CN113127741B - Cache method for reading and writing data of mass users and posts in part-time post recommendation system - Google Patents
Cache method for reading and writing data of mass users and posts in part-time post recommendation system Download PDFInfo
- Publication number
- CN113127741B CN113127741B CN202110473400.4A CN202110473400A CN113127741B CN 113127741 B CN113127741 B CN 113127741B CN 202110473400 A CN202110473400 A CN 202110473400A CN 113127741 B CN113127741 B CN 113127741B
- Authority
- CN
- China
- Prior art keywords
- data
- post
- time
- real
- characteristic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
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
- G06F16/953—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
- 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
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/283—Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP
-
- 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
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/445—Program loading or initiating
- G06F9/44521—Dynamic linking or loading; Link editing at or after load time, e.g. Java class loading
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/105—Human resources
- G06Q10/1053—Employment or hiring
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Software Systems (AREA)
- Human Resources & Organizations (AREA)
- Entrepreneurship & Innovation (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Computational Linguistics (AREA)
- Information Transfer Between Computers (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Memory System Of A Hierarchy Structure (AREA)
Abstract
The invention discloses a cache method for reading and writing mass users and post data in a part-time post recommendation system, which has the characteristics of multiple types of data in a link of the recommendation system, reasonably divides the storage mode of the data in a memory, loads the data as required, and respectively stores the data in a JVM (Java virtual machine) in-pile memory, a JVM (Java virtual machine) out-pile memory and a large-scale distributed cache according to the service characteristics and the data magnitude of the data in the recommendation system. When the recommendation system reads data, the data of each link is stored in the most reasonable position according to the data characteristics and the data magnitude, on one hand, the cache storage of mass data is guaranteed, on the other hand, the low delay and high throughput of the recommendation system are guaranteed, and the real-time performance of the data of each link is also guaranteed to a great extent, so that the quick and accurate recommendation service is provided for users, and the user experience is improved.
Description
Technical Field
The invention relates to the field of data reading and processing application, in particular to a cache method for reading and writing data of mass users and posts in a part-time post recommendation system.
Background
Currently, the recruitment industry is undergoing rapid development and transformation, on one hand, with the surge of internet and the creation of the whole people, each industry faces industrial upgrading, the competition of enterprises focuses on talents competition, and talent recruitment becomes the first major affairs of each enterprise. On the other hand, with the popularization of mobile internet and the maturity of big data technology, online recruitment becomes an efficient and convenient recruitment mode. But one problem is encountered in online recruitment: information overload, that is, a user who recruits online has to face a huge amount of recruiting posts/job resumes, and how to pick out relevant information useful for the user from the huge amount of information is a problem.
Disclosure of Invention
The invention aims to provide a cache method for reading and writing data of mass users and posts in a part-time post recommendation system. The invention can increase the real-time performance of data reading and writing, reduce the risk of data loss and query timeout in large data scenes, increase the high concurrency capability of the system and ensure the low delay and high throughput of the recommendation system.
In order to solve the technical problems, the technical scheme provided by the invention is as follows: the method for caching the reading and writing of the data of the mass users and the posts in the part-time post recommendation system comprises the recommendation system, wherein the recommendation system acquires the data of a data center station through memory cache management, and the data center station is provided with an offline computing platform, a real-time computing platform, an algorithm platform and a data warehouse; the memory cache management is provided with a distributed cache, a JVM out-of-heap memory and a JVM in-heap memory; the off-line computing platform computes data in the data warehouse, writes the computing result into a distributed cache to form user characteristic off-line data, post characteristic off-line data and post recall off-line data, and loads the user characteristic off-line data, the post characteristic off-line data and the post recall off-line data into a JVM (JVM) off-heap memory for cache management;
the real-time computing platform computes data in the data warehouse, and writes the computation result into the distributed cache to form hot spot data, real-time post recall data and real-time user and post characteristic data, wherein the hot spot data is loaded in a memory in a JVM (Java virtual machine) stack for cache management;
the algorithm platform clusters the users at regular time, and writes topN posts selected by each type of users into the distributed cache forming model recall post data;
writing basic configuration data, bottom-of-pocket data and rough rule data in a data warehouse into a distributed cache, and loading the data into a memory in a JVM heap for cache management;
when the recommendation system reads data, the memory cache region corresponding to the data is quickly positioned according to the requirement, and the part-time post recommendation result is immediately made.
