CN103530540A - User identity attribute detection method based on man-machine interaction behavior characteristics - Google Patents
User identity attribute detection method based on man-machine interaction behavior characteristics Download PDFInfo
- Publication number
- CN103530540A CN103530540A CN201310454565.2A CN201310454565A CN103530540A CN 103530540 A CN103530540 A CN 103530540A CN 201310454565 A CN201310454565 A CN 201310454565A CN 103530540 A CN103530540 A CN 103530540A
- Authority
- CN
- China
- Prior art keywords
- man
- user
- attribute
- machine interaction
- identity attribute
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/32—User authentication using biometric data, e.g. fingerprints, iris scans or voiceprints
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2221/00—Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/21—Indexing scheme relating to G06F21/00 and subgroups addressing additional information or applications relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/2133—Verifying human interaction, e.g., Captcha
Landscapes
- Engineering & Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a user identity attribute detection method based on man-machine interaction behavior characteristics. The man-machine interaction behavior characteristics are extracted through analysis of man-machine interaction behaviors generated when man-machine interaction equipment, such as a mouse, a keyboard and a touch screen, operates in an interaction process between a user and an intelligent computer system, a user identity attribute template is established based on the man-machine interaction behavior characteristics, and identity attributes comprising genders, ages, races and the like of the user are detected and distinguished. The user identity attribute detection method has the advantages that the man-machine interaction behavior characteristics fill up blanks in analysis of identity attributes of operators in the intelligent computer system, and a brand new thought is provided for information perception analysis of users of computers and mobile networks. In addition, the user can be continuously analyzed through the method in the interaction process between the user and the intelligent computer system, and normal behaviors of the user cannot be disturbed.
Description
Technical field
The present invention relates to a kind of computing machine and mobile network user information Perception analytical technology, particularly a kind of identity attribute detection method based on computing machine and smart phone user man-machine interaction behavioural characteristic.
Background technology
Propelling along with social informatization, networking spring tide becomes more and more important to the perception analysis of user profile in computing machine and mobile network.On the one hand, in urgent hope of network virtualization economic activity Zhong， businessman such as ecommerce, the Internet banks, can try one's best and understand fully client, thereby to provide commodity targetedly or service to improve the success ratio of business activity; On the other hand, computer network and mobile network information criminal activity are also more and more serious, and the identity attribute such as sex, age, race, language of extracting and analyze the electronic evidence that is present in computational grid system and then definite operator can provide important help for discovery and the containment of network crime activity.
In recent years, there is researchist to propose to detect based on biological characteristic user's information or identity attribute, they detect information such as user's sex, age, races according to physiological characteristics such as people's face, fingerprint, iris, palmmprints, but these class methods need to be used specific biomedical information acquisition equipment, as camera, fingerprint sensor etc., be not suitable for existing computational grid environment.Also do not have at present can be in existing computational grid environment technology or the method for the analyzing and testing user identity attribute of large-scale application.
For the demand, the present invention proposes a kind of technology or method of carrying out analyzing and testing user identity attribute based on man-machine interaction behavioural characteristic.
Summary of the invention
The object of this invention is to provide a kind of computing machine and smart phone user identity attribute detection technique based on man-machine interaction behavioural characteristic, the interbehavior feature of particularly utilizing user to operate to produce in human-computer interaction device's process is as the method according to detecting operator's identity attribute.
For reaching above object, the present invention takes following technical scheme to be achieved:
A user identity detection of attribute method for man-machine interaction behavioural characteristic, is characterized in that, comprises and sets up identity attribute model and detect two parts of identity attribute:
(1) set up identity attribute model, comprise the steps:
The first step, is normally used in human-computer interaction device's process at computing machine and smart phone user, gathers the also man-machine interaction behavioral data of recording user, comprises mouse interbehavior data, keystroke interbehavior data, touches interbehavior data;
Second step, for all computing machines and smart phone user, the kind of definition identity attribute, and the division classification of every kind of identity attribute;
The 3rd step, take fixed observer time span T as the cycle to record man-machine interaction behavioral data divide, forming a plurality of time spans is the man-machine interaction behavioral data piece of T; According to the division of the kind of user's identity attribute and every kind of identity attribute, these behavioral data pieces are carried out to mark;
The 4th step, for the data block of each mark, extracts and mark man-machine interaction behavioural characteristic vector, the man-machine interaction behavioural characteristic vector in different pieces of information piece is combined to form to user's identity attribute proper vector training set;
The 5th step, for every kind of identity attribute, according to the mark of identity attribute kind, obtains each identity attribute characteristic of correspondence vector training set; According to each identity attribute, divide the mark of classification, obtain the mark of each proper vector in training set; Using each identity attribute characteristic of correspondence vector training set as training sample, the mark using the mark of proper vector in training set as training sample, builds respectively identity attribute model to each identity attribute simultaneously respectively.
(2) detect identity attribute, comprise the steps:
The first step, user logins after computing machine or smart mobile phone, catches active user's man-machine interaction behavior, take length T as the cycle, obtain the interior user's man-machine interaction behavioral data of T and extract corresponding man-machine interaction behavioural characteristic vector, and then generating the proper vector of corresponding each identity attribute;
Second step, detects active user's identity attribute: the input using the identity attribute proper vector of generation as the identity attribute presumption model of having set up, obtain the detected value of user identity attribute, user's identity attribute is judged.
In said method, the described concrete steps of setting up the identity attribute proper vector training set that forms user in identity attribute model part the 4th step are as follows:
(1) in the man-machine interaction behavioral data piece that is T at observation interval, traversal man-machine interaction sequence of events, isolates dissimilar interbehavior event successively, comprises mouse interbehavior event, keystroke interbehavior event, touches interbehavior event;
(2) for dissimilar interbehavior event, extract interbehavior proper vector, comprise mouse behavioural characteristic vector, keystroke behavioural characteristic vector, touch behavioural characteristic vector;
(3) the man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms identity attribute proper vector training set.
The man-machine interaction behavioral data that described computing machine or smart phone user produce is the sequence that basic man-machine interaction event forms, the form of basic man-machine interaction event is: { interaction time stamp, interactive screen position, the interactive device type that comprises mouse, keyboard or touch pad, alternative events type }.
Described identity attribute refers to computing machine and intrinsic physiology or the behavioral trait of smart phone user, comprises that user's sex, age, race, language, right-hand man's use habit, schooling, computing machine uses skill level, occupation, finger health status.
The described identity attribute model of setting up is realized by one or more Classifier combinations, and described sorter comprises Weighted random forest classified device, artificial nerve network classifier, support vector machine classifier.Wherein, the concrete steps of setting up identity attribute model by Weighted random forest classified device are:
1) number of the concentrated feature samples of initialization training sample is N, and in each feature samples, the number of characteristic component is M, and the number of decision tree is P, and the number of the decision-making feature of each decision tree is m, and m is much smaller than M;
2) in order to eliminate the impact of different feature dimensions, each dimensional feature vector is normalized, its value is limited between [0,1];
3) use Bagging algorithm to N feature samples sampling P time, obtain P characteristic set;
4) each random tree is chosen to a characteristic set at random, and this decision tree is assessed and error analysis, for each node in tree, the random individual characteristic component based on this point of m of selecting, and for different classes of feature samples, give different weights to find best partitioning scheme;
5) according to the best characteristic node of classifying quality by node division Wei Liangge branch, then recursive call step 4) is until the accurate classification based training sample set of this tree, or all properties was all used; After the complete growth of decision tree, it is not carried out to beta pruning;
6) repeating step 3), 4), 5) until set up whole P decision tree;
7) method that adopts the majority based on weighting to vote comprehensively determines the classification results of a plurality of decision trees, obtains the classification results of Weighted random forest classified device.
