CN106021290A - Method for social network association excavation based on multi-scale geographic information - Google Patents

Method for social network association excavation based on multi-scale geographic information Download PDF

Info

Publication number
CN106021290A
CN106021290A CN201610285422.7A CN201610285422A CN106021290A CN 106021290 A CN106021290 A CN 106021290A CN 201610285422 A CN201610285422 A CN 201610285422A CN 106021290 A CN106021290 A CN 106021290A
Authority
CN
China
Prior art keywords
user
feature
registering
data
sample
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
Application number
CN201610285422.7A
Other languages
Chinese (zh)
Inventor
张凯
张晓宇
云晓春
王树鹏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Information Engineering of CAS
Original Assignee
Institute of Information Engineering of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institute of Information Engineering of CAS filed Critical Institute of Information Engineering of CAS
Priority to CN201610285422.7A priority Critical patent/CN106021290A/en
Publication of CN106021290A publication Critical patent/CN106021290A/en
Pending legal-status Critical Current

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
    • G06F16/242—Query formulation
    • 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
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/245—Classification techniques relating to the decision surface

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Remote Sensing (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention relates to a method for social network association excavation based on multi-scale geographic information. The method comprises the steps that (1) user sign-in data is acquired, pre-processing is carried out to the data, and structured data is obtained; (2) different methods for map division as well as scale standards for each method are set, and multiple corresponding position IDs are computed and acquired according to GPS data in the user sign-in data; (3) weights of the different positions are computed and acquired according to the quantities of sign-in users, and contributions made by the different positions for social relation prediction are represented; (4) characteristic extraction is carried out according to weight information of the positions, and position interaction characteristics of all the users can be obtained; (5) a classifier is trained by the extracted characteristics, and a relation prediction model is obtained; and (6) the obtained relation prediction model is used to predict a target user, and a prediction result of social network relations can be obtained. The method provided by the invention is characterized in that the position sign-in information is fully used to implement training and obtain the prediction model with higher robustness, and the ideal and stable prediction result can be obtained.

