medium of the product, part of product recommending strategy information is dynamically cached in the multi-level cache in real time, and query is preferentially requested in the multi-level cache, so that query pressure on a database is relieved. Therefore, the invention can realize the optimization of the access performance of the product recommendation service interface, lighten the overlarge processing pressure on background service and databases caused by high concurrency access, and improve the data processing capacity of the product recommendation service in the high concurrency environment.
Drawings
FIG. 1 is a flowchart of a method for recommending a product according to embodiment 1 of the present invention;
FIG. 2 is a flowchart of a method for recommending a product according to embodiment 2 of the present invention;
FIG. 3 is a flowchart of a product recommendation deployment method according to embodiment 3 of the present invention;
FIG. 4 is a flowchart of a method for generating a product recommendation policy according to embodiment 4 of the present invention;
fig. 5 is a schematic structural diagram of a recommending apparatus for a product in embodiment 5.
Detailed Description
In order to make the technical scheme of the present invention better understood by those skilled in the art, the following detailed description of the embodiments of the present invention will be given with reference to the accompanying drawings.
It is to be understood that the specific embodiments and figures described herein are merely illustrative of the invention, and are not limiting of the invention.
It is to be understood that the various embodiments of the invention and the features of the embodiments may be combined with each other without conflict.
It is to be understood that only the portions relevant to the present invention are shown in the drawings for convenience of description, and the portions irrelevant to the present invention are not shown in the drawings.
It should be understood that each unit and module in the embodiments of the present invention may correspond to only one physical structure, may be formed by a plurality of physical structures, or may be integrated into one physical structure.
It will be appreciated that, without conflict, the functions and steps noted in the flowcharts and block diagrams of the present invention may occur out of the order noted in the figures.
It is to be understood that the flowcharts and block diagrams of the present invention illustrate the architecture, functionality, and operation of possible implementations of systems, apparatuses, devices, methods according to various embodiments of the present invention. Where each block in the flowchart or block diagrams may represent a unit, module, segment, code, or the like, which comprises executable instructions for implementing the specified functions. Moreover, each block or combination of blocks in the block diagrams and flowchart illustrations can be implemented by hardware-based systems that perform the specified functions, or by combinations of hardware and computer instructions.
It should be understood that the units and modules related in the embodiments of the present invention may be implemented by software, or may be implemented by hardware, for example, the units and modules may be located in a processor.
Example 1:
the embodiment provides a product recommending method which is applied to recommending products to a target user group, wherein the recommended products are marketing products developed by operators or companies. The recommended product information is specific information of a recommended product, and the recommended product information includes: recommended product ID, recommended product name, recommended product profile, etc. The product recommendation policy is a marketing policy for an operator or company to circumscribe a marketing crowd for a newly developed marketing product to realize marketing product sales. The product recommendation policy information is specific information of a product recommendation policy, and the product recommendation policy information includes: product recommendation policy ID, product recommendation policy name, target group feature ID, recommended product ID, recommended contact code, etc. The marketing crowd characteristic pointed out by the target user group characteristic comprises: user behavior characteristics, region characteristics, age characteristics and the like, wherein the user behavior characteristics are the operating frequency of a user for various applications. User characteristics are identification information of a user, including but not limited to: mobile phone number, identification card number. As shown in fig. 1, the recommendation method of the product includes:
Step S101, in response to the product recommendation service call, obtaining product recommendation policy information from a multi-level cache according to the user characteristics, where the multi-level cache includes but is not limited to: a first cache and a second cache.
In this embodiment, the multi-level cache may refer to a second-level cache, a third-level cache, and the like, where the second-level cache includes: the first buffer memory and the second buffer memory, tertiary buffer memory includes: the first cache, the second cache and the third cache. In this embodiment, the multi-level cache refers to a second-level cache, the first cache is a server local memory Map cache (i.e. a local cache), the reading speed is fast, no network transmission is provided, and the server local memory Map cache can adopt different technical frameworks, including but not limited to: open source frameworks such as Guava Cache, google Caffeine, ehCache and the like. In this embodiment, the level one Cache takes a Guava Cache memory Cache as an example. The second cache is a Redis (Remote Dictionary Server, remote dictionary service) non-relational memory cache, network transmission is needed, a distributed architecture is needed, reading speed is high, and storage capacity is large. In the case where no query database is deployed after the secondary cache, the company or platform may use the memory relational database as the third cache, deploying the tertiary cache, but using the memory relational database as the third cache may bring additional complexity and cost: (a) There is a need to deploy and manage database servers and ensure their reliability and scalability. (b) Additional code needs to be written to handle interactions with the database. (c) The memory relational database is suitable for persisting data and providing functions such as querying and transaction support, which may be overly heavy for simple key-value caching. In summary, although it is theoretically possible to use the memory relational database as the third cache and deploy the third cache, in practical situations, it is more common and recommended to use the second cache to meet the multi-level cache requirement.
Optionally, in step S101: in response to the product recommendation service call, before the recommended product information is obtained from the multi-level cache according to the user characteristics, the product recommendation method further comprises the following steps:
step S1001, creating a product recommendation policy according to the recommended product.
Step S1002, the target user group characteristics are defined according to the product recommendation strategy.
