CN106295895A - A kind of position transition Forecasting Methodology based on occupation track data - Google Patents
A kind of position transition Forecasting Methodology based on occupation track data Download PDFInfo
- Publication number
- CN106295895A CN106295895A CN201610669536.1A CN201610669536A CN106295895A CN 106295895 A CN106295895 A CN 106295895A CN 201610669536 A CN201610669536 A CN 201610669536A CN 106295895 A CN106295895 A CN 106295895A
- Authority
- CN
- China
- Prior art keywords
- user
- transition
- data
- position transition
- forecasting methodology
- 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
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2462—Approximate or statistical queries
-
- 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Development Economics (AREA)
- Software Systems (AREA)
- Game Theory and Decision Science (AREA)
- General Engineering & Computer Science (AREA)
- Fuzzy Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of position transition Forecasting Methodology based on occupation track data, on the model analyzing the transition of user's position, find the influence factor of user's position transition, the Forecasting Methodology of research user's position transition.In data base, extract resume, obtain the history occupation track data of user；By large-scale consumer occupation track data being carried out statistical classification contrast, finding user's position transition space-time characteristic, building user's position transition space-time model；From user's position transition space-time model, extract and quantify the influence factor of user's position transition；Data are changed with quantization, the position in conjunction with user, by decision Tree algorithms training position transition forecast model according to position transition influence factor's definition；In S4 based on position transition forecast model, according to the current position of user, it was predicted that the position after user's occupation transition.The present invention can more comprehensively consider the influence factor that user's position changes, thus preferably predicts its accuracy rate changed.
Description
Technical field
The invention belongs to data mining analysis field, a kind of position transition prediction side based on occupation track data
Method.
Background technology
In recent years, the concern of more and more research worker is received based on user's occupation trajectory predictions user's position transition.
So-called position transition prediction, it is common that based on referring to the history occupation track data according to user, in conjunction with the personal feature of user
Information, if user exists the transition behavior of occupation, can be predicted the situation of user's next one position.Wherein, history duty
Industry track data is mainly derived from user and is supplemented perfect data in occupation social networks.By to history occupation track number
According to excavation, build the space-time model of user's position transition, find position transition influence factor, can be accurately by existing algorithm
Prediction user's position transition situation.
By the observation to user's history occupation track data, wherein it is seen that, user's position transition are contained abundant
Time and space idea.For example, position on the middle and senior level, current title and rank is not less than for next title and rank；No
Same academic background, the rank for position can produce impact, well educated, can have the position of higher level；Working experience, with
Title and rank can be impacted by sample.
Summary of the invention
In order to solve when user's occupation changes the problem for obtained position situation, the present invention provides a kind of based on going through
The position transition Forecasting Methodology of history occupation track data, the method, based on user's occupation data, finds position transition impact
Position transition, in conjunction with existing traditional decision-tree, are predicted by factor.
For achieving the above object, the technical scheme that the present invention takes is:
A kind of based on occupation track data position transition Forecasting Methodology, it is characterised in that include following suddenly:
S1, in data base extract resume, obtain user history occupation track data；
S2, by large-scale consumer occupation track data being carried out statistical classification contrast, find user's position transition space-time characteristic,
Build user's position transition space-time model；
S3, from user's position transition space-time model, extract and quantify user's position transition influence factor；
S4, according to position transition influence factor definition with quantify, in conjunction with user position change data, instructed by decision Tree algorithms
Practice position transition forecast model.
S5, in S4 position transition forecast model based on, according to current position P of user, it was predicted that user's occupation change
After position P '.
Further, history occupation track data bag in a kind of position transition Forecasting Methodology S1 based on occupation track data
The data messages such as education background situation containing user, the transition of history position, working time length；Described history position transition bag
Include the scale of place company, the rank of place position.
Further, in a kind of position transition Forecasting Methodology S1 based on occupation track data, space-time characteristic included on the time
The feature of user's position transition in the feature of user's position transition and position.
Further, a kind of position transition Forecasting Methodology S2 based on occupation track data comprises sub-step: S21: first
Situation of change in time and the time span of user job, feature in discovery time is changed to analyzing user's position；
S22: then to different company, the position change of different positions and different educational background user is added up and is measured, and studies and becomes
Law, finds spatially feature；S23: last from the time and spatially to change description in user's position, by number
According to matching, find user's position Changing Pattern.