The behavioral trait that the present invention embodies in interactive process with the formal description user of man-machine interaction sequence of events, detects operator's identity attribute with this, for computing machine and the analysis of mobile network user information Perception provide a kind of brand-new thinking.Its advantage is: first, identity attribute is analyzed desired data and can from interactive process, directly be obtained, without being equipped with extra instrument and equipment; Secondly, identity attribute analysis is based on man-machine interaction behavioural characteristic, without remembering or carrying, is difficult to imitate and forge; In addition, in the process of computer user and smart phone user operating equipment, can continue to catch the human-machine interactive information that user's operation produces, therefore can continue user identity attribute to carry out discriminatory analysis based on man-machine interaction behavioural characteristic, and normal behaviour that can interference user, there is security widely and applicability.
Accompanying drawing explanation
Below in conjunction with the drawings and specific embodiments, the present invention is described in further detail.
Fig. 1 is the step block diagram of the inventive method.
Fig. 2 is that the identity attribute feature of man-machine interaction behavior in the inventive method generates step block diagram.
Fig. 3 is the identity attribute method for establishing model step block diagram based on Weighted random forest in the inventive method.
Fig. 4 is the experimental result picture that adopts the inventive method detection computations machine user identity attribute.In figure, the wrong bar of black is illustrated in the standard deviation of 20 identity attribute accuracys rate after random data sampling.
Embodiment
System architecture
Referring to Fig. 1, the present invention is based on computing machine and the smart phone user identity attribute detection method of man-machine interaction behavioural characteristic, can be used for user identity attribute perception in the e-commerce initiatives such as electronic emporium, the Internet bank, so that commodity or service to be targetedly provided; Also can be used for the information forensics analysis in enterprise information system, important infosystem is carried out to safeguard protection.The present invention comprises and sets up identity attribute model and identity attribute and detect two parts, and concrete implementation step is as follows:
1) setting up identity attribute model part comprises the steps:
(1) at computing machine and smart phone user, normally use in human-computer interaction device's process, gather the also man-machine interaction behavioral data of recording user, comprise mouse interbehavior data, keystroke interbehavior data, touch interbehavior data, and then formation identity attribute model is set up required interbehavior data set; The form of basic human-machine interaction data is: { interaction time stamp, interactive screen position, type of interaction, alternative events type }, type of interaction comprises that mouse is mutual, keyboard mutuality and touch screen interaction, and alternative events type comprises that mouse click and moving event, keyboard keystroke event, finger press and touch moving event on touch-screen;
(2) for all computing machines and smart phone user, the kind of definition identity attribute, and the division classification of every kind of identity attribute;
(3) take fixed observer time span T as the cycle to record man-machine interaction behavioral data divide, forming a plurality of time spans is the man-machine interaction behavioral data piece of T; According to the division of the kind of user's identity attribute and every kind of identity attribute, these behavioral data pieces are carried out to mark;
(4) for the data block of each mark, extract and mark man-machine interaction behavioural characteristic vector, comprise mouse behavioural characteristic vector, keystroke behavioural characteristic vector, touch behavioural characteristic vector.Wherein mouse behavioural characteristic vector refers to by mouse and moves a series of behavior measure amounts that the operations such as the space-time geometric locus of generation and mouse click derive, can use once mobile geometric locus, the relation of rate curve, accelerating curve, average translational speed and distance, the relation of average translational speed and direction, the relation of average translational acceleration and distance, the relation of average translational acceleration and direction, motion track apart from the ratio of displacement as feature; Keystroke behavioural characteristic vector refers to by the derivative a series of behavior measure amounts that obtain of time series that each key of keyboard is pressed and the event of upspringing forms, can use the duration of singly-bound button and the interval time of adjacent key as feature; Touch behavioural characteristic vector and refer to finger mobile space-time geometric locus producing and press a series of behavior measure amounts that wait operation to derive on touch-screen, can use screen touch pressure, touch the click time, touch motion track, touch translational speed curve, touch traveling time be as feature.Man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms user's identity attribute proper vector training set;
(5) for each identity attribute (user's sex, age, race, language, right-hand man's use habit, schooling, computing machine use the attributes such as skill level, occupation, finger health status), according to the mark of identity attribute kind, obtain each identity attribute characteristic of correspondence vector training set; According to each identity attribute, divide the mark of classification, obtain the mark of each proper vector in training set; Using each identity attribute characteristic of correspondence vector training set as training sample, the mark using the mark of proper vector in training set as training sample, builds respectively the identity attribute detection model based on Weighted random forest to each identity attribute simultaneously respectively.User's the sex detection of attribute model of take is example, training data using the proper vector training set with Sex-linked marker as model, the detection of sex attribute is considered as to two classification problems (man or female), thereby builds the sex detection of attribute model based on man-machine interaction behavior.
2) identity attribute test section comprises the steps:
(1) user, use in the process of the intelligent systems such as computing machine or smart mobile phone, catch active user's man-machine interaction behavior, the length T (T generally can be made as 30 seconds or the longer time) of take is the cycle, obtain the interior user's human-machine interaction data of T and extract behavioural characteristic, generating identity attribute proper vector;
(2) input using the identity attribute proper vector generating as identity attribute detection model, obtain the detected value of user identity attribute, by the threshold epsilon of this detected value and corresponding identity attribute model, (ε chooses according to the precision of model training, generally can be set as 50%) compare, differentiate the corresponding identity attribute of user.User's the sex detection of attribute of take is example, input using other identity attribute vector of the correspondence of extracting in the time cycle from T as the sex detection of attribute model of having set up, obtain the detected value of this model, detected value is compared with corresponding threshold value, if detected value is greater than threshold value, the sex of judging active user is the male sex; If detected value is less than threshold value, the sex of judging active user is women.
Identity attribute detection model based on Weighted random forest
Above-mentioned 1) set up that in (5) step of identity attribute model part, the identity attribute detection model process of establishing based on Weighted random forest is referring to Fig. 3, concrete steps are as follows:
(1) it is N that initialization training characteristics is concentrated the number of feature samples, and in each feature samples, the number of characteristic component is M, and the number of decision tree is P, and the number of the decision-making feature of each decision tree is m (m is much smaller than M);
(2) in order to eliminate the impact of different feature dimensions, each dimensional feature vector is normalized, its value is limited between [0,1];
(3) use Bagging algorithm to N feature samples sampling P time, obtain P characteristic set;
(4) each random tree is chosen to a characteristic set at random, and this decision tree is assessed and error analysis.For each node in tree, select at random m the characteristic component based on this point, and for different classes of feature samples, give different weights to find best partitioning scheme;
(5) according to the best characteristic node of classifying quality by node division Wei Liangge branch, then recursive call step (4) is until the accurate classification based training sample set of this tree, or all properties was all used; After the complete growth of decision tree, it is not carried out to beta pruning;
(6) repeating step (3), (4), (5) are until set up whole P decision tree;
(7) method that adopts the majority based on weighting to vote comprehensively determines the classification results of a plurality of decision trees, has obtained the classification results of Weighted random forest classified device.
The description of the partitioning scheme of the system of selection of decision-making characteristic variable number, the best
In (1) step of " the identity attribute detection model based on Weighted random forest ", the selection of decision-making characteristic variable number m refers to that every decision tree of structure be from feature samples, to choose at random m dimensional feature, and chooses the best characteristic node of classifying quality in this m dimensional feature.In the construction process of whole random forest, m is a constant, and we choose m=int (log
2m+1), wherein int is bracket function.
