CN110110180A - A kind of job hunter's recruitment information searching method based on collaborative filtering - Google Patents
A kind of job hunter's recruitment information searching method based on collaborative filtering Download PDFInfo
- Publication number
- CN110110180A CN110110180A CN201910336862.4A CN201910336862A CN110110180A CN 110110180 A CN110110180 A CN 110110180A CN 201910336862 A CN201910336862 A CN 201910336862A CN 110110180 A CN110110180 A CN 110110180A
- Authority
- CN
- China
- Prior art keywords
- user
- post
- information
- collaborative filtering
- job hunter
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- 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
- G06F16/9538—Presentation of query results
-
- 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
Abstract
The invention discloses a kind of job hunter's recruitment information searching method based on collaborative filtering includes the following steps: the post information that existing major recruitment website is crawled by WebMagic web crawlers technology, and storage is in the database；It builds the mobile terminal APP and plain engine platform is searched at computer PC network end, querying condition is set in platform；After user's opening end APP or PC enters platform, carries out registration login and select querying condition；The system background for searching plain platform returns to suitable post information according to querying condition, and is ranked up displaying；User jumps to corresponding recruitment website by click sorted lists and checks post information；The log information that system acquisition is clicked to user, cleans data；Data after cleaning are calculated using the collaborative filtering based on job hunter, generate real-time recommendation list.The present invention provides for user facilitates accurate position search experience service, is quickly found out suitable position convenient for job hunter.
Description
Technical field
The present invention relates to internet search engine technical field, especially a kind of job hunter based on collaborative filtering is recruited
Engage information search method.
Background technique
Internet recruitment website is more and more at present, emerges one after another, also thoroughly change people job hunting mode and
The recruitment mode of enterprise, it is existing in the market just to have: 51job, to hunt and engage the recruitment websites such as net, drag hook net and intelligence connection recruitment.But
User volume on these websites is quite huge but not intercommunication, job hunter are often difficult to find that the post for being suitble to oneself.In addition to this,
Existing recruitment website generally has recruitment information recommendation function, and searching results are often based on the position search condition of job hunter
Screening, the post that this way of recommendation is recommended may not mutually agree with job hunter.Therefore the duty of existing major recruitment website is solved
The problem of position information is not connected, and the search result of return and job hunter do not agree with is to solve the problems, such as that college students'employment is compeled in eyebrow
Eyelash.
Summary of the invention
Technical problem to be solved by the present invention lies in provide a kind of job hunter's recruitment information based on collaborative filtering
Searching method, provides for user and facilitates accurate position search experience service, is quickly found out suitable position convenient for job hunter.
In order to solve the above technical problems, the present invention provides a kind of job hunter's recruitment information search based on collaborative filtering
Method includes the following steps:
(1) the post information that existing major recruitment website is crawled by WebMagic web crawlers technology, is stored in data
In library；
(2) it builds the mobile terminal APP and plain engine platform is searched at computer PC network end, querying condition is set in platform；
(3) it after user's opening end APP or PC enters platform, carries out registration login and selects querying condition；
(4) system background for searching plain platform returns to suitable post information according to querying condition, and is ranked up displaying；
(5) user jumps to corresponding recruitment website by click sorted lists and checks post information；
(6) log information that system acquisition is clicked to user, cleans data；
(7) data after cleaning are calculated using the collaborative filtering based on job hunter, generate real-time recommendation list.
Preferably, in step (7), the data after cleaning are calculated using the collaborative filtering based on job hunter, raw
Specifically comprise the following steps: at real-time recommendation list
(71) code is write using Javascript in system front end bury a little, and user behavior information is stored in
In .log file on nginx server；
(72) collected user journal information is stored in HDFS using flume frame；
(73) data cleansing is carried out using HiveQL to place the data in Hbase database；
(74) building for edge calculations big data platform is carried out using newest spark PC cluster engine, after cleaning
Data using based on job hunter collaborative filtering calculate, be as a result stored in the warehouse Hive, and by Sqoop importing
MySQL database；
(75) it according to calculated result, generates recommendation list and returns to user.