Further, user's position transition spatial model in a kind of position transition Forecasting Methodology based on occupation track data
It is expressed as user's position and changes various information: include the temporal information in user's occupation track data, job information, company information
And individual's education background is used for describing user's position transition information.
Further, described in a kind of position transition Forecasting Methodology S3 based on occupation track data, influence factor comprises:
Company factorWherein no. represents company personnel's quantity；
Position factorWherein po represents the rank of user's position；
Educational factors EF=∑ De, wherein De represents educational background, and the quantization of De is defined as:
Time factor DF=leaving date-hiring date；
Position accumulation factor
Further, position transition prediction mould described in S4 in a kind of position transition Forecasting Methodology based on occupation track data
Type includes three parts, respectively input, it was predicted that part and outfan, and input is the position transition data of user, it was predicted that end
For extracting influence factor in data, being predicted position by calculating process according to the model of training, outfan is user's
Next job information.
Further, influence factor's bag described in S3 in a kind of position transition Forecasting Methodology based on occupation track data
Contain: i.e. change data from the position of each user i and extract above-mentioned influence factor respectively, constitute bit vector Vi=
CFi, PFi, EFi, DFi, PAi}, by Vi and next position Pi one_to_one corresponding, pre-by decision Tree algorithms training position transition
Survey model.
The method have the advantages that
In the method, more comprehensively make use of the occupation data of user's history, it was found that include company, position, education, time
Between, the multiple position such as position accumulation transition influence factor, thus the higher predictablity rate obtained.
Accompanying drawing explanation
Fig. 1 is present invention position based on occupation track data transition Forecasting Methodology flow chart；
Fig. 2 is present invention position based on occupation track data transition Forecasting Methodology temporal characteristics position change profile figure；
Fig. 3 is present invention position based on occupation track data transition Forecasting Methodology space characteristics position change profile figure；
Fig. 4 is present invention position based on occupation track data transition Forecasting Methodology time and space feature position change matching signal
Figure；
Fig. 5 is present invention position based on occupation track data transition Forecasting Methodology position prediction schematic diagram；
Fig. 6 is that present invention position transition based on occupation track data Forecasting Methodology difference position predicts accurate statistics figure.
Detailed description of the invention
In order to make objects and advantages of the present invention clearer, below in conjunction with embodiment, the present invention is carried out further
Describe in detail.Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not used to limit this
Bright.
S1, in data base extract resume, obtain user history occupation track data.
History occupation track data comprises the numbers such as the education background situation of user, the transition of history position, working time length
It is believed that breath；History position transition include the scale of place company, the rank of place position.By these occupation tracks according to user
People's information and position are divided into the data slot of position transition, i.e. from a position to another position.
In the present embodiment, the history occupation track data of totally 10418 users, is divided into 46673 data fragments.
S2, by large-scale consumer occupation track data being carried out statistical classification contrast, find user's position transition space-time
Feature, builds user's position transition space-time model.
Position transition fragment in S1 is researched and analysed, Changing Pattern between the different position of statistics, and in company, time
Between, the rule of position change under educational background different condition.Comprehensive law, proposes the space-time characteristic of position change, builds position change
Space-time model, particularly as follows:
S21: first analysis user's position is changed situation of change in time and the time span of user job, during discovery
Feature between, as shown in Figure 2；
S22: then to different company, the position change of different positions and different educational background user is added up and is measured, and grinds
Study carefully Changing Pattern, find spatially feature, as shown in Figure 3；
S23: last from the time and spatially to change description in user's position, by data matching, finds user's duty
Position Changing Pattern, as shown in Figure 4, illustrates time and space joint effect position situation of change.
Legend illustrated above is the statistical fit of mass data in S1 and analyzes gained.
S3, from user's position transition space-time model, extract and quantify user's position transition influence factor.
Company factor is quantified according to company personnel's quantity by the present embodiment, according to title and rank, position factor is entered
Row quantifies, and measures educational factors according to degree situation, measures time factor according to the working time, according to history
Position accumulation factor is measured by upper position and two factors of time:
Company factorWherein no. represents company personnel's quantity；
Position factorWherein po represents the rank of user's position；
Educational factors EF=∑ De, wherein De represents educational background, and the quantization of De is defined as:
Time factor DF=leaving date-hiring date；
Position accumulation factor
S4, according to position transition influence factor definition with quantify, in conjunction with user position change data, calculated by decision tree
Method training position transition forecast model.
Position transition forecast model includes three parts, respectively input, it was predicted that part and outfan, input is user
Position transition data, it was predicted that hold as extracting influence factor in data, by calculating process, position entered according to the model of training
Row prediction, outfan is the next job information of user.Influence factor is changed, in conjunction with the history position number of user according to position
According to, using these factor data as training input, use decision Tree algorithms training data, generate position transition forecast model originally
Embodiment changes data from the position of each user i and extracts above-mentioned influence factor respectively, constitute bit vector Vi
={ CFi, PFi, EFi, DFi, PAi}, by Vi and next position Pi one_to_one corresponding, train position transition by decision Tree algorithms
Forecast model.
The present embodiment employs the data of 80% in S1 and, as training data, is predicted model training and generates.
S5, in S4 position transition forecast model based on, according to the current position of user, it was predicted that user's occupation transition after
Position.
The history occupation track data of input user, when predecessor company, position, education background, working time, by prediction mould
Type is predicted calculating, and generates next position situation, as shown in Figure 5.Fixed for predicting that the position generated still uses in S3
Justice, judges for difference prediction position is done accuracy, and the present embodiment uses the data in the step 1 of residue 20% for pre-
Surveying, as shown in Figure 6, position predictablity rate is up to 74% in the present embodiment for different position forecasting accuracy results.
The above is only the implementation process of the present invention, it is noted that come for those skilled in the art
Saying, under the premise without departing from the principles of the invention, it is also possible to make some improvements and modifications, these improvements and modifications also should be regarded as
Protection scope of the present invention.
Claims (8)
1. one kind based on occupation track data position transition Forecasting Methodology, it is characterised in that include following suddenly:
S1, in data base extract resume, obtain user history occupation track data；
S2, by large-scale consumer occupation track data being carried out statistical classification contrast, find user's position transition space-time characteristic,
Build user's position transition space-time model；
S3, from user's position transition space-time model, extract and quantify user's position transition influence factor；
S4, according to position transition influence factor definition with quantify, in conjunction with user position change data, instructed by decision Tree algorithms
Practice position transition forecast model；
S5, in S4 position transition forecast model based on, according to the current position of user, it was predicted that user's occupation transition after duty
Position.
The most according to claim 1 in position Forecasting Methodology, it is characterised in that the history occupation track data bag described in S1
The data messages such as education background situation containing user, the transition of history position, working time length；Described history position transition bag
Include the scale of place company, the rank of place position.
The most according to claim 1 in position Forecasting Methodology, it is characterised in that the space-time characteristic described in S2 included on the time
The feature of user's position transition in the feature of user's position transition and position.
The most according to claim 3 in position Forecasting Methodology, it is characterised in that described S2 comprises sub-step:
S21: first analysis user's position is changed situation of change in time and the time span of user job, during discovery
Feature between；
S22: then to different company, the position change of different positions and different educational background user is added up and is measured, and grinds
Study carefully Changing Pattern, find spatially feature；
S23: last from the time and spatially to change description in user's position, by data matching, finds user's duty
Position Changing Pattern.
5. according in position Forecasting Methodology described in claim 1-4, it is characterised in that described user's position transition spatial model
It is expressed as user's position and changes various information: include the temporal information in user's occupation track data, job information, company information
And individual's education background is used for describing user's position transition information.
The most according to claim 1 in position Forecasting Methodology, it is characterised in that described in S3, influence factor comprises:
Company factorAnd no. ＜ 5001 wherein no. represents company personnel's quantity；
Position factorWherein po represents the rank of user's position；
Educational factors EF=Σ De, wherein De represents educational background, and the quantization of De is defined as:
Time factor DF=leaving date-hiring date；
Position accumulation factor
The most according to claim 1 in position Forecasting Methodology, it is characterised in that described in S4, position transition forecast model includes three
Part, respectively input, it was predicted that part and outfan, input is the position transition data of user, it was predicted that hold as extracting number
According to middle influence factor, being predicted position by calculating process according to the model of training, outfan is the next duty of user
Position information.