In (4) step, best partitioning scheme is to instigate grouped data on each node as far as possible from same classification, thereby the impurity level that makes each node reaches minimum partitioning scheme (grouped data in certain node i all comes from same classification, and the impurity level of this node is 0).In the process of every decision tree structure, top-down recurrence division principle is followed in its generation, from root node, starts successively training set to be divided.For each node, according to node impurity level minimum principle, be split into left sibling and have node, they comprise respectively a subset of training data, make node continue division, until branch stops growing according to same rule.If the grouped data in node i all comes from same classification, the impurity level I of this node (i)=0.The measure of impurity level, based on Gini impurity level criterion, is supposed P (w
j) be in node i, to belong to w
jclass number of samples accounts for the frequency of training sample sum, and Gini impurity level criterion is expressed as:
Decision-making technique based on the most ballots of weighting
The method of the majority ballot based on weighting in (7) step of " the identity attribute detection model based on Weighted random forest " refers to gives larger weights to the few classification of feature samples number.In identity attribute testing process, the detection of sex information of take is example (2 class classification problems: man or female), and a sample x is through each decision tree classifier T
iafter, will produce 2 Output rusults, be 2 confidence values, c ∈ { 1,2}, each degree of confidence
p(f (x)=
j) having represented that this sample x belongs to the probable value of j class, the weighted value of final judgement based on all decision tree results, is shown below.
Weights α wherein
icircular for such other votes, be multiplied by the number of times for such other repeated sampling, and it is normalized.
Finally, the decision value obtaining and decision-making value ε are compared, user's identity attribute is differentiated.If F>=ε, active user's sex is the male sex; If F<ε, active user's sex is women.Wherein, choosing of ε can adopt cross validation when model training, adjusted and optimized, to obtain good model training result by the value of continuous variation ε.
According to the present invention, detect the step of sex attribute
The first step, the classification of definition sex attribute, is divided into 2 classes by sex attribute: the first kind is male user in the present embodiment; Equations of The Second Kind is women user.
Second step, catches man-machine interaction behavior training data.Take fixed observer time span T as the cycle to record man-machine interaction behavioral data divide, forming a plurality of time spans is the man-machine interaction behavioral data piece of T, and according to user's sex attribute, these behavioral data pieces is carried out to mark.
The 3rd step, the features training collection of generation sex attribute.For the data block of each mark, extract and mark man-machine interaction behavioural characteristic vector; Man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms the sex attribute feature vector training set with sex attribute flags; Each proper vector in this sex attributive character training set can, for whole man-machine interaction behavioural characteristic vector, can be also the subvector of the man-machine interaction behavioural characteristic vector after feature selecting.
The 4th step, sets up sex detection of attribute model.The test problems of sex attribute is considered as to the problem of 2 classification, using sex attribute characteristic of correspondence vector training set as training sample, using the Sex-linked marker of each proper vector as the mark of training sample simultaneously, sex attribute is built to the identity attribute detection model based on Weighted random forest.
The 5th step, the real time human-machine interaction behavioral data of supervisory user also extracts behavioural characteristic vector.When user accesses computing machine or smart mobile phone, with observation time T, catch new man-machine interaction behavioral data; The human-machine interaction data that is T from time span, extract sex attribute characteristic of correspondence vector.
The 6th step, the testing result of generation sex attribute.Using the vectorial input as the sex detection of attribute model of having set up of the sex attribute characteristic of correspondence of generation, obtain the detected value of user's sex attribute; This detected value and decision-making value are compared, user's sex attribute is judged.
According to the present invention, detect the step of age attribute
The first step, the classification of definition age attribute, is divided into 3 classes by age attribute: the first kind is the user who the age is less than 30 years old in the present embodiment; Equations of The Second Kind is the user of age between 30 years old to 60 years old; The 3rd class is to be greater than the user of 60 years old.
Second step, catches man-machine interaction behavior training data.Take fixed observer time span T as the cycle to record man-machine interaction behavioral data divide, forming a plurality of time spans is the man-machine interaction behavioral data piece of T, and according to user's age attribute, these behavioral data pieces is carried out to mark.
The 3rd step, the features training collection of generation age attribute.For the data block of each mark, extract and mark man-machine interaction behavioural characteristic vector; Man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms the age attribute feature vector training set with has age attribute flags; Each proper vector in this age attributive character training set can, for whole man-machine interaction behavioural characteristic vector, can be also the subvector of the man-machine interaction behavioural characteristic vector after feature selecting.
The 4th step, sets up age detection of attribute model.The test problems of age attribute is considered as to the problem of 3 classification, using age attribute characteristic of correspondence vector training set as training sample, using the age indicator of each proper vector as the mark of training sample simultaneously, age attribute is built to the identity attribute detection model based on Weighted random forest.
The 5th step, the real time human-machine interaction behavioral data of supervisory user also extracts behavioural characteristic vector.When user accesses computing machine or smart mobile phone, with observation time T, catch new man-machine interaction behavioral data; The human-machine interaction data that is T from time span, extract age attribute characteristic of correspondence vector.
The 6th step, the testing result of generation age attribute.Using the vectorial input as the age detection of attribute model of having set up of the age attribute characteristic of correspondence of generation, obtain the detected value of age of user attribute; This detected value and decision-making value are compared, user's age attribute is judged.
The step of detection language attribute according to the present invention
The first step, the classification of definitional language attribute, is divided into linguistic property 2 classes in the present embodiment: the first kind is that English is the user of mother tongue; Equations of The Second Kind is that non-English is the user of mother tongue.
Second step, catches man-machine interaction behavior training data.Take fixed observer time span T as the cycle to record man-machine interaction behavioral data divide, forming a plurality of time spans is the man-machine interaction behavioral data piece of T, and according to user's linguistic property, these behavioral data pieces is carried out to mark.
The 3rd step, the features training collection of production language attribute.For the data block of each mark, extract and mark man-machine interaction behavioural characteristic vector; Man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms the linguistic property proper vector training set with linguistic property mark; Each proper vector that this linguistic property features training is concentrated can, for whole man-machine interaction behavioural characteristic vector, can be also the subvector of the man-machine interaction behavioural characteristic vector after feature selecting.
The 4th step, sets up linguistic property detection model.The test problems of linguistic property is considered as to the problem of 2 classification, using linguistic property characteristic of correspondence vector training set as training sample, using the language tag of each proper vector as the mark of training sample simultaneously, linguistic property is built to the identity attribute detection model based on Weighted random forest.
The 5th step, the real time human-machine interaction behavioral data of supervisory user also extracts behavioural characteristic vector.When user accesses computing machine or smart mobile phone, with observation time T, catch new man-machine interaction behavioral data; The human-machine interaction data that is T from time span, extract linguistic property characteristic of correspondence vector.
The 6th step, the testing result of production language attribute.Using the vectorial input as the linguistic property detection model of having set up of the linguistic property characteristic of correspondence of generation, obtain the detected value of user language attribute; This detected value and decision-making value are compared, user's linguistic property is judged.
According to the present invention, detect the step of right-hand man's use habit attribute
The first step, the classification of definition right-hand man use habit attribute, is divided into 2 classes by right-hand man's use habit attribute in the present embodiment: the first kind is for take the user of left hand as custom operation human-computer interaction device; Equations of The Second Kind is for take the right hand as being accustomed to operating the user that man-machine interaction sets.