In this embodiment, an operator or a company plans a product recommendation policy through a full-volume customer value operation platform (simply referred to as a full-guest platform), and marks out a target user group marketing recommended product, i.e. selects a plurality of recommended products, marks out a target user group respectively, generates a plurality of product recommendation policies, and stores product recommendation policy information and recommended product information into a database. Such as: recommended product 1:511206049563 [ X directed traffic membership package (forward) ], recommended product 2:51202209261701 (X video VIP member-moon card), the delineated target group is characterized by Guangzhou, 20-35 age, commonly used application X video, Y music and the like, and the number of users is 5000 ten thousand; 2 product recommendation policies a and B will be generated, one product recommendation policy corresponding to each recommended product (i.e., product recommendation policy a for marketing of recommended product 1 and product recommendation policy B for marketing of recommended product 2). One recommended product may be selected by different product recommendation strategies to market different target user groups, such as two recommended products: three product recommendation strategies D, E, F can be established for the recommended products 3 and 4, and the recommended products 3, 4 and 3 are selected respectively, and the difference between the product recommendation strategies D and F is that the delineated target user groups are different. The all-passenger platform is based on user behavior characteristics, performs accurate modeling by adopting a big data algorithm, and provides accurate marketing product recommendation service for clients through strategy planning, data insight, real-time tag library and all-channel collaboration and through the full life cycle of the clients. The whole guest platform is in butt joint with four large types of contacts of on-line, off-line, outbound and short messages of the whole country, more than 600 total intensive contacts of headquarters and main operation contacts of provinces are covered, and the requirement on concurrency of a product recommendation service interface is quite large. And as the contact of the operator is accessed to the product recommendation interface of the all-passenger platform, the user access scale is rapidly increased, the daily average access amount reaches more than ten millions of levels, and meanwhile, part of contacts require that the interface access time delay cannot exceed 200ms.
Step S1003, inquiring the user characteristics according to the target user group characteristics.
In this embodiment, the users are filtered from the operator or company related system/database according to the target user group characteristics, so as to obtain filtered user characteristics, for example, according to the above description: the user is filtered by the related system/database of the operator by Guangzhou, age 20-35, common application X video, Y music and the like, the number of the user is assumed to be 5000 ten thousand, and the mobile phone number corresponding to the 5000 ten thousand is taken.
Step S1004, a mapping relation between the product recommendation strategy and the user characteristics is constructed, and the mapping relation is stored in a database.
In this embodiment, a mapping relationship between 5000 ten thousand mobile phone numbers and product recommendation policies a and B is constructed, and the mapping relationship between 5000 ten thousand mobile phone numbers and product recommendation policies a and B is stored in a Database, and an external contact is waited for initiating product recommendation service call, wherein the storage format is a record of one mobile phone number, for example, 100 ten thousand user recommendation policies a are selected, 100 ten thousand records are stored in the Database table, the records include a mobile phone number field and a policy ID field, and the actual storage of the mapping relationship can use technologies such as HBase (Hadoop Database) and the like, so that the time consumption of inquiry is reduced.
Step S1005, summarizing the product recommendation strategies according to the same user characteristics to obtain a product recommendation strategy list.
In this embodiment, according to the mobile phone number field, a corresponding product recommendation policy ID may be queried, and a plurality of product recommendation policies corresponding to the same user may form a list, that is, a product recommendation policy list.
Step S1006, inquiring recent hot spot data, and writing the hot spot data into a multi-level cache, wherein the hot spot data comprises a recent newly created product recommendation policy and a recent frequent inquiry policy list.
In this embodiment, in a preset period, a product recommendation policy is created according to recommended products, and a plurality of product recommendation policies obtained after target user group characteristics are defined, namely, a recently newly created product recommendation policy in recent hotspot data. Wherein the preset period includes, but is not limited to: three days, one week and one month.
When the product recommendation service is started, a SpringBoot starting script of an open source application framework is used for inquiring recent hot spot data (comprising a recent new product recommendation strategy and a recent frequent inquiry strategy list), the recent hot spot data is stored in a multi-level cache in advance, and the product recommendation service is started normally after the cache is ready. According to the method, the device and the system, the recent hot spot data are inquired, the hot spot data are stored in the multi-level cache in advance, and the problem that the pressure of the database is suddenly increased and the recommended service request is slow due to the fact that the instantaneous large flow directly flows to the database after the product recommended service is deployed and restarted is avoided.
Specifically, step S101: in response to the product recommendation service call, product recommendation policy information is obtained from the multi-level cache according to the user characteristics, including steps S1011-S1013:
in step S1011, a product recommendation policy list is queried from the multi-level cache according to the user characteristics in response to the product recommendation service call.
In this embodiment, the contact invokes the product recommendation service, inputs information such as a mobile phone number, and requests are uniformly sent to the server cluster through an Alibaba HSF (Alibaba High-speed Service Framework, an aliba High-speed service framework, which is a distributed RPC service framework widely used in aliba). The product recommendation service firstly queries a product recommendation policy list corresponding to a user in the Guava Cache memory Cache, and if the product recommendation policy list corresponding to the user is not queried in the Guava Cache memory Cache, the product recommendation policy list corresponding to the user is queried in the Reids Cache.
Alternatively, in step S1011: after the product recommendation policy list is queried from the multi-level cache according to the user characteristics in response to the product recommendation service call, the product recommendation method further comprises the following steps:
in step S1014, if the product recommendation policy list is not queried in the multi-level cache, the product recommendation policy list is queried from the database according to the user characteristics.
In this embodiment, the database service may use different cluster implementations or database frameworks, such as drags, oracle, SQL Server, etc. The present embodiment database takes the Drds database as an example.