In the above caching method for reading and writing the mass users and the post data in the part-time post recommendation system, the formation of the user characteristic offline data and the post characteristic offline data is formed by performing user characteristic and post characteristic offline index calculation on data in a data warehouse through an offline calculation platform and writing a calculation result into a distributed cache;
the post recall offline data is formed by the steps that the offline computing platform performs classification computation on posts according to geographic positions, environments and popular degree labels, and then writes computation results into a distributed cache.
According to the cache method for reading and writing the data of the mass users and the posts in the part-time post recommendation system, the off-line indexes of the user characteristics comprise user click post collaboration in historical data, and most common access addresses, ages, sexes and access ports in user history; the off-line indexes of the post features comprise exposure numbers, name reporting numbers, clicking numbers, CTR, CVR, post longitude and latitude and post states in historical data.
According to the cache method for reading and writing the mass users and the post data in the part-time post recommendation system, when the recommendation system reads the user characteristic offline data, the post characteristic offline data and the post recall offline data, the recommendation system loads the user characteristic offline data, the post characteristic offline data and the post recall offline data into the JVM off-pile memory, and when the user characteristic offline data, the post characteristic offline data and the post recall offline data cannot be loaded, the recommendation system reads and updates the JVM off-pile memory from the distributed cache.
According to the cache method for reading and writing the data of the mass users and the posts in the part-time post recommendation system, the hot spot data are formed by calculating the access frequency of the posts in the data warehouse in real time through the real-time calculation platform, and counting the posts of top100 as the hot spot data;
the real-time user and post characteristic data is formed by monitoring the change of a binlog in a data warehouse through a real-time computing platform to compute the user characteristic and post characteristic real-time indexes of the current day and then writing the computed result into a distributed cache;
the formation of the real-time post recall data is formed by monitoring the change of a post table binlog in a data warehouse through a real-time computing platform and storing post data newly sent in the current day into a distributed cache in real time.
According to the cache method for reading and writing the data of the mass users and the posts in the part-time post recommendation system, the user characteristic real-time indexes comprise user click post coordination in data of the day, and most common access addresses, ages, sexes and access ports in user history; the post feature real-time index comprises the exposure number, the name reporting number, the click number, the CTR, the CVR, the longitude and latitude of the post and the state of the post in the data of the current day.
According to the caching method for reading and writing the data of the mass users and the posts in the part-time post recommendation system, the clustering model is arranged in the algorithm platform, and training is performed according to the user characteristics and the post characteristic data through the clustering model at regular time, so that the users are clustered.
According to the cache method for reading and writing the data of the mass users and the posts in the part-time post recommendation system, when the recommendation system reads the data, the execution flow comprises recalling, obtaining the characteristics of the users and the posts, rough typesetting, model fine typesetting and bottom pocketing;
the recalling step can respectively read post recalling offline data and real-time post recalling data from a JVM out-of-pile memory and a distributed cache according to basic configuration data in the JVM in-pile memory;
the step of acquiring the user and post characteristics respectively reads user characteristic offline data, post characteristic offline data and real-time user and post characteristic data from a JVM (Java virtual machine) off-heap memory and a distributed cache;
the coarse-arranging step reads coarse-arranging rule data from the memory in the JVM heap;
the fine arranging step is carried out by utilizing an algorithm platform;
and the bottom-pocketing step reads bottom-pocketing data from the memory in the JVM heap.