Second step, catches man-machine interaction behavior training data.Take fixed observer time span T as the cycle to record man-machine interaction behavioral data divide, forming a plurality of time spans is the man-machine interaction behavioral data piece of T, and according to right-hand man's use habit attribute of user, these behavioral data pieces is carried out to mark.
The 3rd step, the features training collection of generation right-hand man use habit attribute.For the data block of each mark, extract and mark man-machine interaction behavioural characteristic vector; Man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms the right-hand man's use habit attribute feature vector training set with right-hand man's use habit attribute flags; Each proper vector in this right-hand man's use habit attributive character training set can, for whole man-machine interaction behavioural characteristic vector, can be also the subvector of the man-machine interaction behavioural characteristic vector after feature selecting.
The 4th step, sets up right-hand man's use habit detection of attribute model.The test problems of right-hand man's use habit attribute is considered as to the problem of 2 classification, using right-hand man's use habit attribute characteristic of correspondence vector training set as training sample, using right-hand man's use habit mark of each proper vector as the mark of training sample simultaneously, right-hand man's use habit attribute is built to the identity attribute detection model based on Weighted random forest.
The 5th step, the real time human-machine interaction behavioral data of supervisory user also extracts behavioural characteristic vector.When user accesses computing machine or smart mobile phone, with observation time T, catch new man-machine interaction behavioral data; The human-machine interaction data that is T from time span, extract right-hand man's use habit attribute characteristic of correspondence vector.
The 6th step, the testing result of generation right-hand man use habit attribute.Using the vectorial input as right-hand man's use habit detection of attribute model of having set up of right-hand man's use habit attribute characteristic of correspondence of generation, obtain the detected value of user right-hand man's use habit attribute; This detected value and decision-making value are compared, right-hand man's use habit attribute of user is judged.
According to the present invention, detect the step of schooling attribute
The first step, the classification of definition schooling attribute, is divided into 3 classes by schooling attribute: the first kind is that schooling is primary school and following user in the present embodiment; Equations of The Second Kind be schooling in junior middle school the user to senior middle school; The 3rd class is that schooling is university and above user.
Second step, catches man-machine interaction behavior training data.Take fixed observer time span T as the cycle to record man-machine interaction behavioral data divide, forming a plurality of time spans is the man-machine interaction behavioral data piece of T, and according to user's schooling attribute, these behavioral data pieces is carried out to mark.
The 3rd step, the features training collection of generation schooling attribute.For the data block of each mark, extract and mark man-machine interaction behavioural characteristic vector; Man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms the schooling attribute feature vector training set with schooling attribute flags; Each proper vector in this schooling attributive character training set can, for whole man-machine interaction behavioural characteristic vector, can be also the subvector of the man-machine interaction behavioural characteristic vector after feature selecting.
The 4th step, sets up schooling detection of attribute model.The test problems of schooling attribute is considered as to the problem of 3 classification, using schooling attribute characteristic of correspondence vector training set as training sample, using the schooling mark of each proper vector as the mark of training sample simultaneously, schooling attribute is built to the identity attribute detection model based on Weighted random forest.
The 5th step, the real time human-machine interaction behavioral data of supervisory user also extracts behavioural characteristic vector.When user accesses computing machine or smart mobile phone, with observation time T, catch new man-machine interaction behavioral data; The human-machine interaction data that is T from time span, extract schooling attribute characteristic of correspondence vector.
The 6th step, the testing result of generation schooling attribute.Using the vectorial input as the schooling detection of attribute model of having set up of the schooling attribute characteristic of correspondence of generation, obtain the detected value of user's schooling attribute; This detected value and decision-making value are compared, user's schooling attribute is judged.
According to the present invention, detection computations machine uses the step of skill level attribute
The first step, definition computing machine uses the classification of skill level attribute, in the present embodiment computing machine is used skill level attribute to be divided into 3 classes: the first kind is very unskilled user (not having corresponding human-computer interaction device to use experience); Equations of The Second Kind is general skilled user's (using corresponding human-computer interaction device between 1 month to 3 months); The 3rd class is very skilled user's (using corresponding human-computer interaction device over 3 months).
Second step, catches man-machine interaction behavior training data.The fixed observer time span T of take divided the man-machine interaction behavioral data of record as the cycle, and forming a plurality of time spans is the man-machine interaction behavioral data piece of T, and uses skill level attribute to carry out mark to these behavioral data pieces according to user's computing machine.
The 3rd step, generates the features training collection that computing machine uses skill level attribute.For the data block of each mark, extract and mark man-machine interaction behavioural characteristic vector; Man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms with computing machine and use the computing machine of skill level attribute flags to use skill level attribute feature vector training set; It can, for whole man-machine interaction behavioural characteristic vector, can be also the subvector of the man-machine interaction behavioural characteristic vector after feature selecting that this computing machine uses each proper vector in skill level attributive character training set.
The 4th step, sets up computing machine and uses skill level detection of attribute model.Use the test problems of skill level attribute to be considered as the problem of 3 classification in computing machine, the computing machine of usining uses skill level attribute characteristic of correspondence vector training set as training sample, the computing machine of each proper vector of simultaneously usining uses skill level mark as the mark of training sample, to computing machine, uses skill level attribute to build the identity attribute detection model based on Weighted random forest.
The 5th step, the real time human-machine interaction behavioral data of supervisory user also extracts behavioural characteristic vector.When user accesses computing machine or smart mobile phone, with observation time T, catch new man-machine interaction behavioral data; The human-machine interaction data that is T from time span, extract computing machine and use skill level attribute characteristic of correspondence vector.
The 6th step, generates the testing result that computing machine uses skill level attribute.Use skill level attribute characteristic of correspondence vector as the input of the computing machine use skill level detection of attribute model of having set up in the computing machine of generation, obtain the detected value that subscriber computer is used skill level attribute; This detected value and decision-making value are compared, to user's computing machine, use skill level attribute to judge.
According to the present invention, detect the step of professional attribute
The first step, defines the classification of professional attribute, in the present embodiment professional attribute is divided into 2 classes: the first kind is the user of computing machine working; Equations of The Second Kind is the user of non-computer working.
Second step, catches man-machine interaction behavior training data.Take fixed observer time span T as the cycle to record man-machine interaction behavioral data divide, forming a plurality of time spans is the man-machine interaction behavioral data piece of T, and according to user's professional attribute, these behavioral data pieces is carried out to mark.
The 3rd step, generates the features training collection of professional attribute.For the data block of each mark, extract and mark man-machine interaction behavioural characteristic vector; Man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms the professional attribute feature vector training set with professional attribute flags; Each proper vector in this occupation attributive character training set can, for whole man-machine interaction behavioural characteristic vector, can be also the subvector of the man-machine interaction behavioural characteristic vector after feature selecting.
The 4th step, sets up professional detection of attribute model.The test problems of professional attribute is considered as to the problem of 2 classification, using professional attribute characteristic of correspondence vector training set as training sample, using the mark of each proper vector as the mark of training sample simultaneously, professional attribute is built to the identity attribute detection model based on Weighted random forest.
The 5th step, the real time human-machine interaction behavioral data of supervisory user also extracts behavioural characteristic vector.When user accesses computing machine or smart mobile phone, with observation time T, catch new man-machine interaction behavioral data; The human-machine interaction data that is T from time span, extract professional attribute characteristic of correspondence vector.
The 6th step, generates the testing result of professional attribute.Using the vectorial input as the professional detection of attribute model of having set up of the professional attribute characteristic of correspondence generating, obtain the detected value of user's occupation attribute; This detected value and decision-making value are compared, user's professional attribute is judged.
