CN106682686A - User gender prediction method based on mobile phone Internet-surfing behavior - Google Patents
User gender prediction method based on mobile phone Internet-surfing behavior Download PDFInfo
- Publication number
- CN106682686A CN106682686A CN201611127122.2A CN201611127122A CN106682686A CN 106682686 A CN106682686 A CN 106682686A CN 201611127122 A CN201611127122 A CN 201611127122A CN 106682686 A CN106682686 A CN 106682686A
- Authority
- CN
- China
- Prior art keywords
- user
- data
- accuracy
- cell phone
- prediction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000012549 training Methods 0.000 claims abstract description 29
- 239000011159 matrix material Substances 0.000 claims abstract description 25
- 238000012360 testing method Methods 0.000 claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 claims description 24
- 238000007637 random forest analysis Methods 0.000 claims description 9
- 238000003066 decision tree Methods 0.000 claims description 3
- 230000006399 behavior Effects 0.000 abstract description 19
- 238000004458 analytical method Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 238000010276 construction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000007477 logistic regression Methods 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005266 casting Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000013332 literature search Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to a user gender prediction method based on mobile phone Internet-surfing behaviors. The method includes the following steps: counting the number of times of APPs clicked by a user within a period of time; arranging the counted data in the form of a matrix; conducting dimension reduction processing on the matrix; separating the processed data into a training data set and a test data set, and using the training data set to train a prediction model; and using the test data set to verify the prediction model, and calculating accuracy. The method is easy and feasible, and is high in accuracy. The gender of the user is predicted according to the number of times that the user uses the APPs, and supports are provided for relevant individualized services in a later period on the basis of preferences of users of different genders.
Description
Technical field
The invention belongs to computer, communication technical field, and in particular to a kind of user's sex based on surfing Internet with cell phone behavior
Forecasting Methodology.
Background technology
With the development of big data, many network application such as e-commerce website, search engine etc., use is become more concerned with
The primary attribute information at family, by the portrait for building user, the service subsequently to carry out personalization to user provides strong
Support.The personalized shop of Amazon is exactly the good example of personalized service.Amazon shopping center can be according to user's
Browse record and purchaser record recommends corresponding commodity to promote customer consumption to user, or buys identical or phase by analysis
Like product other users purchasing behavior be user's Recommendations.
Ever-increasing simultaneously in the advertising business of Internet firm, many Internet firm is also providing the user
The advertisement of property.Baidu's promotion conference provides the user with different advertising businesses according to the search history keyword of user.Many
Company can analyze the interest model of user according to the historical behavior of user, to provide more preferable personalized service.Portray user
Portrait includes portraying the demographic information of user, geographical location information, search Access Interest hobby etc..Portray user and draw a portrait it
In very important part be exactly user characteristicses identification.User characteristicses identification is in personalized recommendation, suspicious user identification etc.
Aspect all plays an important role.Such as in terms of marketing, it is determined that the individual characteristicss (such as age, sex) of user, just
Autotelic marketing can be carried out for the user.Compared to the formula marketing of casting net of blindness, the success rate of precision marketing will be higher,
Bigger value can be created.Equally, public safety has become one problem that can not be ignored of today's society, some crimes point
Son often hides whereabouts, does not leave any vestige in places such as networks, but mobile phone takes electricity as indispensable instrument
Words, base station can just expose the position of user when switching, while the information such as which APP for using of user and online custom also can be helped
The feature for describing the user is helped, is helped related law enforcement agency to reduce and is detected scope, lock onto target.But the primary attribute of user is believed
Breath such as sex, age, income etc. are typically to be not readily available, because these information are for a user very sensitive
, people are unwilling to disclose this kind of individual privacy attribute.Therefore, the identification of user base feature is the focus studied in recent years.
With the development and the popularization of smart mobile phone of information technology, smart mobile phone application is presented exponential increase, based on position
Service it is also more and more.These location-based network applications permit user and issue the geographical location information of oneself, search for attached
Near people, share experience of individual etc., at the same time network application can be the on-site speciality shop of user recommended user, user
People interested or thing, these need to combine the geographical location information of user, hobby and personal primary attribute information.So
And these information are generally viewed as user privacy information, many network application companies are difficult to obtain.Although some network applications
Require that user fills in the relevant informations such as sex, date of birth, education degree, but these information to user's ratio in user's registration
It is more sensitive, therefore much user never fills in these relevant informations or fills in the information of mistake, these false letters
Breath has negative interaction to personalized recommendation.Practical situation is the primary attribute that most of user does not fill in correlation in registration
Information.