Preferably, in step (74), collaborative filtering specifically comprises the following steps:
(741) it is as shown in Equation 1 to calculate the similarity matrix based on user
Wherein: SuvFor the similarity score of user u and user v, N (u) indicates that user u had the post information of browsing behavior
Set；N (v) indicates that user v had the set of the post information of browsing behavior；The coincidence journey of molecular moiety expression post information
Degree, it is clear that coincidence degree is higher, and post information is more similar；A normalization has been done in denominator part, reduces and operates excessive use
The similarity degree at family and other users；
(742) it is clicked according to the behavior of user and completes post recommendation
Wherein: PuiIt is user u to the recommendation score of post i.SuvFor the similarity score of user u and user v, k is indicated
User v is the preceding k similar users of user u, and the post user u for guaranteeing that user v was browsed was not browsed；rviIt indicates to use
Family v is to the behavior score of post i, for different behaviors (such as: browsing launches resume, participates in interview, enters official rank) to user behavior
Score define it is different.
Preferably, in step (741), the similarity matrix based on user is rewritten are as follows:
Wherein: SuvFor the similarity score of user u and user v, N (u) indicates that user u had the post information of browsing behavior
Set；N (v) indicates that user v had the set of the post information of browsing behavior；U (i) indicates the use for having behavior to post i
Family set, if a post was browsed by many job hunters, it will become lower in the contribution of registration.
Preferably, it in step (742), is clicked according to the behavior of user and completes post recommendation rewriting are as follows:
Wherein: SuvFor the similarity score of user u and user v, N (u) indicates that user u had the post information of browsing behavior
Set；N (v) indicates that user v had the set of the post information of browsing behavior；Each post information contribution degree being overlapped obtains
Divide and be all different, by function f (Δ ti) determine, function is defined as follows:
Wherein: | tui-tvi| user is indicated to the difference of the operating time in the same post, and the time is separated by shorter function
Value is higher, and the weight in registration contribution of response is higher.
The invention has the benefit that the present invention in a manner of search engine, incorporates existing by web crawlers technology
The data of major recruitment website provide convenience for job hunter's job hunting；In conjunction with big data technology, the collaborative filtering based on job hunter is calculated
Method etc. realizes position intelligent recommendation, substantially increases user and hunts for a job efficiency, allows users to more efficiently find and be suitble to oneself
Work.
Detailed description of the invention
Fig. 1 is method flow schematic diagram of the invention.
Fig. 2 is crawler flow diagram of the invention.
Fig. 3 is system architecture schematic diagram of the invention.
Specific embodiment
As shown in figure 3, a kind of job hunter's recruitment information search engine based on collaborative filtering of the invention, including number
According to three parts such as acquisition module 1, data analysis module 2, data display modules 3.
As depicted in figs. 1 and 2, a kind of job hunter's recruitment information searching method based on collaborative filtering, including it is as follows
Step:
Step 1 is crawled the post information of existing major recruitment website by WebMagic web crawlers technology, is stored in
In database.
Step 2, builds the mobile terminal APP and plain engine platform is searched at computer PC network end, the setting inquiry item in platform
Part.
Step 3 carries out registration login and selects querying condition after user's opening end APP or PC enters platform.
Step 4, the system background for searching plain platform returns to suitable post information according to querying condition, and is ranked up exhibition
Show.
Step 5, user jump to corresponding recruitment website by click sorted lists and check post information.
Step 6, the log information that system acquisition to user is clicked, cleans data.
Step 7, the data after cleaning are calculated using the collaborative filtering based on job hunter, generate real-time recommendation
List.
We generate recommendation list using calculating based on collaborative filtering in step 7 are as follows: find and target
There is user the user of similar interests to collect to merge and will find user in this user set and browse, but target user not by
The post information recommended generates offline recommendation list.Since the collaborative filtering (usercf) based on user needs to handle greatly
The data of amount, we calculate by Hadoop frame.