Compared with the prior art, the method and the device have the advantages that the characteristic that various types of data exist in the link of the recommendation system is utilized, the storage mode of the data in the memory is reasonably divided, the data are loaded according to needs, and the data are respectively stored in the JVM in-pile memory, the JVM out-pile memory and the large-scale distributed cache according to the service characteristics and the data magnitude of the data in the recommendation system. The data magnitude of post basic configuration data, bottom-of-pocket data and rough-arrangement rule data is small, and frequently-accessed data are loaded from a data warehouse and then are placed in a memory in a JVM (virtual machine monitor) heap for cache management; the magnitude of data of user characteristic offline data, post characteristic offline data and post recall offline data is relatively large, data which are frequently accessed but not frequently updated are written into a distributed cache by an offline computing platform, and then a JVM (Java virtual machine) off-heap memory is loaded and managed from the distributed cache; writing data which need multi-platform intervention and have strong real-time performance, such as real-time post recall data, real-time user characteristics and post characteristic data and model recall data into a distributed cache for management by a real-time computing platform and an algorithm platform; the hot data can be written into the distributed cache by the real-time computing platform due to the characteristics of strong data real-time performance and access frequency, then the internal memory in the JVM heap is loaded and managed from the distributed cache, so that when the recommendation system reads the data, the internal memory area where the data is located can be quickly positioned to read, and then a part-time position recommendation result is immediately made after a recommendation link is completed.
Drawings
FIG. 1 is a schematic of the present invention.
Detailed Description
The present invention will be further described with reference to the following examples and drawings, but the present invention is not limited thereto.
Example (b): a cache method for reading and writing of massive users and post data in a part-time post recommendation system, as shown in fig. 1, comprises a recommendation system, wherein the recommendation system acquires data of a data center station through memory cache management, the data center station is provided with an offline computing platform (by using a technology stack (mainly Hadoop) of big data, all input data are prepared before computing is started, the input data cannot change, and a computing mode of a computing result is obtained immediately after a problem is solved), a real-time computing platform (a real-time data computing mode based on a cloud server), an algorithm platform and a data warehouse; the memory cache management is provided with a distributed cache, a JVM out-of-pile memory and a JVM in-pile memory;
the off-line computing platform computes data in the data warehouse, writes the computed result into the distributed cache to form user characteristic off-line data, post characteristic off-line data and post recall off-line data, and loads the user characteristic off-line data, the post characteristic off-line data and the post recall off-line data into a JVM off-heap memory for cache management;
the user characteristic offline data and the post characteristic offline data are formed by performing user characteristic and post characteristic offline index calculation on data in a data warehouse through an offline calculation platform and writing calculation results into a distributed cache; the off-line indexes of the user characteristics comprise user clicking position cooperation in historical data, the most common access address, age, gender and access port of user history; the off-line indexes of the post features comprise exposure number, name reporting number, click number, CTR (click number/exposure number x 100%), CVR (name reporting number/click number x 100%), post longitude and latitude and post state in historical data.
The post recall offline data is formed by the steps that the offline computing platform performs classification computation on posts according to geographic positions, environments and popular degree labels, and then writes computation results into a distributed cache.
When reading user characteristic offline data, post characteristic offline data and post recall offline data, the recommendation system loads the data in the JVM out-of-pile memory first, and then reads and updates the JVM out-of-pile memory in the distributed cache when the data cannot be loaded.
The real-time computing platform computes data in the data warehouse, writes the computation result into distributed cache to form hot spot data, real-time post recall data and real-time user and post characteristic data, and loads the hot spot data into a memory in a JVM (JVM) stack for cache management;
the hot spot data is formed by calculating the access frequency of the posts in the data warehouse in real time through a real-time calculation platform, and counting the posts of top100 as the hot spot data; because the reading frequency of the hot spot data is very high, the hot spot data can be divided into two layers of JVM in-pile memory and distributed cache. The loading logic of the system is to load the memory in the JVM heap first, and when the memory is not loaded, the memory in the distributed cache is read and updated. The corresponding expiration time is set when the cache update is made.
The real-time user and post characteristic data is formed by monitoring the change of binlog in a data warehouse through a real-time computing platform to compute the real-time indexes of the user characteristics and the post characteristics of the current day and then writing the computed result into a distributed cache; the formation of the real-time post recall data is formed by monitoring the change of a post table binlog in a data warehouse through a real-time computing platform and storing post data newly sent in the current day into a distributed cache in real time. The real-time performance of the data is high, and the data can be directly updated and read in the distributed cache.
The user characteristic real-time indexes comprise user click post coordination, the most common access address of user history, age, gender and access port in the data of the current day; the post characteristic real-time indexes comprise exposure numbers, name reporting numbers, click numbers, CTR, CVR, post longitude and latitude and post states in data of the day.
The algorithm platform regularly clusters the users, and writes topN posts selected by each type of users into the distributed cache to form model recall post data; and a clustering model is arranged in the algorithm platform, and training is performed according to the user characteristics and the post characteristic data through the clustering model at regular time, so that the users are clustered.
The offline computing platform and the real-time computing platform compute and synchronize basic configuration data, bottom-of-pocket data and rough rule data in a data warehouse into a distributed cache according to agreed service rules, and the basic configuration data, the bottom-of-pocket data and the rough rule data are loaded into a memory in a JVM (JVM) heap for cache management when the distributed cache is used; the data are mainly rule configuration data which can be directly obtained according to a data warehouse, and the bottommost data source is a relational database such as MYSQL, PostgreSQL and the like of the data warehouse; and the loading sequence is JVM in-heap memory-distributed cache-data warehouse, and if the load in the JVM in-heap memory is not enough, the distributed cache is loaded and the JVM in-heap memory is updated. And if the distributed cache is not loaded in the data warehouse, the distributed cache is loaded in the data warehouse and updated. The corresponding expiration time is set when the cache update is made.
When the recommendation system reads data, the memory cache region corresponding to the data is quickly positioned according to the requirement, and the part-time post recommendation result is immediately made. As shown in fig. 1, when an APP client requests a recommendation system to output a recommendation result, an execution process of the recommendation system includes recall, user and post feature acquisition, rough ranking, model refinement and bottoming when reading data;
the recalling step can respectively read post recalling offline data and real-time post recalling data from a JVM out-of-pile memory and a distributed cache according to basic configuration data in the JVM in-pile memory;
the step of acquiring the user and post characteristics respectively reads user characteristic offline data, post characteristic offline data and real-time user and post characteristic data from the JVM off-heap memory and the distributed cache;
the step of coarse-arranging reads coarse-arranging rule data from a memory in a JVM heap;
the fine-arranging step is carried out by using an algorithm platform, namely a clustering model;
and the bottom-pocketing step reads bottom-pocketing data from the memory in the JVM heap.
In summary, when the recommendation system reads data, the data of each link is stored at the most reasonable position according to the data characteristics and the data magnitude, so that on one hand, the cache storage of mass data is ensured, on the other hand, the low delay and high throughput of the recommendation system are ensured, and the real-time performance of the data of each link is also ensured to a great extent, thereby providing a quick and accurate recommendation service for a user and improving the user experience.
Claims (7)
1. The method for caching the reading and writing of the data of the mass users and the posts in the part-time post recommendation system comprises the recommendation system, wherein the recommendation system acquires the data of a data center station through memory cache management, and the data center station is provided with an offline computing platform, a real-time computing platform, an algorithm platform and a data warehouse; the memory cache management is provided with a distributed cache, a JVM out-of-pile memory and a JVM in-pile memory, and is characterized in that: the off-line computing platform computes data in the data warehouse, writes the computing result into a distributed cache to form user characteristic off-line data, post characteristic off-line data and post recall off-line data, and loads the user characteristic off-line data, the post characteristic off-line data and the post recall off-line data into a JVM (JVM) off-heap memory for cache management;
the real-time computing platform computes data in the data warehouse, and writes the computation result into the distributed cache to form hot spot data, real-time post recall data and real-time user and post characteristic data, wherein the hot spot data is loaded in a memory in a JVM (Java virtual machine) stack for cache management;
the algorithm platform clusters the users at regular time, and writes topN posts selected by each type of users into the distributed cache forming model recall post data;
writing basic configuration data, bottom-of-pocket data and rough rule data in a data warehouse into a distributed cache, and loading the data into a memory in a JVM heap for cache management;
when the recommendation system reads data, quickly positioning the data to a memory cache region corresponding to the data according to requirements, and instantly making a part-time post recommendation result;
the recommendation system executes a flow including recalling, obtaining user and post characteristics, rough typesetting, model fine typesetting and bottom pocketing when reading data;
the recalling step can respectively read post recalling offline data and real-time post recalling data from the JVM out-of-pile memory and the distributed cache according to basic configuration data in the JVM in-pile memory;
the step of acquiring the user and post characteristics respectively reads user characteristic offline data, post characteristic offline data and real-time user and post characteristic data from the JVM off-heap memory and the distributed cache;
the step of coarse-arranging reads coarse-arranging rule data from a memory in a JVM heap;
the fine arranging step is carried out by utilizing an algorithm platform;
and the bottom-pocketing step reads bottom-pocketing data from the memory in the JVM heap.
2. The method for caching reading and writing of mass users and post data in the part-time post recommendation system according to claim 1, wherein: the user characteristic offline data and the post characteristic offline data are formed by performing user characteristic and post characteristic offline index calculation on data in a data warehouse through an offline calculation platform and writing calculation results into a distributed cache;
the post recall offline data is formed by the steps that the offline computing platform performs classification computation on posts according to geographic positions, environments and popular degree labels, and then writes computation results into a distributed cache.
3. The method for caching reading and writing of mass users and post data in the part-time post recommendation system according to claim 2, wherein: the off-line indexes of the user characteristics comprise user click post collaboration in historical data, the most common access address, age, gender and access port of user history; the off-line indexes of the post characteristics comprise exposure numbers, name reporting numbers, click numbers, CTR, CVR, post longitude and latitude and post states in historical data.
4. The method for caching reading and writing of mass users and post data in the part-time post recommendation system according to claim 1, wherein: when reading the user characteristic offline data, the post characteristic offline data and the post recall offline data, the recommendation system loads the data into the JVM out-of-pile memory first, and when the data cannot be loaded, reads and updates the JVM out-of-pile memory in the distributed cache.
5. The method for caching reading and writing of mass users and post data in the part-time post recommendation system according to claim 1, wherein: the hot data is formed by calculating the access frequency of the posts in the data warehouse in real time through a real-time calculation platform, and counting the posts of top100 as the hot data;
the real-time user and post characteristic data is formed by monitoring the change of a binlog in a data warehouse through a real-time computing platform to compute the user characteristic and post characteristic real-time indexes of the current day and then writing the computed result into a distributed cache;
the formation of the real-time post recall data is formed by monitoring the change of a post table binlog in a data warehouse through a real-time computing platform and storing post data newly sent in the current day into a distributed cache in real time.
6. The method for caching reading and writing of mass users and post data in the part-time post recommendation system according to claim 1, wherein: the user characteristic real-time indexes comprise user clicking position cooperation, a most frequently used access address, age, gender and an access port of user history in the data of the current day; the post feature real-time index comprises the exposure number, the name reporting number, the click number, the CTR, the CVR, the longitude and latitude of the post and the state of the post in the data of the current day.