If the product recommendation policy list corresponding to the user is not queried in the Guava Cache memory Cache and the references Cache, querying the product recommendation policy list corresponding to the user in the drop database, and finally obtaining the product recommendation policy list A, B, C, D such as the user 186 xxxxxxx.
Step S1012, product recommendation strategy information corresponding to each strategy in the product recommendation strategy list is obtained from the first cache according to the product recommendation strategy list.
Step S1013, if the product recommendation policy information corresponding to each policy in the product recommendation policy list is not recorded in the first cache, the product recommendation policy information corresponding to each policy in the product recommendation policy list is obtained from the second cache.
Step S102, if the product recommendation strategy information is not recorded in the multi-level cache, the product recommendation strategy information is obtained from the database.
Specifically, step S102: if the product recommendation policy information is not recorded in the multi-level cache, obtaining the product recommendation policy information from the database, including:
And respectively acquiring product recommendation strategy information corresponding to each strategy in the product recommendation strategy list from the database according to the product recommendation strategy list.
In this embodiment, after a product recommendation policy list is queried, a multithreading asynchronous query is adopted, for example, for a policy A, B, C, D in a user 186xxxxxxx product recommendation policy list, 4 threads are started to query product recommendation policy information, wherein each thread for querying product recommendation policy information is logically consistent:
(1) The product recommendation policy information is preferentially obtained from the Guava Cache, if the product recommendation policy information is obtained successfully, the product recommendation policy information is directly returned, the Guava Cache can set timeout time, maximum storage quantity and elimination policy, for example, the Cache quantity is set to be 500, and 1 value of the original Cache needs to be eliminated when the 501 th value is written into the Cache. The object (i.e., the originally cached 1 value) is eliminated according to an elimination policy such as: first-in first-out, i.e., eliminate the first stored value; and according to the elimination of the use amount, namely, finding the value with the smallest acquired use frequency after storage in 500 values, effectively eliminating the use, and the like, so as to reduce the memory consumption and ensure the cache hit rate.
(2) When the Guava Cache does not record the product recommendation strategy information, the product recommendation strategy information is acquired from a Redis cluster, and the Redis cluster supports large memory and high concurrency. In this embodiment, the Redis sets and stores the product recommendation policy information for 15 minutes, and the Redis also has a elimination policy, similar to that mentioned in the above (1), and the Redis is deleted when the Redis has insufficient memory, but the memory of the Redis is larger, and the elimination is generally performed only by using a timeout period.
(3) If the product recommendation strategy information is not obtained in Redis, the product recommendation strategy information is queried from the drags database.
According to the method, the product recommendation strategy list and the product recommendation strategy information are sequentially obtained in the multi-level cache (the first cache and the second cache) and the database (namely, the query is preferentially requested in the multi-level cache), so that the query pressure on the database is reduced, the optimization of the access performance of the product recommendation service interface is realized, the overlarge processing pressure on the background service and the database caused by high concurrent access is reduced, and the data processing capacity of the product recommendation service in the high concurrent environment is improved.
Optionally, in step S1013: if the product recommendation policy information corresponding to each policy in the product recommendation policy list is not recorded in the first cache, after the product recommendation policy information corresponding to each policy in the product recommendation policy list is obtained from the second cache, the product recommendation method further comprises the steps of:
in step S1015, if the product recommendation policy information is obtained from the second cache, the cluster broadcasting is performed through the message queue, and the product recommendation policy information is written into the first cache.
Step S1016, if the product recommendation policy information is obtained from the database, writing the product recommendation policy information into the second cache, performing cluster broadcasting through the message queue, and writing the product recommendation policy information into the first cache.
In this embodiment, after the product recommendation policy information is not obtained in the Guava Cache but is obtained in the Redis in the above (2), cluster broadcasting is performed through a RocketMQ message queue, and after receiving the message, the intra-cluster server stores the product recommendation policy information obtained from the Redis in the Guava Cache. The product recommendation policy information is not obtained in the Guava Cache and the Redis in the step (3), after the product recommendation policy information is obtained in the drags, the product recommendation policy information obtained in the drags is written into the Redis, cluster broadcasting is carried out through a RocketMQ message queue, and after the intra-cluster server receives the message, the product recommendation policy information obtained in the drags is stored in the Guava Cache. Other message queues, such as Kafka, rabbitMQ, etc., may also be employed for trunking broadcasting. According to the embodiment, when the first cache has no data, cluster broadcasting is carried out through the message queue, the synchronization of the hot spot data of the first cache and the hot spot data of the second cache are kept, the hot spot product recommendation strategy information can be hit in different machines of the cluster when the request is sent to, and the pressure of the second cache and the database is reduced.
Step S103, recommending product information is obtained according to the product recommending strategy information so as to be used for product recommendation.
In this embodiment, after the thread query waiting for querying the product recommendation policy information is completed, the final product recommendation policy information is assembled and returned, that is, the recommended product IDs in all the product recommendation policy information are acquired, a recommended product ID list (i.e., recommended product list) is obtained by summarizing, the recommended product information is queried from the multi-level cache according to the recommended product ID list, and if the recommended product information is not queried in the multi-level cache, the recommended product information is queried from the database for product recommendation.
Optionally, in step S103: the recommended product information is obtained according to the product recommendation policy information, so that after the recommended product is recommended, the product recommendation method further comprises step S1031-step S1033:
step S1031, recording a product recommendation strategy identification in the product recommendation strategy information to obtain a recommendation record;
step S1032, counting the number of the product recommendation strategy identifiers in the recommendation record;
step S1033, if the number of the product recommendation strategy identifiers exceeds the preset value, summarizing the product recommendation strategies corresponding to the product recommendation strategy identifiers to obtain a recent frequent query strategy list.