Due to the difference of sex for mobile phone application, the APP that masculinity and femininity is used also is not quite similar.Different attribute
User is not quite similar using the frequency of APP, but, the APP species of the user's preferences of same alike result is roughly the same, thus same category
Property user using same APP the frequency it is roughly the same.Therefore use is predicted by studying user using the frequency per a APP
This problem of the primary attribute information at family provides new thinking and method to predict user's unknown message.
Found by literature search to prior art, the user base attribute of early stage be based primarily upon user version data and
Speech data differentiates according to the primary attribute that everyone behavioural habits, writing style etc. carries out age and sex.Eckert etc.
Using sociolinguistics, by the sex for studying the language feature of user to infer user;Koppel etc. is according to user's sex pair
The word of some authors carries out text classification, finds in terms of diction and word content, the author at different sexes and age
Between have larger difference, accordingly he proposes Multi-Class Real Winnow algorithms, according to the Blog content of author
The age and sex of author are classified, preferable effect is achieved.
At present, the research towards the user base attribute identification of mobile terminal internet behavior data is relatively fewer.Prediction user
The method of primary attribute is based primarily upon traditional classifier methods.
1st, LR (logistic regression) algorithm.The algorithm is a kind of sorting technique, is mainly used in two classification.Its predictive value only has
Two, yes/no.Gender prediction can be with it.The algorithm is required relatively strictly, when the independent variable mistake in model independent variable
When many, over-fitting is easily caused.
2nd, NB Algorithm, the method is usually used in text classification.Typically in text classification, first using TF-IDF
Segmentation methods obtain the matrix with regard to term weighing adjusting the weight of word, then classify using NB Algorithm.
For non-text data, the method is less suitable for.
Chinese patent application 201510027957.X discloses a kind of " user base category based on smart mobile phone data on flows
Property Forecasting Methodology ", by the data on flows for analyzing user's smart mobile phone, predict age and sex etc. of user.Its method is point
The accurate service condition of flow of all APP on analysis user mobile phone, analyzes use feature of the user to each APP flow, and calculates
Individual features value.Using all APP traffic characteristics values as characteristic vector, by ID by the primary attribute and feature of user to
Amount is associated, and analyzes the relation between smart mobile phone data on flows and user base attribute, so as to reach according to smart mobile phone
Data on flows predict user base attribute purpose.Predicted using classification of the SVM model realizations to the primary attribute of user.Should
Method is due to the accurate service condition of the flow that analyze all APP on user mobile phone, and analytical data amount is excessive, by all APP streams
Measure feature value can cause variable excessive as characteristic vector, and so as to cause feature unobvious, the accuracy of analyses and prediction is low.
The content of the invention
Present invention aims to the problems of prior art, there is provided a kind of simple, accuracy rate is high
User's gender prediction's method based on surfing Internet with cell phone behavior.
Technical scheme is as follows:A kind of user's gender prediction's method based on surfing Internet with cell phone behavior, including it is as follows
Step:
(1) counting user clicks on the number of times of each APP within a period of time;
(2) statistical data is organized into into matrix form;
(3) dimension-reduction treatment is carried out to the matrix;
(4) data after process are divided into into training dataset and test data set, prediction mould is trained with training dataset
Type;
(5) forecast model, and accuracy in computation are verified with test data set.
Further, the user's gender prediction's method based on surfing Internet with cell phone behavior as above, matrix described in step (2)
Row represent each user record, row represent user use the corresponding number of times of each APP.
Further, the user's gender prediction's method based on surfing Internet with cell phone behavior as above, in step (3), due to
The matrix is sparse matrix, first deletes APP of the miss rate of data more than 99%, then the matrix is carried out at dimensionality reduction again
Reason;Dimension-reduction treatment adopts PCA.