Description

A kind of social networks association mining method based on multi-scale geographic information
Technical field
The invention belongs to information technology, social networks technical field, be specifically related to a kind of based on multi-scale geographic information Social networks association mining method.
Background technology
In social networks research field, social networks association mining is an important research direction, and be other very Many researchs such as the basis of the research such as community discovery and commending system.Such as, people tend to buy the product that relatives and friends recommend, The people that corporations also tend to by being familiar with each other forms.Therefore, social networks association mining has become as social networks research The hot issue in field, and attracted extensive concern.
In traditional sense, social networks association mining utilizes the method for graph model to be predicted, i.e. social network often Network association mining network abstraction becomes graph model, and utilizes topology method to be predicted.In recent years, along with believing based on geographical position Developing rapidly of breath social networks, research worker starts with the positional information of user and predicts that social networks each other closes Connection.
At present, utilize user positional information prediction user between social networks association research, be concentrated mainly on for The extraction aspect of position feature, for position attribution and how to make full use of position attribution and be designed with effect relationship forecast model Studying less, main deficiency is:
A. the openness process shortcoming to the information of registering.The place confirmation method used at present is to utilize fixing grid Or additive method zoning, and the number of registering in each place and number of times of registering are sparse widely different, thus use fixing Size divides map and forms place, may result in registering of some place numerous, and the number of registering in some place Very little.This results in when carrying out social networks association mining, it is difficult to ensure the stability of prediction.The position that number of registering is many, There is social networks association in people, the position that number of registering is few, due between the people that observe between being easier to be predicted to be Less being therefore easier to of interaction be predicted as not having each other friend relation.
B. excavate insufficient for position attribution.When between to people, social networks association is predicted, different positions Put contribution different.The access information of another one user is detected if in user's family, the most permissible Confirm that two people exist social networks association.And detect that two people occur simultaneously if in park or library, then only it is difficult to Infer whether two people exist social networks association accordingly.
C. data tilt problem is processed not.Scale-model investigation currently with positional information prediction customer relationship is less, Especially when Relationship Prediction, the user that there is social networks association is less, causes the positive sample proportion when training forecast model The least, thus cause forecast model robustness the strongest.
Summary of the invention
The present invention proposes a kind of social networks association mining method based on multi-scale geographic information, by making full use of position Put the information of registering, train the more robust forecast model of acquisition, thus obtain predicting the outcome of ideal stability.
The technical solution used in the present invention is as follows:
A kind of social networks association mining method based on multi-scale geographic information, comprises the following steps:
1) obtain user to register data, it is carried out pretreatment and obtains structural data；
2) set distinct methods and the scale calibration of every kind of method of division map, register in data according to user Gps data calculates the multiple position ID obtaining its correspondence；
3) weight of acquisition diverse location is calculated according to number of registering, in order to characterize what social networks was predicted by diverse location Contribution；
4) weight information utilizing position carries out feature extraction, it is thus achieved that the position interaction feature of all users；
5) utilize the features training grader extracted, obtain Relationship Prediction model；
6) targeted customer is predicted by the Relationship Prediction model utilizing gained, it is thus achieved that social network relationships predicts the outcome.
Further, step 1) described structural data includes the time of registering of user, place of registering, the bases such as number of times of registering These data are stored by this information in the matrix form.
Further, step 4) feature extracted includes: common access locations feature based on weighting positional information, based on The Jaccard similarity feature of weighting positional information, cosine similarity feature based on weighting positional information.
Further, step 5) from two angle exercise benchmark graders: on the one hand, with all features of each yardstick It is characterized set, the most each yardstick correspondence one benchmark grader of training；On the other hand, with the single feature calculation of multiple yardsticks Method is characterized set, the most each feature calculation method correspondence one benchmark grader of training.Then the base of training gained is used Test set is classified by quasi-grader, and correspondence obtains confidence level set, and carries out confidence level set respectively from two angles Weighted average, is ranked up according to confidence level average, is respectively adopted top method and chooses the sample that certain proportion confidence level is bigger； If sample all occurs in the sample set that two angle sieves are selected, then set this sample as recommend sample, sample will be recommended Join in training set, and then obtain final grader, i.e. Relationship Prediction model.
Beneficial effects of the present invention is as follows:
1) present invention proposes to utilize multiple Standard Segmentation map, makes each the corresponding multiple position ID of information of registering, with this Overcome the problem being difficult to obtain stability forecast result due to single standard；
2) size that affects of prediction customer relationship is distributed weight according to each position by the present invention, thus more rationally sufficient Application site and information of registering；
3) present invention proposes the positional information method for digging of a kind of recommendation, and training obtains more robust forecast model；Use The present invention carries out social networks association mining, it was predicted that result includes that precision, recall rate, F value and accuracy rate all obtain ideal effect, There is good robustness and Generalization Capability, stable prediction effect can be obtained.
Accompanying drawing explanation
Fig. 1 is the basic step flow chart of the social networks association mining method of the present invention.
Fig. 2 is the concrete steps flow chart of the social networks association mining method of the present invention.
Detailed description of the invention
Understandable for enabling the above-mentioned purpose of the present invention, feature and advantage to become apparent from, below by specific embodiment and Accompanying drawing, the present invention will be further described.
The present invention is defined by data prediction, position ID, position weight calculating, feature extraction, forecast model train five Module (step) forms, and through the process of five modules, the information of registering of user i.e. can be utilized to complete the social networks of user Association mining.Function and principle to five modules are introduced below.