According to the present invention, detect the step of the healthy attribute of finger
The first step, the classification of the healthy attribute of definition finger, is divided into 2 classes by the healthy attribute of finger in the present embodiment: the first kind is the healthy user of finger; Equations of The Second Kind is the user of the non-health of finger.
Second step, catches man-machine interaction behavior training data.The fixed observer time span T of take divided the man-machine interaction behavioral data of record as the cycle, and forming a plurality of time spans is the man-machine interaction behavioral data piece of T, and according to the healthy attribute of user's finger, these behavioral data pieces is carried out to mark.
The 3rd step, generates the features training collection of pointing healthy attribute.For the data block of each mark, extract and mark man-machine interaction behavioural characteristic vector; Man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece is combined, forms the healthy attribute feature vector training set of finger with the healthy attribute flags of finger; This each proper vector of pointing in healthy attributive character training set can, for whole man-machine interaction behavioural characteristic vector, can be also the subvector of the man-machine interaction behavioural characteristic vector after feature selecting.
The 4th step, sets up the healthy detection of attribute model of finger.The test problems of the healthy attribute of finger is considered as to the problem of 2 classification, using and point healthy attribute characteristic of correspondence vector training set as training sample, using the mark of each proper vector as the mark of training sample simultaneously, to pointing healthy attribute, build the identity attribute detection model based on Weighted random forest.
The 5th step, the real time human-machine interaction behavioral data of supervisory user also extracts behavioural characteristic vector.When user accesses computing machine or smart mobile phone, with observation time T, catch new man-machine interaction behavioral data; The human-machine interaction data that is T from time span, extract the healthy attribute characteristic of correspondence vector of finger.
The 6th step, generates the testing result of pointing healthy attribute.Input using the healthy attribute characteristic of correspondence vector of the finger of generation as the healthy detection of attribute model of the finger of having set up, obtains the detected value that user points healthy attribute; This detected value and decision-making value are compared, the healthy attribute of user's finger is judged.
Explanation about identity attribute category setting in embodiment
Identity attribute category setting in an embodiment, only as a kind of embodiment of content of the present invention.The identity attribute category setting scheme that can have other in practical application.As age attribute also can be divided into 4 classes as required: the first kind is the user who the age is less than 15 years old; Equations of The Second Kind be the age at 15-30 the user between year; The 3rd class is to be greater than the 30-50 user in year, and the 4th class is to be greater than the user of 50 years old.Now only need in setting up identity attribute detection model process, use identical category setting, can use method of the present invention.
The experimental result of test section identity attribute according to the present invention
By collecting 58 users' mouse behavioral data and 51 users' keystroke behavioral data is set up man-machine interaction behavioral data collection.The method and the technology that by experiment the present invention are proposed are verified.Table 1 has been listed the testing result of identity attribute information.
The statistics that table 1. identity attribute information detects.
*: upper hurdle represents to use keyboard behavior to carry out the result of sex-screening, and lower hurdle represents to use mouse behavior to carry out the result of sex-screening.
As shown in the experimental result of table 1 and Fig. 4, the method that the present invention proposes can detect user's identity attribute information exactly.When utilizing human-machine interactive information to detect user identity attribute information, accuracy rate is all higher than 85%.When utilizing keyboard mutuality data to detect user's ethnic information, relevant accuracy rate is 87.32%.This result verification the feasibility of method proposed by the invention, show that the method can be computing machine and the analysis of mobile network user information Perception provides a kind of effective technological means.
Claims (6)
1. the user identity detection of attribute method based on man-machine interaction behavioural characteristic, is characterized in that, comprises and sets up identity attribute model and detect two parts of identity attribute:
(1) set up identity attribute model, comprise the steps:
The first step, is normally used in human-computer interaction device's process at computing machine and smart phone user, gathers the also man-machine interaction behavioral data of recording user, comprises mouse interbehavior data, keystroke interbehavior data, touches interbehavior data;
Second step, for all computing machines and smart phone user, the kind of definition identity attribute, and the division classification of every kind of identity attribute;
The 3rd step, take fixed observer time span T as the cycle to record man-machine interaction behavioral data divide, forming a plurality of time spans is the man-machine interaction behavioral data piece of T; According to the division of the kind of user's identity attribute and every kind of identity attribute, these behavioral data pieces are carried out to mark;
The 4th step, for the data block of each mark, extracts and mark man-machine interaction behavioural characteristic vector, the man-machine interaction behavioural characteristic vector in different pieces of information piece is combined to form to user's identity attribute proper vector training set;
The 5th step, for every kind of identity attribute, according to the mark of identity attribute kind, obtains each identity attribute characteristic of correspondence vector training set; According to each identity attribute, divide the mark of classification, obtain the mark of each proper vector in training set; Using each identity attribute characteristic of correspondence vector training set as training sample, the mark using the mark of proper vector in training set as training sample, builds respectively identity attribute model to each identity attribute simultaneously respectively;
(2) detect identity attribute, comprise the steps:
The first step, user logins after computing machine or smart mobile phone, catches active user's man-machine interaction behavior, take length T as the cycle, obtain the interior user's man-machine interaction behavioral data of T and extract corresponding man-machine interaction behavioural characteristic vector, and then generating the proper vector of corresponding each identity attribute;
Second step, detects active user's identity attribute: the input using the identity attribute proper vector of generation as the identity attribute presumption model of having set up, obtain the detected value of user identity attribute, user's identity attribute is judged.
2. the user identity detection of attribute method based on man-machine interaction behavioural characteristic according to claim 1, is characterized in that, the described concrete steps of setting up the identity attribute proper vector training set that forms user in identity attribute model part the 4th step are as follows:
In the man-machine interaction behavioral data piece that the first step is T at observation interval, traversal man-machine interaction sequence of events, isolates dissimilar interbehavior event successively, comprises mouse interbehavior event, keystroke interbehavior event, touches interbehavior event;
Second step, for dissimilar interbehavior event, extracts interbehavior proper vector, comprises mouse behavioural characteristic vector, keystroke behavioural characteristic vector, touches behavioural characteristic vector;
The 3rd step, is combined the man-machine interaction behavioural characteristic Vector Groups in different pieces of information piece, forms identity attribute proper vector training set.
3. the user identity detection of attribute method based on man-machine interaction behavioural characteristic according to claim 1, it is characterized in that, the man-machine interaction behavioral data that described computing machine or smart phone user produce is the sequence that basic man-machine interaction event forms, the form of basic man-machine interaction event is: { interaction time stamp, interactive screen position, the interactive device type that comprises mouse, keyboard or touch pad, alternative events type }.
4. the user identity detection of attribute method based on man-machine interaction behavioural characteristic according to claim 1, it is characterized in that, described identity attribute refers to computing machine and intrinsic physiology or the behavioral trait of smart phone user, comprises that user's sex, age, race, language, right-hand man's use habit, schooling, computing machine uses skill level, occupation, finger health status.
5. the user identity detection of attribute method based on man-machine interaction behavioural characteristic according to claim 1, it is characterized in that, the described identity attribute model of setting up is realized by one or more Classifier combinations, and described sorter comprises Weighted random forest classified device, artificial nerve network classifier, support vector machine classifier.