Further, the user's gender prediction's method based on surfing Internet with cell phone behavior as above, in step (4) instruction is being divided
When practicing data set and test data set, take the method for random division and divide a certain proportion of data for user's different sexes
As training dataset, to avoid some attributes because random division does not have test data.
Further, the data conduct for having 80% in masculinity and femininity data respectively is ensured in random division data procedures
Training dataset, 20% data are used as test data set.
Further, the user's gender prediction's method based on surfing Internet with cell phone behavior as above, adopts in step (4)
RandomForest random forests algorithms are setting up forecast model;The random forests algorithm is one and includes multiple decision trees
Assembled classifier, the classification of its output is determined by the mode of the classification of multiple tree outputs;In modeling process, by not
Disconnected adjustment algorithm parameter (quantity as adjusted CART trees) is improving the accuracy of model prediction result.
Further, the user's gender prediction's method based on surfing Internet with cell phone behavior as above, the standard described in step (5)
Exactness can be represented that the accuracy rate is defined as correctly predicted number and actual prediction by accuracy rate, degree of accuracy, recall rate
The ratio of number;The degree of accuracy is defined as the correctly predicted number of the category and is predicted as the ratio of category number;Recall rate
It is defined as the ratio of the correctly predicted number of the category and category effective strength.
Beneficial effects of the present invention are as follows:One kind that the present invention is provided predicts user's base according to user mobile phone Internet data
The method of plinth attribute, solves because variable is excessive, causes the unconspicuous problem of feature, and can effectively avoid over-fitting
Phenomenon.Meanwhile, it is capable to predict the primary attribute of user for non-text data, such as sex, age.The method is simple,
And accuracy rate is higher.The number of times of the APP used according to user predicting the sex of user, to subsequently according to different sexes user's
Preference carries out the personalized service recommendation of correlation and provides support.
Description of the drawings
Fig. 1 is user gender prediction method flow diagram of the present invention based on surfing Internet with cell phone behavior.
Specific embodiment
The present invention is described in detail with reference to the accompanying drawings and examples.
The present invention proposes a kind of user's gender prediction's method of the number of times data that APP is accessed based on mobile phone end subscriber, can
Using non-text data being effectively user's internet behavior data predicting the sex of user.
The present invention have studied under mobile network environment, and user accesses the internet behavior data that APP is produced, and be dug using data
Dig algorithm to predict the sex of user.The all APP used user are recorded, when user clicks on once under mobile network
APP, then produce an internet log, and counting user clicks on the number of times of each APP within a period of time, then all users are entered
Then data are collected by the similar statistics of row, are write data as matrix form, and row represents each user point within a period of time
Hit the number of times of each APP, row correspondence each APP, because the species of APP is very more, thus the matrix is an individual sparse matrix.
The user gender prediction method based on surfing Internet with cell phone behavior described in this programme mainly includes two parts:Data are pre-
Process part and model construction and predicted portions;Wherein:
Data, including being standardized to above-mentioned sparse matrix, are then carried out by the data prediction part
Data are mainly carried out dimensionality reduction by dimensionality reduction in the stage using principal component analytical method.Reduce the redundancy of data.
The model construction and predicted portions, will carry out training pattern through the data of pretreatment as training data, obtain
For predicting the forecast model of user's sex.Then model, and accuracy in computation are verified using truthful data.In stage master
Will be using RandomForest random forests algorithms come training pattern, the algorithm is the common algorithms in data mining algorithm, category
In known technology, those skilled in the art can realize completely.Recycle the model to predict real user data, and count
Calculate exactness accurately.When model accuracy rate passes through, it is possible to use the model is predicting the sex of user.
Total method flow is as shown in figure 1, comprise the steps:
(1) number of times of each APP is clicked within a period of time according to user's IMEI number counting user;
(2) statistical data is organized into into matrix form;
(3) dimension-reduction treatment is carried out to the matrix;
(4) data after process are divided into into training dataset and test data set, prediction mould is trained with training dataset
Type;
(5) forecast model, and accuracy in computation are verified with test data set.
Embodiment
Below by taking the prediction of specific user's sex as an example, the present invention is further detailed.
(1) data preprocessing phase
This stage is mainly based on data prediction.