Hadoop frame is a distributed system infrastructure developed by apache foundation.With three big groups
Part mapreduce distributed arithmetic frame yarn task schedule platform hdfs distributed file system.We are just in this project
The log information of user is stored in HDFS.Hive is a Tool for Data Warehouse based on Hadoop, can be by structuring
Data file be mapped as a database table, and provide simple sql query function, sql sentence can be converted to
MapReduce task is run.
It is specifically included in the step seven:
S7.1 writes code using Javascript in system front end and bury a little, and user behavior information is stored in
In .log file on nginx server.
Collected user journal information is stored in HDFS (Hadoop distributed field system using flume frame by S7.2
System) in.
S7.3 carries out data cleansing using HiveQL and places the data in Hbase database.
S7.4 carries out building for edge calculations big data platform using newest spark PC cluster engine, after cleaning
Data using based on job hunter collaborative filtering calculate, be as a result stored in the warehouse Hive, and by Sqoop importing
MySQL database.
S7.5 generates recommendation list and returns to user according to calculated result.
Collaborative filtering in the step S7.4 includes:
It is as shown in Equation 1 that S7.4.1 calculates the similarity matrix based on user
Wherein: N (u) indicates that user u had the set of the post information of browsing behavior；N (v) indicates that user v had browsing
The set of the post information of behavior.The coincidence degree of molecular moiety expression post information, it is clear that coincidence degree is higher, post information
It is more similar.A normalization has been done in denominator part, reduces the similarity degree for operating excessive user and other users.
S7.4.2 is clicked according to the behavior of user completes post recommendation
Wherein: rviUser v is indicated to the behavior score of post i, (such as: browsing launches resume, participation for different behaviors
Interview, enter official rank) difference is defined to the score of user behavior.SuvFor the similarity score of user u and user v, user v is user
The preceding k similar users of u, and the post user u for guaranteeing that user v was browsed was not browsed.We can be obtained by this way
Recommendation score of the user u to post i.
In order to carry out more accurate recommendation, we can also carry out improvement at two to formula simultaneously:
1. reducing contribution of the abnormal popular position to user's similarity
Wherein: SuvFor the similarity score of user u and user v, N (u) indicates that user u had the post information of browsing behavior
Set；N (v) indicates that user v had the set of the post information of browsing behavior；U (i) indicates the use for having behavior to post i
Family set, if a post was browsed by many job hunters, it will become lower in the contribution of registration.
2. different job hunters should give appropriate reduction to the period difference of same post behavior
Wherein: SuvFor the similarity score of user u and user v, N (u) indicates that user u had the post information of browsing behavior
Set；N (v) indicates that user v had the set of the post information of browsing behavior；Each post information contribution degree being overlapped obtains
Divide and be all different, by function f (Δ ti) determine.Function is defined as follows:
Wherein: | tui-tvi| user is indicated to the difference of the operating time in the same post, and the time is separated by shorter function
Value is higher, and the weight in registration contribution of response is higher.
In conclusion the invention discloses a kind of job hunter's recruitment information searching method based on collaborative filtering, incorporates
The data of existing major recruitment website, solve the problems, such as data not intercommunication, user are allow to inquire the whole network by this system
Recruitment information；Simultaneously in order to realize intelligent recommendation, collaborative filtering is applied in recruitment industry, the improvement to formula is passed through
Upgrading, the post information being more suitable is provided for user.
Claims (5)
1. a kind of job hunter's recruitment information searching method based on collaborative filtering, which comprises the steps of:
(1) the post information of existing major recruitment website is crawled by WebMagic web crawlers technology, storage is in the database；
(2) it builds the mobile terminal APP and plain engine platform is searched at computer PC network end, querying condition is set in platform；
(3) it after user's opening end APP or PC enters platform, carries out registration login and selects querying condition；
(4) system background for searching plain platform returns to suitable post information according to querying condition, and is ranked up displaying；
(5) user jumps to corresponding recruitment website by click sorted lists and checks post information；
(6) log information that system acquisition is clicked to user, cleans data；
(7) data after cleaning are calculated using the collaborative filtering based on job hunter, generate real-time recommendation list.
2. job hunter's recruitment information searching method based on collaborative filtering as described in claim 1, which is characterized in that step
Suddenly in (7), the data after cleaning are calculated using the collaborative filtering based on job hunter, and it is specific to generate real-time recommendation list
Include the following steps:
(71) code is write using Javascript in system front end bury a little, and user behavior information is stored in nginx
In .log file on server；
(72) collected user journal information is stored in HDFS using flume frame；
(73) data cleansing is carried out using HiveQL to place the data in Hbase database；
(74) building for edge calculations big data platform is carried out using newest spark PC cluster engine, by the number after cleaning
It calculates, is as a result stored in the warehouse Hive, and MySQL is imported by Sqoop according to the collaborative filtering used based on job hunter
Database；
(75) it according to calculated result, generates recommendation list and returns to user.
3. job hunter's recruitment information searching method based on collaborative filtering as claimed in claim 2, which is characterized in that step
Suddenly in (74), collaborative filtering specifically comprises the following steps:
(741) it is as shown in Equation 1 to calculate the similarity matrix based on user
Wherein: SuvFor the similarity score of user u and user v, N (u) indicates that user u had the collection of the post information of browsing behavior
It closes；N (v) indicates that user v had the set of the post information of browsing behavior；Molecular moiety indicates the coincidence degree of post information,
Obviously coincidence degree is higher, and post information is more similar；A normalization has been done in denominator part, reduce operate excessive user with
The similarity degree of other users；
(742) it is clicked according to the behavior of user and completes post recommendation
Wherein: PuiIt is user u to the recommendation score of post i, SuvFor the similarity score of user u and user v, k indicates user v
It is the preceding k similar users of user u, and the post user u for guaranteeing that user v was browsed was not browsed；rviIndicate v pairs of user
The behavior score of post i defines score of the different behaviors to user behavior different.
4. job hunter's recruitment information searching method based on collaborative filtering as claimed in claim 3, which is characterized in that step
Suddenly in (741), the similarity matrix based on user is rewritten are as follows:
Wherein: SuvFor the similarity score of user u and user v, N (u) indicates that user u had the collection of the post information of browsing behavior
It closes；N (v) indicates that user v had the set of the post information of browsing behavior；U (i) indicates to have post i the user of behavior to collect
It closes, if a post was browsed by many job hunters, it will become lower in the contribution of registration.
5. job hunter's recruitment information searching method based on collaborative filtering as claimed in claim 3, which is characterized in that step
Suddenly it in (742), is clicked according to the behavior of user and completes post recommendation rewriting are as follows:
Wherein: SuvFor the similarity score of user u and user v, N (u) indicates that user u had the collection of the post information of browsing behavior
It closes；N (v) indicates that user v had the set of the post information of browsing behavior；The post information contribution degree score that each is overlapped is equal
It is not identical, by function f (Δ ti) determine, function is defined as follows:
Wherein: | tui-tvi| user is indicated to the difference of the operating time in the same post, and the time is separated by the value of shorter function more
The weight in registration contribution of height, response is higher.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910336862.4A CN110110180A (en) | 2019-04-25 | 2019-04-25 | A kind of job hunter's recruitment information searching method based on collaborative filtering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910336862.4A CN110110180A (en) | 2019-04-25 | 2019-04-25 | A kind of job hunter's recruitment information searching method based on collaborative filtering |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110110180A true CN110110180A (en) | 2019-08-09 |
Family
ID=67486520
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910336862.4A Pending CN110110180A (en) | 2019-04-25 | 2019-04-25 | A kind of job hunter's recruitment information searching method based on collaborative filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110110180A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111737557A (en) * | 2020-08-21 | 2020-10-02 | 信度信息技术(苏州)有限公司 | Method and system for recruitment through Internet |