Module one data prediction
The user got in LBSN website data of registering are often unstructured data, it is difficult to carry out data analysis, because of First these data are processed into structural data by this.Through processing, the time of registering of user, place of registering (GPS format), label Store in the matrix form to essential informations such as number of times.
Module two position ID defines
Set n kind method and the scale calibration of every kind of method dividing map, according to the gps data of the information of registering, meter Calculate n the position ID obtaining its correspondence.
The method dividing map includes but not limited to: utilize the grid being staggered to form through parallel to divide map.Chi Scale standard refers to the longitude and latitude distance between every two warps or every two parallels, artificially can refer to according to practical application request Fixed, it is 0.1 degree for example with 0.1 degree as scale calibration, the distance between the most every two warps and between every two parallels.
Module three position weight calculates
Position weight is calculated by number of registering and obtains.When the social networks association between user is predicted, one The number of visiting people in place is the most, and this place is the least to the contribution of Relationship Prediction, such as in the phase of the public place such as park, library Meet, be difficult to prove accordingly between two users, to there is social networks association.Otherwise the number of visiting people of the three unities is the fewest, to relation The contribution of prediction is the biggest.But special circumstances are, as a three unities only user or when not having user to access, this ground The weight setting of point is 0.This is because this kind of place does not has any sign for the interactive information between user, the most not Reference information can be provided to the Relationship Prediction of user.According to description, the computing formula of setting position weight is as follows:
D k 0 , i f Σ i = 1 m d ( i , l k ) ≤ 1 1 ln Σ i = 1 m d ( i , l k ) Σ x = 1 t Σ i = 1 m d ( i , l x ) , i f Σ i = 1 m d ( i , l k ) > 2
Wherein, DkFor position weight, t is total number of positions, and m is total number of users, d (i, lk) represent whether user i can be in position K occurs.If there is then d (i, lk)=1, otherwise d (i, lk)=0.
Module four feature extraction
Feature extraction is the calculating to customer location interactive information, defines three kinds of feature calculation sides at this based on position attribution Method.
A. common access locations feature based on weighting positional information
If two users have the place of common access, social networks association between the two user, may be there is.And The weight of access locations is the biggest, there is a possibility that the biggest.Therefore binding site weight and common access locations set spy Levying extracting method is:
WhereinWithIt is respectively the set of the position ID that user i and user j accessed.
B. Jaccard similarity feature based on weighting positional information
Although common access locations feature is effective, but for any active ues (register number of times and the many user in place of registering) Process deviation easily occurs, it is proposed that binding site weight and Jaccard similarity design feature.Feature calculation formula For:
Wherein DxThe weight of the position that expression user i and user j went jointly, DyRepresent what user i or user j went The weight of position.
C. cosine similarity feature based on weighting positional information
User register custom similarity be another problem merited attention, user a certain place position register time Number can characterize the custom of registering of user to a certain extent.Utilize cosine similarity similar to the custom of registering calculating user at this Degree.With c (i, lk) representing the user i number of times of registering at place k, then the probability that user i occurs at place k can be described as The probability that so this user occurs in each place can use vector representation:
P (i)=< P (i, 1), P (i, 2) ... P (i, t) ＞
Accordingly, probability-weighted vector in position is:
WP (i)=< D1·P(i,1),D2·P(i,2)…DtP (i, t) ＞
Final this feature can be expressed as:
WS i j = W P ( i ) &CenterDot; W P ( j ) | W P ( i ) | | W P ( j ) |
Utilize above-mentioned three kinds of feature calculation methods, every a pair user is calculated, it is thus achieved that position letter between all users The feature of breath.
Module five forecast model is trained
Utilize the feature that a upper module calculates, from two angle exercise benchmark graders.As shown in Figure 2, on the one hand, with often The all of one yardstick are characterized as characteristic set, the most each yardstick correspondence one benchmark grader of training；On the other hand, with multiple The single feature calculation method of yardstick is characterized set, the most each feature calculation method correspondence one benchmark grader of training.Make Classifying test set with the benchmark grader of training gained, correspondence obtains confidence level set.From two angles, to confidence level Set is weighted average respectively, is ranked up according to confidence level average, is respectively adopted top method and chooses certain proportion confidence level Bigger sample.If sample all occurs in the sample set that two angle sieves are selected, then set this sample as recommending sample. Recommendation sample is joined in training set, (under each scale calibration, calculates institute with all feature calculation methods with complete characteristics collection Eigenvalue composition feature set) training obtain final grader, i.e. Relationship Prediction model.
Targeted customer is predicted by the Relationship Prediction model finally utilizing gained, it is thus achieved that Relationship Prediction result.
Provide below an application example.This example uses a location-based social network sites Gowalla to carry out user Social networks association mining.Herein, the definition to social networks association is: if two users exist neighbour in a communication network Contact, then it is assumed that the two user exists social networks association.
Gowalla is social networks based on geographical position, and user can register on website, and the information of registering includes The time of registering of user, place of registering (GPS format), register the time etc..Choosing the most city Jane Austen of data of registering is mesh Mark ground, studies the user on this ground.Choose data of registering 2010 more than the user of 50 as goal in research, total 1585 users, 722,598 data of registering, these users have 7356 between there is call relation.
Through pretreatment, it is structural data by the information processing of registering of user, then with longitude and latitude 0.1 degree, 0.01 degree, 0.001 degree is the square zoning on limit, determines the position ID in place of registering according to the gps data in the information of registering, every Corresponding three the position ID of information of registering.The three kinds of feature calculation methods based on position weight utilizing the present invention to propose extract user Position interaction feature, is then divided into training set and test set by sample set, and wherein 50% sample is as training set, 50% sample This is as test set.Then use Multiple regression model as benchmark grader, utilize training set from two angle exercises Benchmark grader.Sample is recommended to be set as in test set the 20% of the positive sample size of sample to be tested.Prediction reaches ideal effect.Calculate Method false code is as follows:
Above example is only limited in order to technical scheme to be described, the ordinary skill of this area Technical scheme can be modified or equivalent by personnel, without departing from the spirit and scope of the present invention, and this The protection domain of invention should be as the criterion with described in claims.