In this embodiment, assembling and returning the final product recommendation policy information, and further includes obtaining product recommendation policy IDs (i.e., product recommendation policy identifiers) in the final product recommendation policy information, recording the product recommendation policy IDs into a recommendation record table, warehousing the recommendation record table, simultaneously counting the number of each product recommendation policy ID in the recommendation record table, and collecting the product recommendation policy IDs with a large number in the recommendation record table to collect a recent frequent query policy list. According to the method, the device and the system, the product recommendation strategy triggering the product recommendation service is recorded and counted, data support is provided for subsequently updating hot spot data, the response speed of the product recommendation data is accelerated, and the user experience is improved.
According to the product recommendation method provided by the embodiment, part of product recommendation strategy information is dynamically cached in the multi-level cache in real time, query is preferentially requested in the multi-level cache, query pressure on a database is reduced, optimization of access performance of a product recommendation service interface is achieved, overlarge processing pressure on background service and the database caused by high concurrent access is reduced, and data processing capacity of the product recommendation service in a high concurrent environment is improved. By inquiring recent hot spot data and storing the hot spot data into the multi-level cache in advance, the problem that the pressure of the database is suddenly increased and the recommended service request is slow due to the fact that the instantaneous large-flow direct flow flows to the database after the product recommended service is deployed and restarted is avoided. And when the first cache has no data, cluster broadcasting is carried out through the message queue, the synchronization of the hot spot data of the first cache and the second cache is kept, the hot spot product recommendation strategy information can be hit in different machines of the cluster when the request is sent to, and the pressure of the second cache and the database is reduced. And recording and counting a product recommendation strategy triggering the product recommendation service, providing data support for subsequently updating hot spot data, accelerating the response speed of the product recommendation data and improving the user experience.
Example 2:
the embodiment provides a product recommendation method, as shown in fig. 2. The recommending method of the product comprises the following steps:
s1, calling a product recommendation service.
In this embodiment, before the product recommendation service is invoked, the open source application framework SpringBoot startup script is further used to query recent hot spot data (including recent new product recommendation policies and recent frequent query policy lists), and the recent hot spot data is stored in the local cache and the redis cache in advance, and the product recommendation service will be started normally after the cache is ready. In embodiment 1, a product recommendation policy is created according to the recommended product, and a plurality of product recommendation policies obtained after the characteristics of the target user group are defined, namely, a recently newly created product recommendation policy in recently hot spot data. Acquiring product recommendation strategy IDs (namely product recommendation strategy identifiers) in final product recommendation strategy information, recording the product recommendation strategy IDs into a recommendation record table, warehousing the recommendation record table, simultaneously counting the number of the product recommendation strategy IDs in the recommendation record table, taking the product recommendation strategy IDs with more numbers in the recommendation record table, and summarizing the product recommendation strategy IDs into a recent frequent query strategy list. According to the method, the device and the system, the recent hot spot data are inquired, the hot spot data are stored in the multi-level cache in advance, and the problem that the pressure of the database is suddenly increased and the recommended service request is slow due to the fact that the instantaneous large flow directly flows to the database after the product recommended service is deployed and restarted is avoided.
S2, the local cache acquires a user strategy ID list.
In this embodiment, the local cache is the first cache in the multi-level cache in embodiment 1. The local cache (i.e., server local memory Map cache) may employ different technical frameworks including, but not limited to: the local Cache in this embodiment takes the Guava Cache as an example, and has fast reading speed and no network transmission.
In this embodiment, the user policy ID list is the product recommendation policy list in embodiment 1.
S3, if no data exists in the local cache, acquiring a user strategy ID list from the Redis.
In this embodiment, redis is the second cache in the multi-level cache in embodiment 1. The contact calls a product recommendation service, inputs information such as a mobile phone number and the like, and requests are uniformly sent to a server cluster through an Alibaba HSF (Alibaba High-speed Service Framework, an Alibaba High-speed service framework, which is a distributed RPC service framework widely used in Alibaba). The product recommendation service firstly queries a product recommendation policy list corresponding to a user in the Guava Cache memory Cache, and if the product recommendation policy list corresponding to the user is not queried in the Guava Cache memory Cache, the product recommendation policy list corresponding to the user is queried in the Reids Cache.
S4, if the Redis has no data, acquiring a user strategy ID list from the database.
In this embodiment, the database service may use different cluster implementation manners or database frameworks, such as TiDB, tencerting cloud database, and the like, and database frameworks, such as drags, oracle, SQL Server, and the like. The database service of this embodiment takes the drags database as an example.
In this embodiment, if the product recommendation policy list corresponding to the user is not queried in the Guava Cache memory Cache and the references Cache, the product recommendation policy list corresponding to the user is queried in the drags database, and finally, the product recommendation policy list A, B, C, D, such as the user 186xxxxxxx, is obtained.
S5, after the user strategy id list is obtained, multi-thread inquiry strategy information is used.
S6, obtaining strategy information from the local cache.
S7, if no data exists in the local cache, acquiring strategy information from the Redis, and carrying out cluster broadcasting through the RocketMQ to synchronize the strategy information to the local cache.
S8, if the Redis has no data, acquiring strategy information from a database, and synchronizing the strategy information to a local cache through the RocketMQ for cluster broadcasting.