6. the user identity detection of attribute method based on man-machine interaction behavioural characteristic according to claim 5, is characterized in that, the concrete steps of being set up identity attribute model by Weighted random forest classified device are:
1) number of the concentrated feature samples of initialization training sample is N, and in each feature samples, the number of characteristic component is M, and the number of decision tree is P, and the number of the decision-making feature of each decision tree is m, and m is much smaller than M;
2) in order to eliminate the impact of different feature dimensions, each dimensional feature vector is normalized, its value is limited between [0,1];
3) use Bagging algorithm to N feature samples sampling P time, obtain P characteristic set;
4) each random tree is chosen to a characteristic set at random, and this decision tree is assessed and error analysis, for each node in tree, the random individual characteristic component based on this point of m of selecting, and for different classes of feature samples, give different weights to find best partitioning scheme;
5) according to the best characteristic node of classifying quality by node division Wei Liangge branch, then recursive call step 4) is until the accurate classification based training sample set of this tree, or all properties was all used; After the complete growth of decision tree, it is not carried out to beta pruning;
6) repeating step 3), 4), 5) until set up whole P decision tree;
7) method that adopts the majority based on weighting to vote comprehensively determines the classification results of a plurality of decision trees, obtains the classification results of Weighted random forest classified device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310454565.2A CN103530540B (en) | 2013-09-27 | 2013-09-27 | User identity attribute detection method based on man-machine interaction behavior characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310454565.2A CN103530540B (en) | 2013-09-27 | 2013-09-27 | User identity attribute detection method based on man-machine interaction behavior characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103530540A true CN103530540A (en) | 2014-01-22 |
| CN103530540B CN103530540B (en) | 2017-02-22 |
Family
ID=49932544
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310454565.2A Active CN103530540B (en) | 2013-09-27 | 2013-09-27 | User identity attribute detection method based on man-machine interaction behavior characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103530540B (en) |
Cited By (48)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103870547A (en) * | 2014-02-26 | 2014-06-18 | 华为技术有限公司 | Grouping processing method and device of contact persons |
| CN103927467A (en) * | 2014-04-25 | 2014-07-16 | 武汉大学 | Smart phone authentication system and method based on screen touching behavior |
| CN103942489A (en) * | 2014-03-31 | 2014-07-23 | 中国科学院信息工程研究所 | Attack detection method and system on basis of cursor hidden scene |
| CN104239761A (en) * | 2014-09-15 | 2014-12-24 | 西安交通大学 | Continuous identity authentication method based on touch screen slip behavior characteristics |
| CN104537252A (en) * | 2015-01-05 | 2015-04-22 | 深圳市腾讯计算机系统有限公司 | User state single-classification model training method and device |
| CN104778387A (en) * | 2015-04-23 | 2015-07-15 | 西安交通大学 | Cross-platform identity authentication system and method based on human-computer interaction behaviors |
| CN104809377A (en) * | 2015-04-29 | 2015-07-29 | 西安交通大学 | Method for monitoring network user identity based on webpage input behavior characteristics |
| CN104951544A (en) * | 2015-06-19 | 2015-09-30 | 百度在线网络技术（北京）有限公司 | User data processing method and system and method and system for providing user data |
| CN105279691A (en) * | 2014-07-25 | 2016-01-27 | 中国银联股份有限公司 | Financial transaction detection method and equipment based on random forest model |
| CN105306496A (en) * | 2015-12-02 | 2016-02-03 | 中国科学院软件研究所 | User identity detection method and system |
| CN105607741A (en) * | 2015-12-31 | 2016-05-25 | 联想(北京)有限公司 | Control method and electronic equipment |
| CN105760516A (en) * | 2016-02-25 | 2016-07-13 | 广州视源电子科技股份有限公司 | Method and device for distinguishing users |
| CN106095915A (en) * | 2016-06-08 | 2016-11-09 | 百度在线网络技术（北京）有限公司 | The processing method and processing device of user identity |
| CN106156591A (en) * | 2016-07-12 | 2016-11-23 | 中国科学技术大学苏州研究院 | A kind of smart phone user Transparent Authentication method under cloud environment |
| CN106371750A (en) * | 2016-08-30 | 2017-02-01 | 北京奇艺世纪科技有限公司 | Method and device for confirming user gender |
| CN106503499A (en) * | 2016-09-22 | 2017-03-15 | 天津大学 | Smart mobile phone touch-screen input recognition method based on machine learning |
| CN106503722A (en) * | 2016-12-21 | 2017-03-15 | 湖南文理学院 | Operator's skeleton growth rings method based on touch trajectory image procossing |
| CN106572097A (en) * | 2016-11-01 | 2017-04-19 | 南京邮电大学 | Mobile device-based mixed identity authentication method |
| CN106856015A (en) * | 2016-12-20 | 2017-06-16 | 国网山东省电力公司东明县供电公司 | A kind of Work attendance method and device |
| CN107026928A (en) * | 2017-05-24 | 2017-08-08 | 武汉大学 | A kind of behavioural characteristic identification authentication method and device based on mobile phone sensor |
| CN107093076A (en) * | 2016-02-18 | 2017-08-25 | 卡巴斯基实验室股份制公司 | The system and method for detecting fraudulent user transaction |
| CN107194216A (en) * | 2017-05-05 | 2017-09-22 | 中南大学 | A kind of mobile identity identifying method and system of the custom that swiped based on user |
| CN107451437A (en) * | 2016-05-31 | 2017-12-08 | 百度在线网络技术（北京）有限公司 | The locking means and device of a kind of mobile terminal |
| CN107623715A (en) * | 2017-08-08 | 2018-01-23 | 阿里巴巴集团控股有限公司 | A kind of identity information acquisition methods and device |
| CN107644106A (en) * | 2017-10-17 | 2018-01-30 | 厦门市美亚柏科信息股份有限公司 | The internuncial method of automatic mining business, terminal device and storage medium |
| CN107688827A (en) * | 2017-08-24 | 2018-02-13 | 西安交通大学 | A kind of user identity attribute forecast method based on user's daily behavior feature |
| CN107798446A (en) * | 2016-09-07 | 2018-03-13 | 北京嘀嘀无限科技发展有限公司 | The evaluation process method and device of rich ore route |
| CN108090216A (en) * | 2017-12-29 | 2018-05-29 | 咪咕文化科技有限公司 | A kind of Tag Estimation method, apparatus and storage medium |
| CN108600258A (en) * | 2018-05-09 | 2018-09-28 | 华东师范大学 | A kind of method for auditing safely towards Integrated Electronic System self-generating white list |
| CN109067970A (en) * | 2018-06-27 | 2018-12-21 | 上海擎感智能科技有限公司 | Based on the smart phone display methods and system of onboard instruments screen, car-mounted terminal |
| CN109145544A (en) * | 2018-09-05 | 2019-01-04 | 郑州云海信息技术有限公司 | A kind of human-computer behavior detection system and method |
| CN109583161A (en) * | 2018-11-27 | 2019-04-05 | 咪咕文化科技有限公司 | A kind of information processing method and device, storage medium |
| WO2019109786A1 (en) * | 2017-12-06 | 2019-06-13 | Oppo广东移动通信有限公司 | User gender recognition method and device |