Claims (9)

1. a social networks association mining method based on multi-scale geographic information, it is characterised in that comprise the following steps:
1) obtain user to register data, it is carried out pretreatment and obtains structural data；
2) distinct methods and the scale calibration of every kind of method, the GPS number in data of registering according to user of division map are set According to calculating the multiple position ID obtaining its correspondence；
3) weight of acquisition diverse location is calculated according to number of registering, in order to characterize the tribute that social networks is predicted by diverse location Offer；
4) weight information utilizing position carries out feature extraction, it is thus achieved that the position interaction feature of all users；
5) utilize the features training grader extracted, obtain Relationship Prediction model；
6) targeted customer is predicted by the Relationship Prediction model utilizing gained, it is thus achieved that social network relationships predicts the outcome.
2. the method for claim 1, it is characterised in that: step 1) described structural data includes when registering of user Between, place of registering, number of times of registering, these data are stored in the matrix form.
3. the method for claim 1, it is characterised in that: step 3) according to equation below calculating position weight:
D k = 0 , i f &Sigma; i = 1 m d ( i , l k ) &le; 1 1 l n &Sigma; i = 1 m d ( i , l k ) &Sigma; x = 1 t &Sigma; i = 1 m d ( i , l x ) , i f &Sigma; i = 1 m d ( i , l k ) > 2
Wherein, DkFor position weight, t is total number of positions, and m is total number of users；d(i,lk) represent whether user i can go out at position k Existing, if there is then d (i, lk)=1, otherwise d (i, lk)=0.
4. method as claimed in claim 3, it is characterised in that step 4) feature extracted includes: based on weighting positional information Common access locations feature, based on weighting positional information Jaccard similarity feature, based on weighting positional information cosine Similarity feature.
5. method as claimed in claim 4, it is characterised in that described common access locations feature based on weighting positional information Computing formula be:
WhereinWithIt is respectively the set of the position ID that user i and user j accessed.
6. method as claimed in claim 4, it is characterised in that described Jaccard similarity based on weighting positional information is special The computing formula levied is:
Wherein DxThe weight of the position that expression user i and user j went jointly, DyRepresent the position that user i or user j went Weight.
7. method as claimed in claim 4, it is characterised in that described cosine similarity feature based on weighting positional information Computing formula is:
WS i j = W P ( i ) &CenterDot; W P ( j ) | W P ( i ) | | W P ( j ) | ,
Wherein, the position probability-weighted vector that WP (i) and WP (j) user i and j occurs in each place.
8. the method for claim 1, it is characterised in that: step 5) from two angle exercise benchmark graders: on the one hand, All with each yardstick are characterized as characteristic set, the most each yardstick correspondence one benchmark grader of training；On the other hand, with The single feature calculation method of multiple yardsticks is characterized set, and the most each feature calculation method correspondence trains a benchmark classification Device.
9. method as claimed in claim 8, it is characterised in that: step 5) use the benchmark grader of training gained to test set Classifying, correspondence obtains confidence level set, and is weighted averagely from two angles respectively to confidence level set, according to confidence Degree average is ranked up, and is respectively adopted top method and chooses the sample that certain proportion confidence level is bigger；If sample is two angles The sample set filtered out all occurs, then set this sample as recommend sample, recommendation sample is joined in training set, and then Obtain final grader, i.e. Relationship Prediction model.
CN201610285422.7A 2016-04-29 2016-04-29 Method for social network association excavation based on multi-scale geographic information Pending CN106021290A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610285422.7A CN106021290A (en) 2016-04-29 2016-04-29 Method for social network association excavation based on multi-scale geographic information