In this embodiment, the policy information is the product recommendation policy information in embodiment 1, and after the product recommendation policy list is queried, a multithreading asynchronous query is adopted, for example, for the policy A, B, C, D in the user 186xxxxxxx product recommendation policy list, 4 threads are started to query the product recommendation policy information, where each thread for querying the product recommendation policy information is logically consistent with each other:
(1) The product recommendation policy information is preferentially obtained from the Guava Cache, if the product recommendation policy information is obtained successfully, the product recommendation policy information is directly returned, the Guava Cache can set timeout time, maximum storage quantity and elimination policy, for example, the Cache quantity is set to be 500, and 1 value of the original Cache needs to be eliminated when the 501 th value is written into the Cache. The object (i.e., the originally cached 1 value) is eliminated according to an elimination policy such as: first-in first-out, i.e., eliminate the first stored value; and according to the elimination of the use amount, namely, finding the value with the smallest acquired use frequency after storage in 500 values, effectively eliminating the use, and the like, so as to reduce the memory consumption and ensure the cache hit rate.
(2) When the Guava Cache does not record the product recommendation strategy information, the product recommendation strategy information is acquired from a Redis cluster, and the Redis cluster supports large memory and high concurrency. In this embodiment, the Redis sets and stores the product recommendation policy information for 15 minutes, and the Redis also has a elimination policy, similar to that mentioned in the above (1), and the Redis is deleted when the Redis has insufficient memory, but the memory of the Redis is larger, and the elimination is generally performed only by using a timeout period.
(3) If the product recommendation strategy information is not obtained in Redis, the product recommendation strategy information is queried from the drags database.
According to the method, the product recommendation strategy list and the product recommendation strategy information are sequentially obtained in the multi-level cache (the first cache and the second cache) and the database (namely, the query is preferentially requested in the multi-level cache), so that the query pressure on the database is reduced, the optimization of the access performance of the product recommendation service interface is realized, the overlarge processing pressure on the background service and the database caused by high concurrent access is reduced, and the data processing capacity of the product recommendation service in the high concurrent environment is improved.
After the product recommendation policy information is not obtained in the Guava Cache but is obtained in the Redis, the product recommendation policy information is subjected to cluster broadcasting through a RocketMQ message queue, and after the intra-cluster server receives the message, the product recommendation policy information obtained from the Redis is stored in the Guava Cache. The product recommendation policy information is not obtained in the Guava Cache and the Redis in the step (3), after the product recommendation policy information is obtained in the drags, the product recommendation policy information obtained in the drags is written into the Redis, cluster broadcasting is carried out through a RocketMQ message queue, and after the intra-cluster server receives the message, the product recommendation policy information obtained in the drags is stored in the Guava Cache. Other message queues, such as Kafka, rabbitMQ, etc., may also be employed for trunking broadcasting. According to the embodiment, when the first cache has no data, cluster broadcasting is carried out through the message queue, the synchronization of the hot spot data of the first cache and the hot spot data of the second cache are kept, the hot spot product recommendation strategy information can be hit in different machines of the cluster when the request is sent to, and the pressure of the second cache and the database is reduced.
S9, combining strategy information, recording a product recommendation request, and returning a product recommendation result.
In this embodiment, after the thread query waiting for querying the product recommendation policy information is completed, the final product recommendation policy information is assembled and returned, that is, the recommended product IDs in all the product recommendation policy information are acquired, a recommended product ID list (i.e., recommended product list) is obtained by summarizing, the recommended product information is queried from the multi-level cache according to the recommended product ID list, and if the recommended product information is not queried in the multi-level cache, the recommended product information is queried from the database for product recommendation.
Assembling and returning the final product recommendation strategy information, further comprising the steps of obtaining product recommendation strategy IDs (namely product recommendation strategy identifiers) in the final product recommendation strategy information, recording the product recommendation strategy IDs into a recommendation record table, warehousing the recommendation record table, simultaneously counting the number of the product recommendation strategy IDs in the recommendation record table, taking the product recommendation strategy IDs with more numbers in the recommendation record table, and summarizing the product recommendation strategy IDs into a recent frequent query strategy list. According to the method, the device and the system, the product recommendation strategy triggering the product recommendation service is recorded and counted, data support is provided for subsequently updating hot spot data, the response speed of the product recommendation data is accelerated, and the user experience is improved.
According to the product recommendation method provided by the embodiment, part of product recommendation strategy information is dynamically cached in the multi-level cache in real time, query is preferentially requested in the multi-level cache, query pressure on a database is reduced, optimization of access performance of a product recommendation service interface is achieved, overlarge processing pressure on background service and the database caused by high concurrent access is reduced, and data processing capacity of the product recommendation service in a high concurrent environment is improved. By inquiring recent hot spot data and storing the hot spot data into the multi-level cache in advance, the problem that the pressure of the database is suddenly increased and the recommended service request is slow due to the fact that the instantaneous large-flow direct flow flows to the database after the product recommended service is deployed and restarted is avoided. And when the first cache has no data, cluster broadcasting is carried out through the message queue, the synchronization of the hot spot data of the first cache and the second cache is kept, the hot spot product recommendation strategy information can be hit in different machines of the cluster when the request is sent to, and the pressure of the second cache and the database is reduced. And recording and counting a product recommendation strategy triggering the product recommendation service, providing data support for subsequently updating hot spot data, accelerating the response speed of the product recommendation data and improving the user experience.