1st, by the recognition methodss of APP rule bases in system, the APP that user uses is identified.Analysis user's is upper
Data are cleaned by net daily record, delete unnecessary field.According to the internet log of user, counting user is each in one day
Period uses the number of times of APP.
Because the sex of user belongs to privacy information, some users are ready disclosure, some disclosures of being unwilling, so of the invention
Seek to predict the sex of relative users by user's internet log data, completion is carried out to the information.
2nd, by the data compilation after statistics into matrix form, row represents user, is classified as the access times of corresponding APP.Phase
Pass form is as follows:
PopStar disappears | QQ is interconnected | Between QQ rooms | QQ input methods | QQ synchronization assistants | ||
15 | 19 | 22 | 0 | 0 | 0 | 0 |
20 | 0 | 5 | 0 | 0 | 16 | 0 |
30 | 0 | 2 | 0 | 0 | 16 | 0 |
40 | 0 | 7 | 2 | 0 | 0 | 0 |
12 | 1 | 12 | 1 | 0 | 15 | 0 |
25 | 0 | 17 | 0 | 0 | 0 | 0 |
17 | 0 | 0 | 0 | 0 | 0 | 0 |
18 | 50 | 162 | 0 | 0 | 1 | 31 |
19 | 0 | 95 | 1 | 0 | 0 | 0 |
20 | 0 | 59 | 2 | 0 | 1 | 0 |
21 | 0 | 46 | 0 | 0 | 0 | 7 |
22 | 0 | 38 | 0 | 3 | 22 | 0 |
23 | 3 | 421 | 7 | 386 | 59 | 0 |
24 | 0 | 329 | 0 | 0 | 92 | 16 |
25 | 0 | 4 | 0 | 0 | 0 | 0 |
3rd, because the species of APP is very more, everyone APP for using also is not quite similar, generic APP, such as QQ, wechat
All can use Deng substantially people, and the number that the APP of some minorities is used is relatively fewer, therefore the data matrix of above-mentioned process
It is a sparse matrix, has more than 80% null value.Because the number of APP is excessive, APP of the miss rate more than 99% is deleted.
Although the 4, deleting some APP, the dimension of data is reduced, and the dimension of matrix is still very big, if
Data are all put model into, then the accuracy rate of model is very low, so data needed to carry out data before into model
Dimensionality reduction, the method for dimensionality reduction has a lot, such as principal component analysiss, odd value analysis, factorial analyses etc..Employ in the present embodiment
The method of principal component analysiss carries out dimensionality reduction to data.The dimension of the user's internet behavior data after Data Dimensionality Reduction is 150.Greatly
The complexity of calculating is reduced greatly.
(2) model training and forecast period
It is that the machine learning for having supervision is asked for the forecasting problem of user base attribute is defined as a classification problem
Topic.Supervised learning to be referred to and train grader using one group of known class target sample data, made by adjusting the parameter of grader
It reaches the process of required estimated performance.Here, we train mould according to the data of the sex of the user for having collected
Type.When training pattern starts, it is necessary first to which the corresponding sex of user is added to the last of the data that pass over pretreatment
Then data are divided again by string.For the division of data needs to follow following principle:
1) two parts are divided data into:Training set D1With test set D2, wherein training set accounts for 80%, and test set accounts for 20%.
2) for training dataset D1With test data set D2D should be met1+D2=D and D1∩D2=φ.
3) used as the training data of model, test set is used for verifying the accuracy of model training set.
4) data that ensure to there is 80% in masculinity and femininity data respectively in data procedures are randomly selected as training
Collection, 20% used as test set.
According to ready-portioned training data above come training pattern.The algorithm that the model is used is random forests algorithm,
During training pattern, constantly adjustment algorithm parameter, the quantity of CART trees, enables model more such as in adjustment grader
Good fitting data.Random forest is a grader comprising multiple decision trees, and the classification of its output is by multiple trees
Depending on the mode of the classification of output, Expired Drugs can be avoided using the algorithm.
The accuracy of prediction is the basic index of classification of assessment algorithm, to a certain extent can measure algorithm classification
Energy.User base attribute forecast problem is classification problem in the present invention, and classification accuracy index the most frequently used at present has accuracy rate
(Accuracy), degree of accuracy (Precision), recall rate (Recall).For two categorizing systems, the situation of prediction
Have 4 kinds, by taking the sex primary attribute of user as an example, i.e., user be male and predict user's sex for man, user be male but
It is to predict that user's sex is female, user's sex still predicts that user's sex is man for female, and user's sex is that female and prediction are somebody's turn to do
User's sex is female.Table 1-1 has made a summary, wherein f to this 4 kinds of situations++,f+-,f-+,f--This 4 kinds of situations are represented respectively
Number.M represents the quantity of male user in test set, and F represents that test data concentrates the quantity of female user, it is clear that M=f+++
f+-, F=f-++f--。
The confusion matrix of table 1-1 user's gender attribute forecasting problem
For the primary attribute of user, the accuracy rate (Accuracy) of its prediction is defined as correctly predicted number and reality
The ratio of prediction number.The accuracy rate of user gender prediction is:
For a classification of user base attribute, it is correct that its degree of accuracy (Precision) predicted is defined as the category
Prediction number and the ratio of category number is predicted as, the prediction accuracy of user's sex primary attribute male's classification is:
For a classification of user base attribute, it is correctly predicted that its recall rate (Recall) predicted is defined as the category
Number and category effective strength ratio, the prediction recall rate computing formula of user's sex primary attribute male's classification is:
The present invention is using accuracy rate come evaluation model.
It is analyzed according to the intraday internet behavior data logging of 2900 Beijing mobile subscribers to having been gathered by, Jing
After crossing data cleansing, dimension-reduction treatment, carry out training pattern.80% data are randomly selected here as training set, 20% number
According to as test set.Predict the outcome as Acc=70%.
Obviously, those skilled in the art can carry out the essence of various changes and modification without deviating from the present invention to the present invention
God and scope.So, if these modifications and modification to the present invention belong to the model of the claims in the present invention and its equivalent technology
Within enclosing, then the present invention is also intended to comprising these changes and modification.
Claims (10)
1. a kind of user's gender prediction's method based on surfing Internet with cell phone behavior, comprises the steps:
(1) counting user clicks on the number of times of each APP within a period of time;
(2) statistical data is organized into into matrix form;
(3) dimension-reduction treatment is carried out to the matrix;
(4) data after process are divided into into training dataset and test data set, with training dataset forecast model is trained;
(5) forecast model, and accuracy in computation are verified with test data set.
2. user's gender prediction's method of surfing Internet with cell phone behavior is based on as claimed in claim 1, it is characterised in that:Step (2)
Described in matrix row represent each user record, row represent user use the corresponding number of times of each APP.
3. user's gender prediction's method of surfing Internet with cell phone behavior is based on as claimed in claim 1 or 2, it is characterised in that:In step
Suddenly in (3), because the matrix is sparse matrix, APP of the miss rate of data more than 99% is first deleted, then again to the square
Battle array carries out dimension-reduction treatment.
4. user's gender prediction's method of surfing Internet with cell phone behavior is based on as claimed in claim 3, it is characterised in that:Dimension-reduction treatment
Using PCA.
5. user's gender prediction's method of surfing Internet with cell phone behavior is based on as claimed in claim 1, it is characterised in that:Step (4)
In dividing training dataset and during test data set, taking the method for random division and divide necessarily for user's different sexes
The data of ratio as training dataset, to avoid some attributes because random division does not have test data.
6. user's gender prediction's method of surfing Internet with cell phone behavior is based on as claimed in claim 5, it is characterised in that:Random division
To ensure the data for having 80% in masculinity and femininity data respectively as training dataset, 20% data conduct in data procedures
Test data set.
7. the user's gender prediction's method based on surfing Internet with cell phone behavior as described in claim 1 or 5, it is characterised in that:Step
(4) forecast model is set up using random forests algorithm in;The random forests algorithm is a group comprising multiple decision trees
Grader is closed, the classification of its output is determined by the mode of the classification of multiple tree outputs.
8. user's gender prediction's method of surfing Internet with cell phone behavior is based on as claimed in claim 7, it is characterised in that:Step (4)
In modeling process, the accuracy of model prediction result is improved by continuous adjustment algorithm parameter.
9. user's gender prediction's method of surfing Internet with cell phone behavior is based on as claimed in claim 8, it is characterised in that:By adjustment
The quantity of CART trees is improving the accuracy of model prediction result in algorithm model.
10. user's gender prediction's method of surfing Internet with cell phone behavior is based on as claimed in claim 1, it is characterised in that:Step (5)
Described in accuracy can be represented that the accuracy rate is defined as correctly predicted number by accuracy rate, degree of accuracy, recall rate
With the ratio of actual prediction number;The degree of accuracy is defined as the correctly predicted number of the category and is predicted as the ratio of category number
Example;Recall rate is defined as the ratio of the correctly predicted number of the category and category effective strength.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611127122.2A CN106682686A (en) | 2016-12-09 | 2016-12-09 | User gender prediction method based on mobile phone Internet-surfing behavior |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611127122.2A CN106682686A (en) | 2016-12-09 | 2016-12-09 | User gender prediction method based on mobile phone Internet-surfing behavior |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106682686A true CN106682686A (en) | 2017-05-17 |
Family
ID=58868978
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611127122.2A Pending CN106682686A (en) | 2016-12-09 | 2016-12-09 | User gender prediction method based on mobile phone Internet-surfing behavior |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106682686A (en) |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107341200A (en) * | 2017-06-20 | 2017-11-10 | 北京微影时代科技有限公司 | Judge viewing user property method for distinguishing and film recommendation method |
CN107704547A (en) * | 2017-09-26 | 2018-02-16 | 硕诺科技(深圳)有限公司 | One kind passes through mobile phone usage behavior identity method for distinguishing |
CN107886366A (en) * | 2017-11-22 | 2018-04-06 | 深圳市金立通信设备有限公司 | Generation method, sex fill method, terminal and the storage medium of Gender Classification model |
CN108564220A (en) * | 2018-04-19 | 2018-09-21 | 广州优视网络科技有限公司 | User gender prediction method, apparatus, storage medium and computer equipment |
CN108596241A (en) * | 2018-04-20 | 2018-09-28 | 西北工业大学 | User gender rapid classification method based on multidimensional perception data |
CN109145932A (en) * | 2017-06-28 | 2019-01-04 | 中兴通讯股份有限公司 | User's gender prediction's method, device and equipment |
WO2019120019A1 (en) * | 2017-12-20 | 2019-06-27 | Oppo广东移动通信有限公司 | User gender prediction method and apparatus, storage medium and electronic device |
CN110134722A (en) * | 2019-05-22 | 2019-08-16 | 北京小度信息科技有限公司 | Target user determines method, apparatus, equipment and storage medium |
CN110570254A (en) * | 2019-09-12 | 2019-12-13 | 名创优品(横琴)企业管理有限公司 | Commodity structure control method and system based on face recognition data |
WO2020082855A1 (en) * | 2018-10-25 | 2020-04-30 | 阿里巴巴集团控股有限公司 | Service recommendation method, apparatus and device |
CN111178983A (en) * | 2020-01-03 | 2020-05-19 | 北京搜狐新媒体信息技术有限公司 | User gender prediction method, device, equipment and storage medium |
CN111510422A (en) * | 2020-01-09 | 2020-08-07 | 中国石油大学(华东) | Identity authentication method based on terminal information extension sequence and random forest model |
CN112765479A (en) * | 2021-04-06 | 2021-05-07 | 北京神州泰岳智能数据技术有限公司 | Information recommendation method and device, electronic equipment and readable storage medium |
CN113268654A (en) * | 2020-02-17 | 2021-08-17 | 北京搜狗科技发展有限公司 | User gender identification method and device and electronic equipment |
CN113657917A (en) * | 2020-05-12 | 2021-11-16 | 上海佳投互联网技术集团有限公司 | Visitor gender analysis method and system based on USER-AGENT |
US11468536B2 (en) | 2018-05-18 | 2022-10-11 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for recommending a personalized pick-up location |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102279851A (en) * | 2010-06-12 | 2011-12-14 | 阿里巴巴集团控股有限公司 | Intelligent navigation method, device and system |
CN102289447A (en) * | 2011-06-16 | 2011-12-21 | 北京亿赞普网络技术有限公司 | Website webpage evaluation system based on communication network message |
CN103020087A (en) * | 2011-09-26 | 2013-04-03 | 百度在线网络技术(北京)有限公司 | Method and device for generating clicking journals, as well as method and device for adjusting search results |
CN104090886A (en) * | 2013-12-09 | 2014-10-08 | 深圳市腾讯计算机系统有限公司 | Method and device for constructing real-time portrayal of user |
CN104281634A (en) * | 2014-03-13 | 2015-01-14 | 电子科技大学 | Neighborhood-based mobile subscriber basic attribute forecasting method |
CN105224681A (en) * | 2015-10-23 | 2016-01-06 | 武汉大学 | Based on the customer requirement retrieval method and system of family's place of working context environmental |
US20160048887A1 (en) * | 2014-08-18 | 2016-02-18 | Fuji Xerox Co., Ltd. | Systems and methods for gaining knowledge about aspects of social life of a person using visual content associated with that person |
CN105408894A (en) * | 2014-06-25 | 2016-03-16 | 华为技术有限公司 | Method and device for determining user identity category |
CN106126597A (en) * | 2016-06-20 | 2016-11-16 | 乐视控股(北京)有限公司 | User property Forecasting Methodology and device |
-
2016
- 2016-12-09 CN CN201611127122.2A patent/CN106682686A/en active Pending
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102279851A (en) * | 2010-06-12 | 2011-12-14 | 阿里巴巴集团控股有限公司 | Intelligent navigation method, device and system |
CN102289447A (en) * | 2011-06-16 | 2011-12-21 | 北京亿赞普网络技术有限公司 | Website webpage evaluation system based on communication network message |
CN103020087A (en) * | 2011-09-26 | 2013-04-03 | 百度在线网络技术(北京)有限公司 | Method and device for generating clicking journals, as well as method and device for adjusting search results |
CN104090886A (en) * | 2013-12-09 | 2014-10-08 | 深圳市腾讯计算机系统有限公司 | Method and device for constructing real-time portrayal of user |
CN104281634A (en) * | 2014-03-13 | 2015-01-14 | 电子科技大学 | Neighborhood-based mobile subscriber basic attribute forecasting method |
CN105408894A (en) * | 2014-06-25 | 2016-03-16 | 华为技术有限公司 | Method and device for determining user identity category |
US20160048887A1 (en) * | 2014-08-18 | 2016-02-18 | Fuji Xerox Co., Ltd. | Systems and methods for gaining knowledge about aspects of social life of a person using visual content associated with that person |
CN105224681A (en) * | 2015-10-23 | 2016-01-06 | 武汉大学 | Based on the customer requirement retrieval method and system of family's place of working context environmental |
CN106126597A (en) * | 2016-06-20 | 2016-11-16 | 乐视控股(北京)有限公司 | User property Forecasting Methodology and device |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107341200A (en) * | 2017-06-20 | 2017-11-10 | 北京微影时代科技有限公司 | Judge viewing user property method for distinguishing and film recommendation method |
CN109145932A (en) * | 2017-06-28 | 2019-01-04 | 中兴通讯股份有限公司 | User's gender prediction's method, device and equipment |
CN107704547A (en) * | 2017-09-26 | 2018-02-16 | 硕诺科技(深圳)有限公司 | One kind passes through mobile phone usage behavior identity method for distinguishing |
CN107704547B (en) * | 2017-09-26 | 2022-01-14 | 英望科技(山东)有限公司 | Method for identifying gender through mobile phone using behaviors |
CN107886366A (en) * | 2017-11-22 | 2018-04-06 | 深圳市金立通信设备有限公司 | Generation method, sex fill method, terminal and the storage medium of Gender Classification model |
WO2019120019A1 (en) * | 2017-12-20 | 2019-06-27 | Oppo广东移动通信有限公司 | User gender prediction method and apparatus, storage medium and electronic device |
CN109948633A (en) * | 2017-12-20 | 2019-06-28 | 广东欧珀移动通信有限公司 | User gender prediction method, apparatus, storage medium and electronic equipment |
CN108564220A (en) * | 2018-04-19 | 2018-09-21 | 广州优视网络科技有限公司 | User gender prediction method, apparatus, storage medium and computer equipment |
CN108596241A (en) * | 2018-04-20 | 2018-09-28 | 西北工业大学 | User gender rapid classification method based on multidimensional perception data |
US11468536B2 (en) | 2018-05-18 | 2022-10-11 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for recommending a personalized pick-up location |
WO2020082855A1 (en) * | 2018-10-25 | 2020-04-30 | 阿里巴巴集团控股有限公司 | Service recommendation method, apparatus and device |
CN110134722A (en) * | 2019-05-22 | 2019-08-16 | 北京小度信息科技有限公司 | Target user determines method, apparatus, equipment and storage medium |
CN110570254A (en) * | 2019-09-12 | 2019-12-13 | 名创优品(横琴)企业管理有限公司 | Commodity structure control method and system based on face recognition data |
CN110570254B (en) * | 2019-09-12 | 2023-04-07 | 创优数字科技(广东)有限公司 | Commodity structure control method and system based on face recognition data |
CN111178983A (en) * | 2020-01-03 | 2020-05-19 | 北京搜狐新媒体信息技术有限公司 | User gender prediction method, device, equipment and storage medium |
CN111178983B (en) * | 2020-01-03 | 2024-03-12 | 北京搜狐新媒体信息技术有限公司 | User gender prediction method, device, equipment and storage medium |
CN111510422A (en) * | 2020-01-09 | 2020-08-07 | 中国石油大学(华东) | Identity authentication method based on terminal information extension sequence and random forest model |
CN111510422B (en) * | 2020-01-09 | 2021-07-09 | 中国石油大学(华东) | Identity authentication method based on terminal information extension sequence and random forest model |
CN113268654A (en) * | 2020-02-17 | 2021-08-17 | 北京搜狗科技发展有限公司 | User gender identification method and device and electronic equipment |
CN113657917A (en) * | 2020-05-12 | 2021-11-16 | 上海佳投互联网技术集团有限公司 | Visitor gender analysis method and system based on USER-AGENT |
CN112765479A (en) * | 2021-04-06 | 2021-05-07 | 北京神州泰岳智能数据技术有限公司 | Information recommendation method and device, electronic equipment and readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106682686A (en) | User gender prediction method based on mobile phone Internet-surfing behavior | |
Di Clemente et al. | Sequences of purchases in credit card data reveal lifestyles in urban populations | |
CN105224699B (en) | News recommendation method and device | |
CN107862022B (en) | Culture resource recommendation system | |
KR101871747B1 (en) | Similarity tendency based user-sightseeing recommendation system and method thereof | |
CN104573054B (en) | A kind of information-pushing method and equipment | |
WO2018223719A1 (en) | Method for predicting insurance purchasing behavior of a user, device, computing apparatus, and medium | |
CN106168953B (en) | Bo-Weak-relationship social network-oriented blog recommendation method | |
CN103235824A (en) | Method and system for determining web page texts users interested in according to browsed web pages | |
CN110110225B (en) | Online education recommendation model based on user behavior data analysis and construction method | |
CN111177559B (en) | Text travel service recommendation method and device, electronic equipment and storage medium | |
CN103246725A (en) | Wireless network based data traffic pushing system and method | |
CN103177129B (en) | Internet real-time information recommendation prognoses system | |
CN107908616B (en) | Method and device for predicting trend words | |
CN107291755B (en) | Terminal pushing method and device | |
US8838435B2 (en) | Communication processing | |
CN104281634B (en) | A kind of mobile subscriber's primary attribute Forecasting Methodology based on neighborhood | |
CN107368499B (en) | Client label modeling and recommending method and device | |
CN111447575A (en) | Short message pushing method, device, equipment and storage medium | |
CN115098650A (en) | Comment information analysis method based on historical data model and related device | |
CN103854206A (en) | Method and device for analyzing group characteristics | |
CN105389714B (en) | Method for identifying user characteristics from behavior data | |
Ozdikis et al. | Spatial statistics of term co-occurrences for location prediction of tweets | |
WO2019242453A1 (en) | Information processing method and device, storage medium, and electronic device | |
Li et al. | Modeling and analysis of group user portrait through WeChat mini program |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170517 |