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610285422.7A CN106021290A (en) 2016-04-29 2016-04-29 Method for social network association excavation based on multi-scale geographic information

Publications (1)

Publication Number Publication Date
CN106021290A true CN106021290A (en) 2016-10-12

Family

ID=57081708

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610285422.7A Pending CN106021290A (en) 2016-04-29 2016-04-29 Method for social network association excavation based on multi-scale geographic information

Country Status (1)

Country Link
CN (1) CN106021290A (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106778851A (en) * 2016-12-05 2017-05-31 公安部第三研究所 Social networks forecasting system and its method based on Mobile Phone Forensics data
CN107038649A (en) * 2017-05-10 2017-08-11 广东小天才科技有限公司 Friend recommendation method and device for terminal user
CN107169088A (en) * 2017-05-12 2017-09-15 中国矿业大学 A kind of user social contact relationship strength computational methods and system interacted based on space-time
CN108074016A (en) * 2017-12-25 2018-05-25 苏州大学 Customer relationship intensity prediction method, device and equipment based on position social networks
CN108268519A (en) * 2016-12-30 2018-07-10 阿里巴巴集团控股有限公司 A kind of method and apparatus of recommendation network object
CN108345662A (en) * 2018-02-01 2018-07-31 福建师范大学 A kind of microblog data weighted statistical method of registering considering user distribution area differentiation
CN112800111A (en) * 2021-01-26 2021-05-14 重庆邮电大学 Position prediction method based on training data mining

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103154993A (en) * 2010-08-18 2013-06-12 费斯布克公司 Location ranking using social graph information
CN103795613A (en) * 2014-01-16 2014-05-14 西北工业大学 Method for predicting friend relationships in online social network
WO2015014281A1 (en) * 2013-07-31 2015-02-05 Tencent Technology (Shenzhen) Company Limited Method and apparatus for displaying living service information
CN104541273A (en) * 2012-08-20 2015-04-22 微软公司 Social relevance to infer information about points of interest
CN104657434A (en) * 2015-01-30 2015-05-27 中国科学院信息工程研究所 Construction method for social network structure

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103154993A (en) * 2010-08-18 2013-06-12 费斯布克公司 Location ranking using social graph information
CN104541273A (en) * 2012-08-20 2015-04-22 微软公司 Social relevance to infer information about points of interest
WO2015014281A1 (en) * 2013-07-31 2015-02-05 Tencent Technology (Shenzhen) Company Limited Method and apparatus for displaying living service information
CN103795613A (en) * 2014-01-16 2014-05-14 西北工业大学 Method for predicting friend relationships in online social network
CN104657434A (en) * 2015-01-30 2015-05-27 中国科学院信息工程研究所 Construction method for social network structure

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106778851A (en) * 2016-12-05 2017-05-31 公安部第三研究所 Social networks forecasting system and its method based on Mobile Phone Forensics data
CN106778851B (en) * 2016-12-05 2020-05-01 公安部第三研究所 Social relationship prediction system and method based on mobile phone evidence obtaining data
CN108268519A (en) * 2016-12-30 2018-07-10 阿里巴巴集团控股有限公司 A kind of method and apparatus of recommendation network object
CN107038649A (en) * 2017-05-10 2017-08-11 广东小天才科技有限公司 Friend recommendation method and device for terminal user
CN107038649B (en) * 2017-05-10 2021-03-26 广东小天才科技有限公司 Friend recommendation method and device for terminal user
CN107169088A (en) * 2017-05-12 2017-09-15 中国矿业大学 A kind of user social contact relationship strength computational methods and system interacted based on space-time
CN107169088B (en) * 2017-05-12 2020-05-12 中国矿业大学 User social relationship strength calculation method and system based on time-space interaction
CN108074016A (en) * 2017-12-25 2018-05-25 苏州大学 Customer relationship intensity prediction method, device and equipment based on position social networks
CN108074016B (en) * 2017-12-25 2021-07-30 苏州大学 User relationship strength prediction method, device and equipment based on location social network
CN108345662A (en) * 2018-02-01 2018-07-31 福建师范大学 A kind of microblog data weighted statistical method of registering considering user distribution area differentiation
CN112800111A (en) * 2021-01-26 2021-05-14 重庆邮电大学 Position prediction method based on training data mining

Similar Documents

Publication Publication Date Title
Yang et al. Integrated flood vulnerability assessment approach based on TOPSIS and Shannon entropy methods
CN106021290A (en) Method for social network association excavation based on multi-scale geographic information
He et al. Mining transition rules of cellular automata for simulating urban expansion by using the deep learning techniques
Guzzetti et al. Estimating the quality of landslide susceptibility models
CN103795613B (en) Method for predicting friend relationships in online social network
Feng et al. Incorporation of spatial heterogeneity-weighted neighborhood into cellular automata for dynamic urban growth simulation
Yan et al. Optimizing landslide susceptibility mapping in the Kongtong District, NW China: comparing the subdivision criteria of factors
Pulugurtha et al. Modeling AADT on local functionally classified roads using land use, road density, and nearest nonlocal road data
Li et al. A Bayesian method to mine spatial data sets to evaluate the vulnerability of human beings to catastrophic risk
Stachelek et al. Application of inverse path distance weighting for high-density spatial mapping of coastal water quality patterns
Giannakidou et al. Implementing a flood vulnerability index in urban coastal areas with industrial activity
CN104679942A (en) Construction land bearing efficiency measuring method based on data mining
Wang et al. Detecting village‐level regional development differences: A GIS and HLM method
Martínez et al. Atlas of surface currents in the Mediterranean and Canary–Iberian–Biscay waters
Koga et al. Assessing impervious area ratios of grid-based land-use classifications on the example of an urban watershed
Wang et al. A comparison of proximity and accessibility drivers in simulating dynamic urban growth
CN110716998B (en) Fine scale population data spatialization method
Mathew et al. Comparative assessment of geospatial and statistical methods to estimate local road annual average daily traffic
CN109784602A (en) A kind of disaster-ridden kind of coupling physical vulnerability assessment method based on PTVA model
Alhazzani et al. Urban Attractors: Discovering patterns in regions of attraction in cities
Kobayashi et al. Development of a distributed rainfall‐run‐off/flood‐inundation simulation and economic risk assessment model
Li et al. Using spatial analysis and Bayesian network to model the vulnerability and make insurance pricing of catastrophic risk
Zhang et al. Incorporation of spatial anisotropy in urban expansion modelling with cellular automata
Wang et al. Gully erosion susceptibility prediction in Mollisols using machine learning models
CN113901348A (en) Oncomelania snail distribution influence factor identification and prediction method based on mathematical model

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

Application publication date: 20161012

RJ01 Rejection of invention patent application after publication