Example 3:
the present embodiment provides a product recommendation deployment method, as shown in fig. 3. The deployment method of the product recommendation comprises the following steps: s01, starting a product recommendation service.
S02, inquiring recent hot spot data, wherein the hot spot data comprises: newly created product recommendation policies and frequently queried product recommendation policies.
In this embodiment, in embodiment 1, a product recommendation policy is created according to the recommended product, and a plurality of product recommendation policies obtained after the target user group feature is outlined, namely, a recently newly created product recommendation policy in recent hotspot data. Acquiring product recommendation strategy IDs (namely product recommendation strategy identifiers) in final product recommendation strategy information, recording the product recommendation strategy IDs into a recommendation record table, warehousing the recommendation record table, simultaneously counting the number of the product recommendation strategy IDs in the recommendation record table, taking the product recommendation strategy IDs with more numbers in the recommendation record table, and summarizing the product recommendation strategy IDs into a recent frequent query strategy list.
S03, synchronizing the recent hot spot data to the Redis and the local cache.
In this embodiment, redis is the second cache in the multi-level cache in embodiment 1, and the local cache is the first cache in the multi-level cache in embodiment 1.
S04, product recommendation deployment is finished, and product recommendation service is ready.
In this embodiment, when the product recommendation service is started, the open source application framework SpringBoot startup script is used to query recent hot spot data (including recent new product recommendation policies and recent frequent query policy lists), and the recent hot spot data is stored in the multi-level cache in advance, and the product recommendation service will be started normally after the cache is ready. According to the method, the device and the system, the recent hot spot data are inquired, the hot spot data are stored in the multi-level cache in advance, and the problem that the pressure of the database is suddenly increased and the recommended service request is slow due to the fact that the instantaneous large flow directly flows to the database after the product recommended service is deployed and restarted is avoided.
According to the product recommendation deployment method, the recent hot spot data are inquired, the hot spot data are stored in the multi-level cache in advance, and the problem that the pressure of the database is suddenly increased and the recommended service request is slow due to the fact that the instantaneous large flow directly flows to the database after the product recommendation service is deployed and restarted is avoided. And recording and counting a product recommendation strategy triggering the product recommendation service, providing data support for subsequently updating hot spot data, accelerating the response speed of the product recommendation data and improving the user experience.
Example 4:
the embodiment provides a method for generating a product recommendation strategy, as shown in fig. 4. The method for generating the product recommendation strategy comprises the following steps:
S11, creating a marketing strategy.
In this embodiment, the marketing strategy is the product recommendation strategy in embodiment 1. An operator or a company plans a product recommendation strategy through a full-quantity customer value operation platform (called a full-customer platform for short), and marks out a target user group marketing recommended product, namely, a plurality of recommended products are selected, a target user group is respectively marked out, a plurality of product recommendation strategies are generated, and product recommendation strategy information and recommended product information are stored in a database. Such as: recommended product 1:511206049563 [ X directed traffic membership package (forward) ], recommended product 2:51202209261701 (X video VIP member-moon card), the delineated target group is characterized by Guangzhou, 20-35 age, commonly used application X video, Y music and the like, and the number of users is 5000 ten thousand; 2 product recommendation policies a and B will be generated, one product recommendation policy corresponding to each recommended product (i.e., product recommendation policy a for marketing of recommended product 1 and product recommendation policy B for marketing of recommended product 2). One recommended product may be selected by different product recommendation strategies to market different target user groups, such as two recommended products: three product recommendation strategies D, E, F can be established for the recommended products 3 and 4, and the recommended products 3, 4 and 3 are selected respectively, and the difference between the product recommendation strategies D and F is that the delineated target user groups are different.
The all-passenger platform is based on user behavior characteristics, performs accurate modeling by adopting a big data algorithm, and provides accurate marketing product recommendation service for clients through strategy planning, data insight, real-time tag library and all-channel collaboration and through the full life cycle of the clients. The whole guest platform is in butt joint with four large types of contacts of on-line, off-line, outbound and short messages of the whole country, more than 600 total intensive contacts of headquarters and main operation contacts of provinces are covered, and the requirement on concurrency of a product recommendation service interface is quite large. And as the contact of the operator is accessed to the product recommendation interface of the all-passenger platform, the user access scale is rapidly increased, the daily average access amount reaches more than ten millions of levels, and meanwhile, part of contacts require that the interface access time delay cannot exceed 200ms.
S12, filtering by a user.
In this embodiment, the user features are identification information of the user, and the user features include, but are not limited to: mobile phone number, identification card number. Filtering users in the operator or company related system/database according to the target user group characteristics to obtain filtered user characteristics, for example, according to the above: the user is filtered by the related system/database of the operator by Guangzhou, age 20-35, common application X video, Y music and the like, the number of the user is assumed to be 5000 ten thousand, and the mobile phone number corresponding to the 5000 ten thousand is taken.
S13, storing strategy information.
In this embodiment, a mapping relationship between 5000 ten thousand mobile phone numbers and product recommendation policies a and B is constructed, and the mapping relationship between 5000 ten thousand mobile phone numbers and product recommendation policies a and B is stored in a Database, and an external contact is waited to initiate product recommendation service call, wherein the storage format is a record of one mobile phone number, for example, 100 ten thousand user recommendation policies a are selected, 100 ten thousand records are stored in the Database table, the records include a mobile phone number field and a policy ID field, and the actual storage of the mapping relationship can use technologies such as HBase (Hadoop Database) and the like, so that the time consumption of query is reduced.