8. according in position Forecasting Methodology described in claim 6-7, it is characterised in that described in S3, influence factor comprises: i.e. from
The position transition data of each user i extract above-mentioned influence factor respectively, constitutes bit vector Vi={CFi, PFi,
EFi, DFi, PAi}, by Vi and next position Pi one_to_one corresponding, by decision Tree algorithms training position transition forecast model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610669536.1A CN106295895A (en) | 2016-08-15 | 2016-08-15 | A kind of position transition Forecasting Methodology based on occupation track data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610669536.1A CN106295895A (en) | 2016-08-15 | 2016-08-15 | A kind of position transition Forecasting Methodology based on occupation track data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106295895A true CN106295895A (en) | 2017-01-04 |
Family
ID=57671562
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610669536.1A Pending CN106295895A (en) | 2016-08-15 | 2016-08-15 | A kind of position transition Forecasting Methodology based on occupation track data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106295895A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108834079A (en) * | 2018-09-21 | 2018-11-16 | 北京邮电大学 | A kind of load balance optimization method based on mobility prediction in heterogeneous network |
| CN108921497A (en) * | 2018-06-05 | 2018-11-30 | 北京纳人网络科技有限公司 | Information processing method and device |
| CN113742563A (en) * | 2020-05-28 | 2021-12-03 | 北京百度网讯科技有限公司 | Method, device, equipment and medium for establishing work prediction model and recommending work |
| CN114048243A (en) * | 2021-10-19 | 2022-02-15 | 盐城金堤科技有限公司 | Method and device for mining personnel transition process, storage medium and electronic equipment |
-
2016
- 2016-08-15 CN CN201610669536.1A patent/CN106295895A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108921497A (en) * | 2018-06-05 | 2018-11-30 | 北京纳人网络科技有限公司 | Information processing method and device |
| CN108834079A (en) * | 2018-09-21 | 2018-11-16 | 北京邮电大学 | A kind of load balance optimization method based on mobility prediction in heterogeneous network |
| CN113742563A (en) * | 2020-05-28 | 2021-12-03 | 北京百度网讯科技有限公司 | Method, device, equipment and medium for establishing work prediction model and recommending work |
| CN113742563B (en) * | 2020-05-28 | 2023-08-18 | 北京百度网讯科技有限公司 | Work prediction model establishment and work recommendation method, device, equipment and medium |
| CN114048243A (en) * | 2021-10-19 | 2022-02-15 | 盐城金堤科技有限公司 | Method and device for mining personnel transition process, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109741112B (en) | User purchase intention prediction method based on mobile big data | |
| Xiao et al. | On Modeling and Predicting Individual Paper Citation Count over Time. | |
| Zhu et al. | Forecasting carbon price using empirical mode decomposition and evolutionary least squares support vector regression | |
| CN106295895A (en) | A kind of position transition Forecasting Methodology based on occupation track data | |
| Abdel Sabour et al. | Incorporating geological and market uncertainties and operational flexibility into open pit mine design | |
| Fagundes et al. | Robust regression with application to symbolic interval data | |
| CN104063429A (en) | Predicting method for user behavior in e-commerce | |
| Ladino et al. | Travel time forecasting from clustered time series via optimal fusion strategy | |
| Gay et al. | Novel resilience assessment methodology for water distribution systems | |
| CN110852792B (en) | Route value evaluation method based on big data analysis and related products | |
| El Maghraoui et al. | Smart energy management system: a comparative study of energy consumption prediction algorithms for a hotel building | |
| Kunjir et al. | A comparative study of predictive machine learning algorithms for COVID-19 trends and analysis | |
| Cao et al. | Efficient fine-grained location prediction based on user mobility pattern in lbsns | |
| CN110019563B (en) | Portrait modeling method and device based on multi-dimensional data | |
| Julian et al. | Application of machine learning to link prediction | |
| Herrema et al. | A novel machine learning model to predict abnormal Runway Occupancy Times and observe related precursors | |
| Zhu et al. | Validating rail transit assignment models with cluster analysis and automatic fare collection data | |
| CN103544503A (en) | Behavior recognition method based on multi-instance markov model | |
| Zhou et al. | Traffic conduction analysis model with time series rule mining | |
| Singh et al. | A hybrid surrogate based algorithm (HSBA) to solve computationally expensive optimization problems | |
| JP2013156691A (en) | Purchase prediction device, method, and program | |
| Xie et al. | Walmart Sale Forecasting Model Based On LSTM And LightGBM | |
| Panda et al. | Machine learning using exploratory analysis to predict taxi fare | |
| CN110019166A (en) | Screen the method and customer defection early warning method of attribute data | |
| Baas et al. | Predicting virality on networks using local graphlet frequency distribution |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170104 |