7. The method for caching reading and writing of mass users and post data in the part-time post recommendation system according to claim 1, wherein: and a clustering model is arranged in the algorithm platform, and training is performed according to the user characteristics and the post characteristic data through the clustering model at regular time, so that the users are clustered.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110473400.4A CN113127741B (en) | 2021-04-29 | 2021-04-29 | Cache method for reading and writing data of mass users and posts in part-time post recommendation system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110473400.4A CN113127741B (en) | 2021-04-29 | 2021-04-29 | Cache method for reading and writing data of mass users and posts in part-time post recommendation system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113127741A CN113127741A (en) | 2021-07-16 |
| CN113127741B true CN113127741B (en) | 2022-08-23 |
Family
ID=76780669
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110473400.4A Active CN113127741B (en) | 2021-04-29 | 2021-04-29 | Cache method for reading and writing data of mass users and posts in part-time post recommendation system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113127741B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113486245B (en) * | 2021-07-26 | 2023-07-07 | 杭州弧途科技有限公司 | Part-time job recommendation engine recall method combining big data offline calculation and real-time monitoring |
| CN113487295B (en) * | 2021-07-26 | 2024-01-19 | 杭州弧途科技有限公司 | Off-line post recall method of part-time post recommendation system |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016189317A1 (en) * | 2015-05-26 | 2016-12-01 | Jagex Limited | An online game having a computerized recommender system |
| CN105488662B (en) * | 2016-01-07 | 2021-09-03 | 北京华品博睿网络技术有限公司 | Online recruitment system based on bidirectional recommendation |
| CN107657427A (en) * | 2017-09-26 | 2018-02-02 | 成都兼职地带信息技术有限公司 | One kind is based on the part-time post commending system of data acquisition |
| CN108182213A (en) * | 2017-12-20 | 2018-06-19 | 福建新大陆软件工程有限公司 | A kind of data processing optimization device and method based on distributed system |
| CN109558429B (en) * | 2018-11-16 | 2021-01-01 | 广东百城人才网络股份有限公司 | Talent service bidirectional recommendation method and system based on Internet big data |
| CN111651374A (en) * | 2020-04-14 | 2020-09-11 | 北京齐尔布莱特科技有限公司 | Data processing method and device, computing equipment and readable storage medium |
-
2021
- 2021-04-29 CN CN202110473400.4A patent/CN113127741B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113127741A (en) | 2021-07-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108600321A (en) | A kind of diagram data storage method and system based on distributed memory cloud | |
| CN104933112B (en) | Distributed interconnection Transaction Information storage processing method | |
| CN109656958B (en) | Data query method and system | |
| CN113127741B (en) | Cache method for reading and writing data of mass users and posts in part-time post recommendation system | |
| CN106649687B (en) | Big data online analysis processing method and device | |
| CN102779138B (en) | The hard disk access method of real time data | |
| Schaffner et al. | A hybrid row-column OLTP database architecture for operational reporting | |
| CN103631922A (en) | Hadoop cluster-based large-scale Web information extraction method and system | |
| CN110837520A (en) | Data processing method, platform and system | |
| Gupta et al. | Faster as well as early measurements from big data predictive analytics model | |
| CN111552885A (en) | System and method for realizing automatic real-time message pushing operation | |
| CN113157943A (en) | Distributed storage and visual query processing method for large-scale financial knowledge map | |
| CN103365987A (en) | Clustered database system and data processing method based on shared-disk framework | |
| CN116719822B (en) | Method and system for storing massive structured data | |
| CN106776810B (en) | Big data processing system and method | |
| CN115599871A (en) | Lake and bin integrated data processing system and method | |
| CN116126973A (en) | Meteorological lattice data management method and device | |
| CN102597969A (en) | Database management device using key-value store with attributes, and key-value-store structure caching-device therefor | |
| CN114138735A (en) | Method for quickly loading Janus graph data in batches | |
| Qi et al. | The consistency analysis of secondary index on distributed ordered tables | |
| CN111399838A (en) | Data modeling method and device based on spark SQ L and materialized view | |
| CN110569310A (en) | Management method of relational big data in cloud computing environment | |
| CN111813800A (en) | Streaming data real-time approximate calculation method based on deep reinforcement learning | |
| CN110781210A (en) | Data processing platform for multi-dimensional aggregation real-time query of large-scale data | |
| CN114416738B (en) | Data aggregation method and device based on relational database |
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 |