According to the method for generating the product recommendation strategy, the user strategy relation and the strategy information are constructed for product recommendation through the cooperation of the full-quantity customer value operation platform (hereinafter referred to as the full-passenger platform) and the related system/database of the operator, and the accurate marketing product recommendation service is provided for the customers.
Example 5:
as shown in fig. 5, this embodiment provides a recommendation device for a product, including: the system comprises a first obtaining module 21, a second obtaining module 22 and a recommending module 23, wherein the first obtaining module 21 is used for responding to the product recommending service call and obtaining product recommending strategy information from a multi-level cache according to user characteristics, the first obtaining module 22 is connected with the first obtaining module 21 and used for obtaining the product recommending strategy information from a database if the product recommending strategy information is not recorded in the multi-level cache, and the recommending module 23 is connected with the first obtaining module 21 and the second obtaining module 23 and used for obtaining the recommending product information according to the product recommending strategy information and used for recommending the product, wherein the multi-level cache comprises but is not limited to: a first cache and a second cache.
Specifically, the first acquisition module 21 includes: the first query unit 211, the first acquisition unit 212 and the second acquisition unit 213, the first query unit 211 is configured to query a product recommendation policy list from the multi-level cache according to a user feature in response to a product recommendation service call, the first acquisition unit 212 is connected to the first query unit 211 and configured to respectively acquire product recommendation policy information corresponding to each policy in the product recommendation policy list from the first cache according to the product recommendation policy list, and the second acquisition unit 213 is connected to the first acquisition unit 212 and the first query unit 211 and is configured to acquire product recommendation policy information corresponding to each policy in the product recommendation policy list from the second cache if product recommendation policy information corresponding to each policy in the product recommendation policy list is not recorded in the first cache.
Optionally, the first acquisition module 21 further includes: the second query unit 214 is configured to query the product recommendation policy list from the database according to the user characteristics if the product recommendation policy list is not queried in the multi-level cache.
Optionally, the first acquisition module 21 further includes: the first updating unit 215 and the second updating unit 216, the first updating unit 215 is configured to perform cluster broadcasting through the message queue if the product recommendation policy information is obtained from the second cache or the database, write the product recommendation policy information into the first cache, and the second updating unit 216 is configured to write the product recommendation policy information into the second cache if the product recommendation policy information is obtained from the database.
Specifically, the second acquisition module 22 includes: the third obtaining unit 221 is configured to obtain, from the database, product recommendation policy information corresponding to each policy in the product recommendation policy list according to the product recommendation policy list.
Optionally, the product recommending device further includes: the recording module 24 is configured to record product recommendation policy identifiers in the product recommendation policy information, obtain a recommendation record, count the number of each product recommendation policy identifier in the recommendation record, and aggregate product recommendation policies corresponding to the product recommendation policy identifiers if the number of the product recommendation policy identifiers exceeds a preset value, so as to obtain a recent frequent query policy list.
Optionally, the product recommending device further includes: the system comprises a policy generation module 20a and a deployment module 20b, wherein the policy generation module 20a is used for creating a product recommendation policy according to recommended products, defining target user group characteristics according to the product recommendation policy, inquiring user characteristics according to the target user group characteristics, constructing a mapping relation between the product recommendation policy and the user characteristics, storing the mapping relation in a database, summarizing the product recommendation policy according to the same user characteristics to obtain a product recommendation policy list, and the deployment module 20b is used for inquiring recent hot spot data and writing the recent hot spot data into a multi-level cache, wherein the hot spot data comprises a recently newly built product recommendation policy and a recently frequently inquired policy list.
It can be appreciated that the above-provided recommending apparatus for a product is used for executing the method corresponding to the embodiment 1 provided above, and therefore, the advantages achieved by the recommending apparatus for a product can refer to the method of the embodiment 1 and the advantages of the corresponding scheme in the following detailed description, which are not repeated herein.
Example 6:
the present embodiment also provides an electronic device comprising a memory in which a computer program is stored and a processor arranged to run the computer program to implement the method of recommending a product of the above embodiments.
Example 7:
the present embodiment also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the recommendation method for products in the above embodiments.
It is to be understood that the above embodiments are merely illustrative of the application of the principles of the present invention, but not in limitation thereof. Various modifications and improvements may be made by those skilled in the art without departing from the spirit and substance of the invention, and are also considered to be within the scope of the invention.
Claims (10)
1. A method of recommending a product, comprising:
responding to the product recommendation service call, and acquiring product recommendation strategy information from the multi-level cache according to the user characteristics;
if the product recommendation strategy information is not recorded in the multi-level cache, acquiring the product recommendation strategy information from a database;
acquiring recommended product information according to the product recommendation strategy information for product recommendation, wherein the multi-level cache comprises but is not limited to: a first cache and a second cache.
2. The method for recommending products according to claim 1, wherein the obtaining product recommendation policy information from the multi-level cache according to the user characteristics in response to the product recommendation service call specifically comprises:
responding to the product recommendation service call, and inquiring a product recommendation strategy list from the multi-level cache according to the user characteristics;
respectively acquiring product recommendation policy information corresponding to each policy in the product recommendation policy list from the first cache according to the product recommendation policy list;
if the product recommendation policy information corresponding to each policy in the product recommendation policy list is not recorded in the first cache, the product recommendation policy information corresponding to each policy in the product recommendation policy list is obtained from the second cache.
3. The method of claim 2, further comprising, after said querying the product recommendation policy list from the multi-level cache based on the user characteristics in response to the product recommendation service call:
if the product recommendation policy list is not queried in the multi-level cache, querying the product recommendation policy list from the database according to the user characteristics.
4. The method for recommending products according to claim 2, wherein after the product recommendation policy information corresponding to each policy in the product recommendation policy list is obtained from the second cache if the product recommendation policy information corresponding to each policy in the product recommendation policy list is not recorded in the first cache, further comprising:
if the product recommendation strategy information is obtained from the second cache, cluster broadcasting is carried out through the message queue, and the product recommendation strategy information is written into the first cache;
if the product recommendation strategy information is obtained from the database, the product recommendation strategy information is written into the second cache, cluster broadcasting is carried out through the message queue, and the product recommendation strategy information is written into the first cache.
5. The method for recommending products according to claim 2, wherein if no product recommendation policy information is recorded in the multi-level cache, obtaining the product recommendation policy information from the database, specifically comprises:
And respectively acquiring product recommendation strategy information corresponding to each strategy in the product recommendation strategy list from the database according to the product recommendation strategy list.
6. The method of recommending products according to claim 1, further comprising, after said acquiring recommended product information for product recommendation according to the product recommendation policy information:
recording a product recommendation strategy identification in the product recommendation strategy information to obtain a recommendation record;
counting the number of recommendation strategy identifiers of each product in the recommendation record;
if the number of the product recommendation strategy identifiers exceeds the preset value, summarizing the product recommendation strategies corresponding to the product recommendation strategy identifiers to obtain a recent frequent query strategy list.
7. The method of claim 1, further comprising, prior to said retrieving recommended product information from the multi-level cache based on the user characteristics in response to the product recommendation service call:
creating a product recommendation strategy according to the recommended product;
defining target user group characteristics according to a product recommendation strategy;
inquiring user characteristics according to the target user group characteristics;
building a mapping relation between a product recommendation strategy and user characteristics, and storing the mapping relation into a database;
Summarizing product recommendation strategies according to the same user characteristics to obtain a product recommendation strategy list;
and inquiring recent hot spot data, and writing the hot spot data into the multi-level cache, wherein the hot spot data comprises a recent new product recommendation strategy and a recent frequent inquiry strategy list.
8. A recommendation device for a product, comprising: a first acquisition module, a second acquisition module and a recommendation module,
a first acquisition module, configured to acquire product recommendation policy information from the multi-level cache according to user characteristics in response to a product recommendation service call,
the first acquisition module is connected with the first acquisition module and is used for acquiring the product recommendation strategy information from the database if the product recommendation strategy information is not recorded in the multi-level cache,
the recommending module is connected with the first acquiring module and the second acquiring module and is used for acquiring recommended product information according to the product recommending strategy information so as to be used for product recommendation, wherein the multi-level cache comprises but is not limited to: a first cache and a second cache.
9. An electronic device comprising a memory and a processor, the memory having stored therein a computer program, the processor being arranged to run the computer program to implement the recommendation method for a product according to any of claims 1-7.
10. A computer readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, implements a recommendation method for a product according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311227793.6A CN117290392A (en) | 2023-09-21 | 2023-09-21 | Product recommendation method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311227793.6A CN117290392A (en) | 2023-09-21 | 2023-09-21 | Product recommendation method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117290392A true CN117290392A (en) | 2023-12-26 |
Family
ID=89251174
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311227793.6A Pending CN117290392A (en) | 2023-09-21 | 2023-09-21 | Product recommendation method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117290392A (en) |
-
2023
- 2023-09-21 CN CN202311227793.6A patent/CN117290392A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102638584B (en) | Data distributing and caching method and data distributing and caching system | |
| CN111258978B (en) | Data storage method | |
| CN105138571B (en) | Distributed file system and method for storing massive small files | |
| CN105159845A (en) | Memory reading method | |
| US10579597B1 (en) | Data-tiering service with multiple cold tier quality of service levels | |
| US11151081B1 (en) | Data tiering service with cold tier indexing | |
| US11775480B2 (en) | Method and system for deleting obsolete files from a file system | |
| CN112685499B (en) | Method, device and equipment for synchronizing flow data of working service flow | |
| CN107888666A (en) | A kind of cross-region data-storage system and method for data synchronization and device | |
| CN111274004B (en) | Process instance management method and device and computer storage medium | |
| US10951465B1 (en) | Distributed file system analytics | |
| CN101778131A (en) | Data synchronization system | |
| CN109246102B (en) | System and method for supporting large-scale authentication data rapid storage and retrieval | |
| CN102724301B (en) | Cloud database system and method and equipment for reading and writing cloud data | |
| US11210212B2 (en) | Conflict resolution and garbage collection in distributed databases | |
| EP3649532B1 (en) | Methods, systems, databases and network nodes of data communication networks for handling data posts | |
| CN101789963A (en) | Data synchronization system | |
| CN107590199B (en) | Memory-oriented multithreading database design method | |
| CN112800066A (en) | Index management method, related device and storage medium | |
| CN112148745A (en) | Multi-HBase cluster access method, device and storage medium | |
| CN117290392A (en) | Product recommendation method and device, electronic equipment and storage medium | |
| US11886439B1 (en) | Asynchronous change data capture for direct external transmission | |
| CN112214460A (en) | High-performance storage control method based on distributed large-capacity fragmentation | |
| CN117648297B (en) | Method, system, equipment and medium for offline merging of small files based on object storage | |
| CN114168560A (en) | Data management method and device |
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 |