| CN111861361A (en) * | 2020-04-09 | 2020-10-30 | 河北利至人力资源服务有限公司 | Intelligent resume pushing system and method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104933239A (en) * | 2015-06-09 | 2015-09-23 | 江苏大学 | Hybrid model based personalized position information recommendation system and realization method therefor |
| CN105005880A (en) * | 2015-08-19 | 2015-10-28 | 郭文峰 | Internet based job application and recruitment system |
| CN105893641A (en) * | 2016-07-01 | 2016-08-24 | 中国传媒大学 | Job recommending method |
| CN107562818A (en) * | 2017-08-16 | 2018-01-09 | 中国工商银行股份有限公司 | Information recommendation system and method |
-
2019
- 2019-04-25 CN CN201910336862.4A patent/CN110110180A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104933239A (en) * | 2015-06-09 | 2015-09-23 | 江苏大学 | Hybrid model based personalized position information recommendation system and realization method therefor |
| CN105005880A (en) * | 2015-08-19 | 2015-10-28 | 郭文峰 | Internet based job application and recruitment system |
| CN105893641A (en) * | 2016-07-01 | 2016-08-24 | 中国传媒大学 | Job recommending method |
| CN107562818A (en) * | 2017-08-16 | 2018-01-09 | 中国工商银行股份有限公司 | Information recommendation system and method |
Non-Patent Citations (4)
| Title |
|---|
| 文哥的学习日记: "推荐系统理论（二）-利用用户行为数据进行推荐（协同过滤）", 《简书》 * |
| 欧阳裕洁等: "一种基于评分时间差的协同过滤算法", 《信息与电脑》 * |
| 蔡立志等: "《大数据测评》", 30 January 2015 * |
| 路小瑞: "基于Hadoop平台的职位推荐系统的设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111861361A (en) * | 2020-04-09 | 2020-10-30 | 河北利至人力资源服务有限公司 | Intelligent resume pushing system and method |
| CN111861361B (en) * | 2020-04-09 | 2021-07-27 | 河北利至人力资源服务有限公司 | Intelligent resume pushing system and method |
| CN111737557A (en) * | 2020-08-21 | 2020-10-02 | 信度信息技术(苏州)有限公司 | Method and system for recruitment through Internet |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11580104B2 (en) | Method, apparatus, device, and storage medium for intention recommendation | |
| CN108509551B (en) | Microblog network key user mining system and method based on Spark environment | |
| Rakhra et al. | Implementing Machine Learning for Smart Farming to Forecast Farmers' Interest in Hiring Equipment. | |
| US20190197416A1 (en) | Information recommendation method, apparatus, and server based on user data in an online forum | |
| Salehi et al. | Personalized recommendation of learning material using sequential pattern mining and attribute based collaborative filtering | |
| US10417301B2 (en) | Analytics based on scalable hierarchical categorization of web content | |
| CN104268292B (en) | The label Word library updating method of portrait system | |
| US8504411B1 (en) | Systems and methods for online user profiling and segmentation | |
| CN112307762B (en) | Search result sorting method and device, storage medium and electronic device | |
| CN111159341B (en) | Information recommendation method and device based on user investment and financial management preference | |
| US20150287051A1 (en) | System and method for identifying growing companies and monitoring growth using non-obvious parameters | |
| Khalid et al. | A literature review of implemented recommendation techniques used in massive open online courses | |
| Xiao et al. | Job recommendation with hawkes process: an effective solution for recsys challenge 2016 | |
| Liu et al. | QA document recommendations for communities of question–answering websites | |
| CN110110180A (en) | A kind of job hunter's recruitment information searching method based on collaborative filtering | |
| CN103814353A (en) | Search-based universal navigation | |
| Su et al. | Identifying and tracking topic-level influencers in the microblog streams | |
| Kim et al. | Topic-Driven SocialRank: Personalized search result ranking by identifying similar, credible users in a social network | |
| Alam et al. | Confluence of social network, social question and answering community, and user reputation model for information seeking and experts generation | |
| Sridhar et al. | A comparative study on how big data is scaling business intelligence and analytics | |
| CN110717089A (en) | User behavior analysis system and method based on weblog | |
| CN104809253A (en) | Internet data analysis system | |
| Becheru et al. | Towards social data analytics for smart tourism: A network science perspective | |
| Mathias et al. | Personalized sightseeing tours: a model for visits in art museums | |
| Nouvellet et al. | Discovery of usage patterns in digital library web logs using Markov modeling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190809 |
|
| RJ01 | Rejection of invention patent application after publication |