| WO2019120023A1 (en) * | 2017-12-22 | 2019-06-27 | Oppo广东移动通信有限公司 | Gender prediction method and apparatus, storage medium and electronic device |
| 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 |
| CN110110766A (en) * | 2019-04-23 | 2019-08-09 | 西安交通大学 | A kind of online character analysis method and device based on motion planning controlling feature |
| CN110347928A (en) * | 2019-07-16 | 2019-10-18 | 韶关市启之信息技术有限公司 | A kind of recommendation method for personalized information and system |
| CN110363283A (en) * | 2019-06-06 | 2019-10-22 | 哈尔滨工业大学（深圳） | User property prediction technique and relevant apparatus based on deep learning |
| CN111046374A (en) * | 2019-11-26 | 2020-04-21 | 山东浪潮人工智能研究院有限公司 | Method and system for improving user password security based on machine learning |
| CN111147431A (en) * | 2018-11-06 | 2020-05-12 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating information |
| US10859387B2 (en) | 2016-04-27 | 2020-12-08 | Beijing Didi Infinity Technology And Development Co., Ltd. | System and method for determining routes of transportation service |
| WO2021093367A1 (en) * | 2019-11-13 | 2021-05-20 | 支付宝(杭州)信息技术有限公司 | Model training and risk identification method, apparatus and device |
| WO2021109761A1 (en) * | 2019-12-03 | 2021-06-10 | 深圳市联谛信息无障碍有限责任公司 | Method and system for determining disabled user in webpage |
| WO2021175010A1 (en) * | 2020-03-02 | 2021-09-10 | 深圳壹账通智能科技有限公司 | User gender identification method and apparatus, electronic device, and storage medium |
| WO2021243534A1 (en) * | 2020-06-02 | 2021-12-09 | 深圳市欢太科技有限公司 | Behavior control method and apparatus and storage medium |
| CN114747179A (en) * | 2020-06-25 | 2022-07-12 | 谷歌有限责任公司 | Abnormal user interface input detection |
| CN115081334A (en) * | 2022-06-30 | 2022-09-20 | 支付宝(杭州)信息技术有限公司 | Method, system, apparatus and medium for predicting age bracket or gender of user |
| CN116418587A (en) * | 2023-04-19 | 2023-07-11 | 中国电子科技集团公司第三十研究所 | Data cross-domain switching behavior audit trail method and data cross-domain switching system |
-
2013
- 2013-09-27 CN CN201310454565.2A patent/CN103530540B/en active Active
Non-Patent Citations (4)
| Title |
|---|
| 张明烁 等: "基于击键特征的手机用户身份认证系统", 《软件开发与设计》, vol. 2009, no. 22, 30 November 2009 (2009-11-30) * |
| 房超 等: "基于鼠标动力学模型的用户身份认证与监控", 《西安交通大学学报》, vol. 42, no. 10, 31 October 2008 (2008-10-31), pages 1235 - 1239 * |
| 朱明 等: "基于击键特征的用户身份认证新方法", 《计算机工程》, vol. 28, no. 10, 31 October 2002 (2002-10-31) * |
| 沈超 等: "基于鼠标行为特征的用户身份认证与监控", 《通信学报》, vol. 31, no. 7, 31 July 2010 (2010-07-31), pages 68 - 75 * |
Cited By (65)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015127739A1 (en) * | 2014-02-26 | 2015-09-03 | 华为技术有限公司 | Method and apparatus for grouping contacts |
| CN103870547A (en) * | 2014-02-26 | 2014-06-18 | 华为技术有限公司 | Grouping processing method and device of contact persons |
| TWI684148B (en) * | 2014-02-26 | 2020-02-01 | 華為技術有限公司 | Grouping processing method and device of contact person |
| CN103942489A (en) * | 2014-03-31 | 2014-07-23 | 中国科学院信息工程研究所 | Attack detection method and system on basis of cursor hidden scene |
| CN103927467A (en) * | 2014-04-25 | 2014-07-16 | 武汉大学 | Smart phone authentication system and method based on screen touching behavior |
| CN105279691A (en) * | 2014-07-25 | 2016-01-27 | 中国银联股份有限公司 | Financial transaction detection method and equipment based on random forest model |
| CN104239761A (en) * | 2014-09-15 | 2014-12-24 | 西安交通大学 | Continuous identity authentication method based on touch screen slip behavior characteristics |
| CN104537252B (en) * | 2015-01-05 | 2019-09-17 | 深圳市腾讯计算机系统有限公司 | User Status list disaggregated model training method and device |
| CN104537252A (en) * | 2015-01-05 | 2015-04-22 | 深圳市腾讯计算机系统有限公司 | User state single-classification model training method and device |
| CN104778387A (en) * | 2015-04-23 | 2015-07-15 | 西安交通大学 | Cross-platform identity authentication system and method based on human-computer interaction behaviors |
| CN104778387B (en) * | 2015-04-23 | 2017-12-08 | 西安交通大学 | Cross-platform identity authorization system and method based on man-machine interaction behavior |
| CN104809377A (en) * | 2015-04-29 | 2015-07-29 | 西安交通大学 | Method for monitoring network user identity based on webpage input behavior characteristics |
| CN104809377B (en) * | 2015-04-29 | 2018-01-05 | 西安交通大学 | Network user identity monitoring method based on webpage input behavior feature |
| CN104951544A (en) * | 2015-06-19 | 2015-09-30 | 百度在线网络技术（北京）有限公司 | User data processing method and system and method and system for providing user data |
| WO2016201933A1 (en) * | 2015-06-19 | 2016-12-22 | 百度在线网络技术（北京）有限公司 | User data processing method, providing method, system and computer device |
| CN105306496A (en) * | 2015-12-02 | 2016-02-03 | 中国科学院软件研究所 | User identity detection method and system |
| CN105607741A (en) * | 2015-12-31 | 2016-05-25 | 联想(北京)有限公司 | Control method and electronic equipment |
| CN105607741B (en) * | 2015-12-31 | 2019-01-15 | 联想(北京)有限公司 | A kind of control method and electronic equipment |
| CN107093076A (en) * | 2016-02-18 | 2017-08-25 | 卡巴斯基实验室股份制公司 | The system and method for detecting fraudulent user transaction |
| CN107093076B (en) * | 2016-02-18 | 2021-07-13 | 卡巴斯基实验室股份制公司 | System and method for detecting fraudulent user transactions |
| CN105760516A (en) * | 2016-02-25 | 2016-07-13 | 广州视源电子科技股份有限公司 | Method and device for distinguishing users |
| US10859387B2 (en) | 2016-04-27 | 2020-12-08 | Beijing Didi Infinity Technology And Development Co., Ltd. | System and method for determining routes of transportation service |
| CN107451437B (en) * | 2016-05-31 | 2021-04-16 | 百度在线网络技术（北京）有限公司 | Locking method and device of mobile terminal |
| CN107451437A (en) * | 2016-05-31 | 2017-12-08 | 百度在线网络技术（北京）有限公司 | The locking means and device of a kind of mobile terminal |
| CN106095915A (en) * | 2016-06-08 | 2016-11-09 | 百度在线网络技术（北京）有限公司 | The processing method and processing device of user identity |
| CN106156591A (en) * | 2016-07-12 | 2016-11-23 | 中国科学技术大学苏州研究院 | A kind of smart phone user Transparent Authentication method under cloud environment |
| CN106371750A (en) * | 2016-08-30 | 2017-02-01 | 北京奇艺世纪科技有限公司 | Method and device for confirming user gender |
| CN107798446A (en) * | 2016-09-07 | 2018-03-13 | 北京嘀嘀无限科技发展有限公司 | The evaluation process method and device of rich ore route |
| CN106503499A (en) * | 2016-09-22 | 2017-03-15 | 天津大学 | Smart mobile phone touch-screen input recognition method based on machine learning |
| CN106572097B (en) * | 2016-11-01 | 2019-07-09 | 南京邮电大学 | A kind of mobile device-based hybrid identity identifying method |
| CN106572097A (en) * | 2016-11-01 | 2017-04-19 | 南京邮电大学 | Mobile device-based mixed identity authentication method |
| CN106856015A (en) * | 2016-12-20 | 2017-06-16 | 国网山东省电力公司东明县供电公司 | A kind of Work attendance method and device |
| CN106503722A (en) * | 2016-12-21 | 2017-03-15 | 湖南文理学院 | Operator's skeleton growth rings method based on touch trajectory image procossing |
| CN106503722B (en) * | 2016-12-21 | 2019-09-24 | 湖南文理学院 | Operator's skeleton growth rings method based on touch trajectory image procossing |
| CN107194216A (en) * | 2017-05-05 | 2017-09-22 | 中南大学 | A kind of mobile identity identifying method and system of the custom that swiped based on user |
| CN107026928A (en) * | 2017-05-24 | 2017-08-08 | 武汉大学 | A kind of behavioural characteristic identification authentication method and device based on mobile phone sensor |
| CN107623715A (en) * | 2017-08-08 | 2018-01-23 | 阿里巴巴集团控股有限公司 | A kind of identity information acquisition methods and device |
| CN107623715B (en) * | 2017-08-08 | 2020-06-09 | 阿里巴巴集团控股有限公司 | Identity information acquisition method and device |
| CN107688827A (en) * | 2017-08-24 | 2018-02-13 | 西安交通大学 | A kind of user identity attribute forecast method based on user's daily behavior feature |
| CN107644106A (en) * | 2017-10-17 | 2018-01-30 | 厦门市美亚柏科信息股份有限公司 | The internuncial method of automatic mining business, terminal device and storage medium |
| CN107644106B (en) * | 2017-10-17 | 2020-10-09 | 厦门市美亚柏科信息股份有限公司 | Method, terminal device and storage medium for automatically mining service middleman |
| WO2019109786A1 (en) * | 2017-12-06 | 2019-06-13 | Oppo广东移动通信有限公司 | User gender recognition method and device |
| CN110020155A (en) * | 2017-12-06 | 2019-07-16 | 广东欧珀移动通信有限公司 | User's gender identification method and device |
| US11544583B2 (en) | 2017-12-06 | 2023-01-03 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method for gender recognition of user and related products |
| 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 |
| WO2019120023A1 (en) * | 2017-12-22 | 2019-06-27 | Oppo广东移动通信有限公司 | Gender prediction method and apparatus, storage medium and electronic device |
| CN108090216A (en) * | 2017-12-29 | 2018-05-29 | 咪咕文化科技有限公司 | A kind of Tag Estimation method, apparatus and storage medium |
| CN108600258A (en) * | 2018-05-09 | 2018-09-28 | 华东师范大学 | A kind of method for auditing safely towards Integrated Electronic System self-generating white list |
| CN109067970A (en) * | 2018-06-27 | 2018-12-21 | 上海擎感智能科技有限公司 | Based on the smart phone display methods and system of onboard instruments screen, car-mounted terminal |
| CN109145544A (en) * | 2018-09-05 | 2019-01-04 | 郑州云海信息技术有限公司 | A kind of human-computer behavior detection system and method |
| CN111147431A (en) * | 2018-11-06 | 2020-05-12 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating information |
| CN109583161A (en) * | 2018-11-27 | 2019-04-05 | 咪咕文化科技有限公司 | A kind of information processing method and device, storage medium |
| CN110110766A (en) * | 2019-04-23 | 2019-08-09 | 西安交通大学 | A kind of online character analysis method and device based on motion planning controlling feature |
| CN110363283B (en) * | 2019-06-06 | 2021-06-22 | 哈尔滨工业大学（深圳） | User attribute prediction method based on deep learning and related device |
| CN110363283A (en) * | 2019-06-06 | 2019-10-22 | 哈尔滨工业大学（深圳） | User property prediction technique and relevant apparatus based on deep learning |
| CN110347928A (en) * | 2019-07-16 | 2019-10-18 | 韶关市启之信息技术有限公司 | A kind of recommendation method for personalized information and system |
| WO2021093367A1 (en) * | 2019-11-13 | 2021-05-20 | 支付宝(杭州)信息技术有限公司 | Model training and risk identification method, apparatus and device |
| CN111046374A (en) * | 2019-11-26 | 2020-04-21 | 山东浪潮人工智能研究院有限公司 | Method and system for improving user password security based on machine learning |
| WO2021109761A1 (en) * | 2019-12-03 | 2021-06-10 | 深圳市联谛信息无障碍有限责任公司 | Method and system for determining disabled user in webpage |
| WO2021175010A1 (en) * | 2020-03-02 | 2021-09-10 | 深圳壹账通智能科技有限公司 | User gender identification method and apparatus, electronic device, and storage medium |
| WO2021243534A1 (en) * | 2020-06-02 | 2021-12-09 | 深圳市欢太科技有限公司 | Behavior control method and apparatus and storage medium |
| CN114747179A (en) * | 2020-06-25 | 2022-07-12 | 谷歌有限责任公司 | Abnormal user interface input detection |
| CN115081334A (en) * | 2022-06-30 | 2022-09-20 | 支付宝(杭州)信息技术有限公司 | Method, system, apparatus and medium for predicting age bracket or gender of user |
| CN116418587A (en) * | 2023-04-19 | 2023-07-11 | 中国电子科技集团公司第三十研究所 | Data cross-domain switching behavior audit trail method and data cross-domain switching system |
| CN116418587B (en) * | 2023-04-19 | 2024-04-30 | 中国电子科技集团公司第三十研究所 | Data cross-domain switching behavior audit trail method and data cross-domain switching system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103530540B (en) | 2017-02-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103530540B (en) | User identity attribute detection method based on man-machine interaction behavior characteristics | |
| CN103793484B (en) | The fraud identifying system based on machine learning in classification information website | |
| DE112016004922B4 (en) | Design and analysis system of the touch screen user button behavior mode and its identity recognition method | |
| CN104809377B (en) | Network user identity monitoring method based on webpage input behavior feature | |
| CN103530546B (en) | A kind of identity identifying method based on the behavior of user's mouse | |
| CN110443274A (en) | Method for detecting abnormality, device, computer equipment and storage medium | |
| CN109784091A (en) | A kind of list data method for secret protection merging difference privacy GAN and PATE model | |
| CN110781308B (en) | Anti-fraud system for constructing knowledge graph based on big data | |
| CN107577682A (en) | Users' Interests Mining and user based on social picture recommend method and system | |
| CN105760649B (en) | A kind of credible measure towards big data | |
| CN109447099A (en) | A kind of Combining Multiple Classifiers based on PCA dimensionality reduction | |
| CN104636631A (en) | Diabetes mellitus probability calculation method based on large data of diabetes mellitus system | |
| CN103412557A (en) | Industrial fault detection and diagnostic method suitable for nonlinear process on-line monitoring | |
| CN111368926B (en) | Image screening method, device and computer readable storage medium | |
| Mondal et al. | Continuous authentication in a real world settings | |
| Du et al. | An integrated framework based on latent variational autoencoder for providing early warning of at-risk students | |
| CN112329816A (en) | Data classification method and device, electronic equipment and readable storage medium | |
| CN107885857B (en) | A kind of search results pages user's behavior pattern mining method, apparatus and system | |
| CN113706251B (en) | Model-based commodity recommendation method, device, computer equipment and storage medium | |
| CN106537387B (en) | Retrieval/storage image associated with event | |
| CN109327480A (en) | A kind of multi-step attack scene method for digging based on neural network and Bayesian network attack graph | |
| CN110110663A (en) | A kind of age recognition methods and system based on face character | |
| CN109472462A (en) | A kind of project risk ranking method and device based on the fusion of multi-model storehouse | |
| CN110135487A (en) | A kind of computer user mouse Behavior modeling method | |
| CN106383586A (en) | Training system for children suffering from autistic spectrum disorders |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |