CN110197435B

CN110197435B - Object recognition method and device, storage medium and electronic device

Info

Publication number: CN110197435B
Application number: CN201810368880.6A
Authority: CN
Inventors: 潘杰; 魏雪
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2018-04-23
Filing date: 2018-04-23
Publication date: 2023-09-26
Anticipated expiration: 2038-04-23
Also published as: CN110197435A

Abstract

The invention discloses an object identification method and device, a storage medium and an electronic device. Wherein the method comprises the following steps: acquiring a target operation characteristic value of a specified operation characteristic of a specified object, wherein the specified operation characteristic is used for identifying operation information of a specified operation executed by the specified object; determining target probability that the specified object belongs to a target group according to the target operation characteristic value and a target model, wherein the target model is a model obtained by training an initial model by using a sample operation characteristic value of a sample object; and determining an object recognition result according to the determined target probability, wherein the object recognition result is used for indicating whether the specified object belongs to the target group. The invention solves the technical problem of poor accuracy of the identification result caused by the fact that the user identification mode depends on the personal data registered by the user.

Description

Object recognition method and device, storage medium and electronic device

Technical Field

The present invention relates to the field of computers, and in particular, to an object recognition method and apparatus, a storage medium, and an electronic apparatus.

Background

Currently, in order to identify whether a user belongs to a specific group, an account number of a social network platform used by a login application program needs to be acquired, and then personal data registered when the user uses the account number is queried to determine whether the user belongs to the specific group.

Since the above-described user identification method depends on the personal data registered by the user, the registration of the personal data has subjectivity, resulting in a problem of poor accuracy of the identification result.

In view of the above problems, no effective solution has been proposed at present.

Disclosure of Invention

The embodiment of the invention provides an object identification method and device, a storage medium and an electronic device, which are used for at least solving the technical problem of poor accuracy of an identification result caused by the fact that a user identification mode depends on personal data registered by a user.

According to an aspect of an embodiment of the present invention, there is provided an object recognition method including: acquiring a target operation characteristic value of a specified operation characteristic of a specified object, wherein the specified operation characteristic is used for identifying operation information of a specified operation executed by the specified object; determining target probability that the specified object belongs to a target group according to the target operation characteristic value and a target model, wherein the target model is a model obtained by training an initial model by using a sample operation characteristic value of a sample object; and determining an object recognition result according to the determined target probability, wherein the object recognition result is used for indicating whether the specified object belongs to the target group.

According to another aspect of the embodiment of the present invention, there is also provided an object recognition apparatus including: a first acquisition unit configured to acquire a target operation feature value of a specified operation feature of a specified object, where the specified operation feature is used to identify operation information of a specified operation performed by the specified object; the first determining unit is used for determining the target probability that the specified object belongs to a target group according to the target operation characteristic value and a target model, wherein the target model is a model obtained by training an initial model by using the sample operation characteristic value of the sample object; and a second determining unit, configured to determine an object recognition result according to the determined target probability, where the object recognition result is used to indicate whether the specified object belongs to the target group.

According to a further aspect of embodiments of the present invention, there is also provided a storage medium having stored therein a computer program, wherein the computer program is arranged to perform the above method when run.

According to still another aspect of the embodiments of the present invention, there is also provided an electronic device including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the above method by the computer program.

In the embodiment of the application, a mode of automatic identification is adopted, and a target operation characteristic value of a specified operation characteristic of a specified object is obtained, wherein the specified operation characteristic is used for identifying operation information of specified operation executed by the specified object; according to the target operation characteristic value and the target model, the probability that the appointed object belongs to the target group is determined, whether the appointed object belongs to the target group is determined according to the determined probability, and the aim of automatically identifying whether the appointed object belongs to the target group is fulfilled.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:

FIG. 1 is a schematic illustration of an application environment for an object recognition method according to an embodiment of the present application;

FIG. 2 is a flow chart of an alternative object recognition method according to an embodiment of the invention;

FIG. 3 is a schematic diagram of an alternative logistic regression according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of an alternative Sigmoid function according to an embodiment of the invention;

FIG. 5 is a schematic diagram of another alternative object recognition method according to an embodiment of the present invention;

FIG. 6 is a flow chart of another alternative object recognition method according to an embodiment of the invention;

FIG. 7 is a schematic diagram of an alternative object recognition apparatus according to an embodiment of the present invention;

fig. 8 is a schematic structural view of an alternative electronic device according to an embodiment of the present invention.

Detailed Description

In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

According to an aspect of an embodiment of the present invention, there is provided an object recognition method. Alternatively, the above-described object recognition method may be applied, but is not limited to, in an application environment as shown in fig. 1. As shown in fig. 1, a target operation feature value of a specified operation feature of a specified object is acquired in the terminal 102, wherein the specified operation feature is used to identify operation information of a specified operation performed by the specified object, and the acquired target operation feature value is transmitted to the server 104 through a network. The server 104 determines the target probability that the specified object belongs to the target group according to the target operation characteristic value and a target model, wherein the target model is a model obtained by training an initial model by using the sample operation characteristic value of the sample object; and determining an object recognition result according to the determined target probability, wherein the object recognition result is used for indicating whether the specified object belongs to the target group.

Alternatively, after determining the object recognition result according to the determined target probability, the server 104 may send, to the processing device 106 through the network, target information to instruct the processing device 106 to configure a service corresponding to the target information for the specified object, where the target information includes an identification of the specified object and an identification of the target group, in a case where the object recognition result is that the specified object belongs to the target group.

Alternatively, in this embodiment, the above terminal may include, but is not limited to, at least one of: a mobile phone, a tablet computer, etc. The network may include, but is not limited to, a wireless network, wherein the wireless network includes: bluetooth, WIFI, and other networks that enable wireless communications. The server may include, but is not limited to, at least one of: PCs and other devices for computing services. The above is merely an example, and the present embodiment is not limited thereto.

Optionally, in this embodiment, as an optional implementation manner, as shown in fig. 2, the above object identifying method may include:

s202, acquiring a target operation characteristic value of a specified operation characteristic of a specified object, wherein the specified operation characteristic is used for identifying operation information of a specified operation executed by the specified object;

S204, determining target probability that the specified object belongs to a target group according to the target operation characteristic value and a target model, wherein the target model is a model obtained by training an initial model by using the sample operation characteristic value of the sample object;

s206, determining an object recognition result according to the determined target probability, wherein the object recognition result is used for indicating whether the specified object belongs to the target group.

Alternatively, the above object recognition method may be, but not limited to, a process of recognizing whether an object belongs to a target group. For example, the method is applied to service development and popularization, and for example, by taking service development as an example, the identification characteristics of a target packet are analyzed by outputting objects belonging to the target packet, so as to form a "portrait" (common characteristics) of the target packet. Thus, the preference and the demand of the object of the target group are more accurately positioned, and the service research and development aiming at the target object is performed.

In the related art, the object recognition depends on the personal information registered by the user, and the registration of the personal information is subjective and has poor accuracy of the recognition result. In the application, the target operation characteristic value of the specified operation characteristic of the specified object is obtained, wherein the specified operation characteristic is used for identifying the operation information of the specified operation executed by the specified object; according to the target operation characteristic value and the target model, the probability that the appointed object belongs to the target group is determined, whether the appointed object belongs to the target group is determined according to the determined probability, the aim of automatically identifying the object can be fulfilled, and the accuracy of object identification can be improved due to the fact that the object is identified according to the appointed operation characteristic of the object, and further the problem that in the related art, the accuracy of an identification result is poor due to the fact that a user identification mode depends on personal data registered by a user is solved.

Alternatively, in the present embodiment, a target operation feature value of a specified operation feature of a specified object is acquired, wherein the specified operation feature is operation information for identifying a specified operation performed by the specified object.

Alternatively, the specified object may be a user of the target software, and the operation behavior of the specified object is acquired through the terminal in which the target software is installed. The designation operation may be any operation performed by a designated user. The specified operation feature may be operation information for identifying a specified operation performed by the specified object, and the operation information may be any information related to the specified operation, for example, indication information of whether the specified operation is performed, information of the number of times, frequency, time, and the like of performing the specified operation.

Alternatively, the specified operating characteristics may include, but are not limited to:

(1) A first feature for identifying operation information specifying a use operation performed by an object on the target software;

the target software may be software related to the target packet. The corresponding target software may be different for different target packets. Alternatively, the target software may include, but is not limited to: dial-up software, free WIFI software, game software, video software, etc.

Alternatively, the first feature may be the use of operational information related to the target software, which may include, but is not limited to: indication information of using the target software, the number of times of using the target software, frequency, time interval, and the like. For example, the first feature may be: the target operation feature value corresponding to the first feature may be: 1, representing use target software; 0, indicating that the target software is not used.

Alternatively, the target operation feature value of the first feature may be obtained by: the current process of the target device (corresponding to the specified object) is enumerated periodically.

For example, specifying operational characteristics includes: a first feature, the first feature being: operation information specifying a use operation performed by a user on dial software, the operation information being: specifying indication information of an application operation executed by an object on the dialing software (indicating whether the dialing software is used or not), enumerating a current process of the target device at regular time, using a target API (for example, ntQuerySystemInformation and the like), and if the application operation is the target software, specifying that a user uses the dialing software, wherein a target operation characteristic value is 1; if the target software is not the target software, the user is specified not to use the dialing software, and the target operation characteristic value is 0.

(2) A second feature for identifying operation information specifying an access operation performed by the object on the target web site;

the target web address may be software related to the target packet. The corresponding target web address may be different for different target packets. Alternatively, the target web site may be: a web site containing a specific field, or a specific web site.

For example, if the target packet is: college students, then the target web address may include, but is not limited to: a website with an edu.cn domain name, an intranet address of the edu.cn, a teaching website, and the like. For another example, if the target packet is: the target website may be: a website for selling mother and infant products.

Alternatively, the second feature may be operation information related to accessing the target web site, and the operation instruction information may include, but is not limited to: indication information of the access target website, the number of times of accessing the target website, frequency, time interval and the like. For example, the second feature may be: the indication information of the access operation performed by the specified object on the target website may be the target operation feature value corresponding to the second feature: 1, representing accessing a target website; 0, indicating that the target web site is not accessed.

Alternatively, the target operation feature value of the second feature may be obtained by: and determining the website accessed by the appointed user through the process of entering the browser.

For example, specifying operational characteristics includes: a second feature, the second feature being: operation information of access operation executed by a user on the teaching website is specified, wherein the operation information is as follows: the instruction information of the access operation executed by the appointed object on the teaching website (indicating whether the teaching website is accessed or not) enters the process of the browser through the injection module, and if the teaching website is accessed, the target operation characteristic value is 1; if the target software is not the target software, if the teaching website is not accessed, the target operation characteristic value is 0.

(3) A third feature for identifying operation information specifying a download operation performed by an object on a first target file containing a first specific keyword;

the first target file may be a file related to the target group. For different target packets, the corresponding first target file may be different. The first target file contains a first specific keyword related to the target packet. The first specific key may be included in: in the file name of the first target file, in the file body, in the file abstract, etc.

For example, if the target packet is: college students, the specific keywords contained in the first target file may include, but are not limited to: work, school timetables, experimental reports, and the like. For another example, if the target packet is: the mother with child, the specific keywords contained in the first target file may include, but are not limited to: infant, milk powder, child's brand, etc.

The third feature may be to download operation information related to the first target file, optionally, the operation information may include, but is not limited to: the indication information of the first target file is downloaded, the number of times the first target file is downloaded, the frequency, the time interval and the like. For example, the third feature may be: the indication information of the downloading operation performed by the specified object on the first target file may be that the target operation feature value corresponding to the third feature is: 1, representing downloading a first target file; and 0, indicating that the first target file is not downloaded.

For example, specifying operational characteristics includes: a third feature, the third feature being: specifying operation information of a download operation performed by an object on a first target file containing "course", the operation information being: specifying information (indicating whether or not to download the first target file) indicating a download operation performed by the object on the first target file containing "course", and if the first target file is downloaded by the specified user, setting the target operation characteristic value to 1; if the first target file is not downloaded, the target operation characteristic value is 0.

(4) A fourth feature for identifying operation information specifying an opening operation performed by an object on a second target file containing a second specific keyword;

the second target file may be a file related to the target group. For different target packets, the corresponding second target file may be different. The second target file contains a second specific keyword related to the target packet. The second specific key may be included in: in the file name of the second target file, in the file body, in the file abstract, etc.

For example, if the target packet is: college students, the specific keywords contained in the second target file may include, but are not limited to: work, school timetables, experimental reports, and the like. For another example, if the target packet is: the mother with child, the specific keywords contained in the second target file may include, but are not limited to: infant, milk powder, child's brand, etc.

The fourth feature may be operation information related to opening the second target file, optionally, the operation information may include, but is not limited to: indication information of opening the second target file, the number of times of opening the second target file, frequency, time interval, and the like. For example, the fourth feature may be: the indication information of the opening operation performed by the specified object on the second target file may be that the target operation feature value corresponding to the fourth feature is: 1, a second target file is opened; and 0, indicating that the second target file is not opened.

For example, specifying operational characteristics includes: a fourth feature, the fourth feature being: operation information of opening operation executed by the specified object on the second target file containing "course", the operation information being: indication information of an opening operation performed by a specified object on a second target file containing "course" (indicating whether or not the second target file is opened), and if the specified user opens the second target file, the target operation feature value is 1; if the second target file is not opened, the target operation feature value is 0.

(5) A fifth feature in which operation information for identifying an open operation performed by the specified object on the target window containing the third specific keyword;

the third specific key contained in the target window is a key related to the target packet. The corresponding target window may be different for different target packets. The third specific key is contained in: title of the target window, display content of the target window, and the like.

For example, if the target packet is: college students, then the specific keywords contained in the target window may include, but are not limited to: work, school timetables, experimental reports, and the like. For another example, if the target packet is: the mother with child, the specific keywords contained in the target window may include, but are not limited to: infant, milk powder, child's brand, etc.

The fifth feature may be operation information related to opening the target window, for example, instruction information to open the target window, the number of times the target window is opened, frequency, time interval, and the like. Optionally, the fifth feature may be: indication information specifying an open operation performed by the object on the target window (indicating whether or not to open the target window), the target operation feature value corresponding to the fifth feature may be: 1, representing opening a target window; 0, indicating that the target window is not opened.

Alternatively, the target operation feature value of the fifth feature may be obtained by: and (3) obtaining keywords contained in the current window by timing enumeration of window titles by using APIs such as EnumWindows and EnumChildWindows.

For example, specifying operational characteristics includes: fifth feature, fifth feature is: operation information specifying an opening operation performed by an object on a target window containing "course", the operation information being: specifying indication information (indicating whether or not to open a target window) of an opening operation performed by an object on the target window containing "course", enumerating window titles at regular time, and if the target window is opened by a specified user, setting a target operation characteristic value to be 1; if the target window is not opened, the target operation feature value is 0.

Optionally, specifying the operational characteristics may further include: operation information (sixth feature) specifying a user's joining operation to the target group, for example, the target group may be a group containing a fourth specific keyword. The fourth specific key may be a key related to the target packet. For example, the target group is college students, and the fourth specific keyword may be "class", "expiration".

Alternatively, the target operation feature value of the sixth feature may be obtained by: by timing enumeration of window titles, APIs such as EnumWindows and EnumChildWindows are used, and if the current window is a specific Internet media platform (corresponding to the target group), keywords contained in the current window or a process corresponding to the specific Internet media platform are acquired.

Optionally, specifying the operational characteristics may further include: operation information (seventh feature) specifying an access operation performed by a user on the network, which may include, but is not limited to: the time of joining the network, the time of exiting the network, the duration of accessing the network, etc.

Optionally, specifying the operational characteristics may further include: operation information (eighth feature) specifying an opening operation or a closing operation performed by a user on a target device, the operation information may include, but is not limited to: time to turn on or off the target device, total open time of the target device, etc. The opening operation of the target device can be obtained through analysis of event information of the system, the closing operation of the target device can be obtained through GetTicketCount,

Optionally, specifying the operational characteristics may further include: operation information (ninth feature) specifying an opening operation or a closing operation performed by a user on a target game or a target video, the operation information may include, but is not limited to: time to turn on or off the target game, total open time of the target game, etc. The opening operation or closing operation of the target game or the target video may be acquired in a similar manner as the above-described designating operation performed on the target software.

Alternatively, different designated operational characteristics may be used for different target packets. For a target packet, one or more of the specified operating characteristics may be used. For example, the specified operational characteristics may include one or more first characteristics, one or more second characteristics, one or more third characteristics, one or more fourth characteristics, one or more fifth characteristics. The above is merely an example, and is not limited thereto in the present embodiment.

For example, the target packet is: college students, assigned operating characteristics are: (1) using dialing software; (2) accessing a website for an edu.cn domain name; (3) accessing a teaching website; (4) Downloaded files containing "textbooks", "jobs" or "experimental reports".

Optionally, for different specified operation features, the range of values of the target operation feature values of the specified operation feature of the specified object may be the same or different, and the specific target operation feature value is determined according to the execution condition of the specified operation by the user.

Optionally, in this embodiment, the target probability that the specified object belongs to the target group is determined according to the target operation feature value and a target model, where the target model is a model obtained by training the initial model using the sample operation feature value of the sample object.

Alternatively, the target model may be a mathematical algorithm model for determining the probability that the sample output is a particular value, and a data algorithm model for classifying the sample. In the target model, a target operation feature value (may be plural) as a specified object is input, the input domain may be [ - +++ ], the output is a packet to which the specified object belongs, typically discrete, i.e. having a limited number of output values, for example, the range of values may have only two values {0,1},1 indicating that the specified object belongs to the target group and 0 indicating that the specified object does not belong to the target group. The output result is: the probability that the object belongs to the target group is specified.

Alternatively, the target model may be obtained by training the initial model using sample values of the sample objects. When training to obtain the target model, two groups of sample objects can be established first, one group is a first sample object belonging to the target group, the other group is a second sample object not belonging to the target group, and the number of the two groups of sample objects can be the same. The manner in which two types of sample objects are obtained may include, but is not limited to, one of the following:

1) Two types of sample objects can be obtained in an investigation mode, and firstly, whether each sample object belongs to a target group or not is determined by receiving feedback information of the sample object on whether the sample object belongs to the target group;

2) By receiving grouping information of sample objects from other devices or a target database, it is determined that each sample object belongs to a target group or does not belong to a target group.

Alternatively, after two sets of sample objects are acquired, the sample objects of the two sets of sample objects may be further grouped, respectively, one set for training the data model (training sample) and one set for testing the data model (test sample) obtained after training. The grouping rules of the two sets of sample data (sample objects belonging to the target group and sample objects not belonging to the target group) may be the same such that the number of sample objects belonging to the target group and sample objects not belonging to the target group are equal in the sample objects for the training data model and for the test data model.

Optionally, in the present embodiment, a first operation feature value of a first specified operation feature of a first sample object belonging to the target group and a second operation feature value of a second specified operation feature of a second sample object not belonging to the target group are acquired, wherein the first specified operation feature is used for identifying operation information of a specified operation performed by the first sample object, and the second specified operation feature is used for identifying operation information of a specified operation performed by the second sample object; and training the initial model by using the first operation characteristic value and the second operation characteristic value to obtain a target model.

Alternatively, for the initial model, the number of variables (specified operating characteristics) may be varied, each variable corresponding to a specified operating characteristic. The specific value of the number of variables can be set as needed.

Optionally, when training the initial model, the input of the initial model is a first operation feature value of a first designated operation feature of a first sample object belonging to the target group and a second operation feature value of a second designated operation feature of a second sample object not belonging to the target group, and the first designated operation feature and the second operation feature are operation information of an object performing a designated operation, but the objects corresponding to the first designated operation feature and the second operation feature are different. The number of the first sample objects and the second sample objects may be one or more, and may be set according to actual needs, which is not limited in this embodiment. The initial model is trained using the first operational characteristic value of the first sample object and the second operational characteristic value of the second sample object to obtain a target model.

Optionally, in this embodiment, training the initial model using the first operation feature value and the second operation feature value, to obtain the target model includes: training an initial model by using the first operation characteristic value, the second operation characteristic value and the initial weight of the appointed operation characteristic to obtain a target model, wherein the initial weight is calculated according to the number of the sample objects which are subjected to the appointed operation in the first sample object and the number of the sample objects which are subjected to the appointed operation in the second sample object.

In the initial model, each specified operation feature has a coefficient (i.e., weight) corresponding to it, and the initial value of each coefficient can be calculated from the number of sample objects in the first sample object on which the specified operation corresponding to the coefficient is performed and the number of sample objects in the second sample object on which the specified operation corresponding to the coefficient is performed. For example, the ratio of the number of sample objects in the first sample object on which the specified operation corresponding to the coefficient is performed to the number of sample objects in the test sample (including the first sample object and the second sample object) on which the specified operation corresponding to the coefficient is performed may be calculated.

The initial model and the target model are described with particular reference to the following examples. The initial model and the target model can be logistic regression models, the logistic regression is a supervised statistical learning method, and the logistic regression models are generalized linear regression analysis models which are commonly used in the fields of data mining, automatic disease diagnosis, economic prediction and the like.

In a linear regression model, the output is typically continuous, for example: y=f (x) =ax+b, there is a corresponding y output for each input x. Both the domain and the value range of the model may be [ - ≡, ++ infinity ].

In connection with fig. 3, logistic regression is to solve the classification problem, train a logistic regression model according to some known training set (e.g., triangles and snowflakes in fig. 3), and predict new data (e.g., circles in fig. 3) to predict which class the data belongs to. It can be seen that the goal of logistic regression is to find decision boundaries with sufficient discrimination to be able to distinguish the various types of data well.

For the logistic regression model, the input may be continuous [ - ] infinity, +++ ], but the output is typically discrete, i.e. only a limited number of output values. For example, it may have only two values {0,1}, which may represent some sort of sample (high/low, sick/healthy, negative/positive, etc.), which is the most common logistic regression of the two classes. Thus, overall, x over the whole real range can be mapped to a limited number of points by means of a logistic regression model, thus achieving classification of x. For a certain x, using a logistic regression model, the probability of being assigned to a certain class can be determined, and then the probability of being assigned to a certain class y can be assigned to a certain class y through logistic regression analysis.

Logistic regression is also known as a generalized linear regression model, which is essentially the same form as a linear regression model, having ax+b (this is an example, the number of variables x may contain multiple), where a and b are the parameters to be solved. The difference between them is that the multiple linear regression directly takes ax+b as the dependent variable, i.e. y=ax+b, whereas the logistic regression corresponds ax+b to one hidden state p, p=s (ax+b) by a function S, and then determines the value of the dependent variable according to the magnitudes of p and 1-p. The function S here is a Sigmoid function:

one schematic function image of the Sigmoid function is shown in fig. 4.

And changing t into ax+b, so as to obtain the parameter form of the logistic regression model:

by the action of the function S, the output value can be limited to the interval 0,1, p (x) can be used to represent the probability p (y= 1|x) (probability of y=1), i.e. the probability that y is divided into 1 group when one x occurs, a threshold value can be set, e.g. 0.5, and when y >0.5, this x is classified as 1, and if y <0.5, x is classified as 0. The threshold may be adjustable, for example, it is possible to set the threshold to 0.8, that is to say there is a probability of more than 80%, and this x is considered to belong to the category 1.

Training the initial model by using the operation characteristic values of the specified operation characteristics of the sample objects in the test sample, so as to obtain coefficient values and constant values of the specified operation characteristics, and thus obtain the target model.

Optionally, in this embodiment, an object identification result is determined according to the determined target probability, where the object identification result is used to indicate whether the specified object belongs to the target group.

Optionally, after determining the target probability that the specified object belongs to the target group, the target probability may be compared with a preset threshold, and if the target probability is greater than or equal to the preset threshold, the object recognition result may be determined as follows: the preset object belongs to a target group, and if the target probability is smaller than a preset threshold value, the object identification result can be determined as follows: the preset object does not belong to the target group.

Optionally, the preset threshold may be set as required, and in a default case, the preset threshold may be set to 0.5; if the accuracy requirement for the predictions belonging to the target packet is not high, the preset threshold may be set to be less than 0.5, e.g., 0.4; if the accuracy requirement for the predictions belonging to the target packet is high, the preset threshold may be set to be greater than 0.5, e.g., 0.8.

Optionally, in this embodiment, after determining the object recognition result according to the determined target probability, determining a target IP address range in a case where the object recognition result is that the specified object belongs to the target packet, where the target IP address range includes the first IP address used by the specified object; it is determined that the object using the second IP address belongs to the target packet, wherein the second IP address is included in the IP address range.

Alternatively, in the case where it is determined that the specified object belongs to the target packet, if the IP addresses used by the users in the target packet have relevance (for example, the IP addresses are the same, the IP addresses are similar, or the IP addresses are consecutive) according to the characteristics of the users in the target packet, IP address diffusion may be performed based on the IP address of the specified object, and the target IP address range including the IP address used by the specified object may be determined. An object using an IP address contained in the target IP address range also belongs to the target packet.

Alternatively, after determining that the object using the second IP address belongs to the target packet, the target model may be subjected to multidimensional learning based on the operation feature values of the specified operation feature of the object using the second IP address, thereby optimizing the model parameters.

Optionally, in this embodiment, after determining the object recognition result according to the determined target probability, in a case where the object recognition result is that the specified object belongs to the target group, the target information is sent to the processing device, so as to instruct the processing device to configure, for the specified object, a service corresponding to the target information, where the target information includes an identification of the specified object and an identification of the target group.

Alternatively, in the case where it is determined that the specified object belongs to the target packet, the target packet including the identification of the specified object and the identification of the target packet may be transmitted to the processing apparatus to instruct the processing apparatus to configure the service corresponding to the target information for the specified object. The target information may also include:

1) Specifying an operation feature of the object and a target operation feature value of the operation feature;

2) An operation feature value of another operation feature of the specified object other than the specified operation feature.

After receiving the target information, the processing device may analyze behavior habits (commonalities) of objects in the target group according to information contained in target information corresponding to different specified objects belonging to the target group, form a user portrayal of "target group" so as to configure a service corresponding to the target group (e.g., perform product development, function configuration), or configure a service for the target group in combination with user portrayal in other target groups. After receiving the target information, the processing device may select, according to the target information, a part of services from the services corresponding to the target packet as the designated object to configure, so as to enhance the user experience.

By the embodiment, the target operation characteristic value of the specified operation characteristic of the specified object is obtained, wherein the specified operation characteristic is used for identifying the operation information of the specified operation executed by the specified object; according to the target operation characteristic value and the target model, the probability that the appointed object belongs to the target group is determined, and whether the appointed object belongs to the target group or not is determined according to the determined probability, so that the aim of automatically identifying the object can be fulfilled, and the accuracy of object identification is improved.

As an alternative, after determining the object recognition result according to the determined target probability, the method further includes:

and sending target information to the processing equipment to instruct the processing equipment to configure a service corresponding to the target information for the specified object under the condition that the specified object belongs to the target group as a result of object identification, wherein the target information comprises the identification of the specified object and the identification of the target group.

According to the method and the device, the target information comprising the identification of the designated object and the identification of the target group is sent to the processing equipment, so that the processing equipment is instructed to configure the service corresponding to the target information for the designated object, the pertinence of service configuration is improved, and the user experience is improved.

s1, determining a target IP address range under the condition that an object identification result is that a specified object belongs to a target packet, wherein the target IP address range comprises a first IP address used by the specified object;

s2, determining that the object using the second IP address belongs to the target packet, wherein the second IP address is contained in the IP address range.

According to the embodiment, more users in the target group are determined through IP address diffusion, so that target model optimization is conveniently performed, the operation characteristics of objects in the target group are analyzed, and further the accuracy of prediction of the target model and the comprehensiveness of operation characteristic analysis are guaranteed.

As an alternative, before acquiring the target operation feature value of the specified operation feature of the specified object, the method further includes:

s1, acquiring a first operation characteristic value of a first specified operation characteristic of a first sample object belonging to a target group and a second operation characteristic value of a second specified operation characteristic of a second sample object not belonging to the target group, wherein the first specified operation characteristic is used for identifying operation information of a specified operation executed by the first sample object, and the second specified operation characteristic is used for identifying operation information of the specified operation executed by the second sample object;

S2, training the initial model by using the first operation characteristic value and the second operation characteristic value to obtain a target model.

Optionally, training the initial model using the first operational feature value and the second operational feature value, and obtaining the target model includes:

training an initial model by using the first operation characteristic value, the second operation characteristic value and the initial weight of the appointed operation characteristic to obtain a target model, wherein the initial weight is calculated according to the number of the sample objects which are subjected to the appointed operation in the first sample object and the number of the sample objects which are subjected to the appointed operation in the second sample object.

According to the method and the device, the initial model is trained by using the operation characteristic values of the specified operation characteristics of the sample objects belonging to the target group and the operation characteristic values of the specified operation characteristics of the sample objects not belonging to the target group, so that the target model is obtained, and the rationality of target model determination and the accuracy of a prediction result can be ensured due to the comprehensive consideration of the operation characteristics of the sample objects belonging to the target group and the operation characteristics of the sample objects not belonging to target molecules.

As an alternative, the specified operating characteristics include at least one of:

A first feature for identifying operation information specifying a use operation performed by an object on the target software;

a second feature for identifying operation information specifying an access operation performed by the object on the target web site;

a third feature, wherein the third feature is used for identifying operation information of a downloading operation executed by a specified object on a first target file, and the first target file contains a first specific keyword;

a fourth feature, wherein the fourth feature is used for identifying operation information of an opening operation executed by a specified object on a second target file, and the second target file contains a second specific keyword;

and a fifth feature for identifying operation information specifying an opening operation performed by the object on the target window, the target window including a third specific keyword.

According to the method and the device, different user behaviors are analyzed, and the aspects of software features, website features, document features and the like are considered to determine the specified operation features, so that the rationality of the specified operation features is ensured, and the accuracy of the prediction result is further ensured.

The above object recognition method will be described below with reference to fig. 5 and 6. In this example, the target group is "college student".

Referring to fig. 5 and 6, the object recognition method in this example includes the steps of:

s602, the server screens out college student user samples from the user big data.

To obtain a college student user sample, a batch of accurate data is required, i.e., it is explicitly known whether a batch of users are college students. In this example, the acquisition is performed by way of a questionnaire. The process is as follows:

firstly, tips (small stickers) are issued to a group of users through channels, after clicking, the users can open a page, and the page is a questionnaire page, and the questionnaire can comprise the following problems: whether college students, what university, etc. Thus, a group of accurate college student users are screened out.

S604, after screening out college student user samples, data of 50% of known users (college students and non-college students) is selected as a sample set for calculation.

When the sample set is selected, only the user sample which is a college student can be selected as the sample set so as to perform data algorithm model training.

S606, collecting characteristic values of operation characteristics of each sample in the sample set.

For a user, there are many features, and for a feature specifying whether it is a college student, the features shown in table 1 are selected for analysis.

TABLE 1

And acquiring the execution condition of each sample on each feature through a feature process or a specific API on the terminal corresponding to the sample.

The identification capability of different features to 'college students' is different, the accuracy of a single feature is analyzed first, and the analysis result is as follows:

1) Accessing a school website: the differentiation is very high and the coverage is relatively large.

After analyzing the sample to perform the operation of "visit school website", the samples to perform the operation are: 433, among which, samples are college students: 332 samples, not college students, were: 101, the ratio of the number of samples of the college students to the total number of samples for performing the operation is: 76.7%.

2) Dialing software: the differentiation degree is high, and the coverage rate is also more.

After analyzing the sample execution using the "dialing software", the sample executing the operation is: 863, among them, samples of college students are: 835, samples not for college students are: the ratio of the number of samples, which are college students, to the total number of samples for performing the operation is 28: 96.8%.

3) Start-up time, shut-down time, game time, video time: college and non-college students have no obvious distinction.

4) The name of wifi has keywords: library, class

The amount is small and cannot be verified.

5) The download file has corresponding keywords (operation, course, experiment report): the differentiation is higher and the coverage ratio is lower.

After analyzing the operation of executing the "file downloading the corresponding keyword (job, course, experiment report)" on the sample, the sample executing the operation is: 77 samples, among which are college students: the 60 samples, not college students, were: the ratio of the number of samples, which are college students, to the total number of samples for performing the operation is 17: 77.9%.

6) Open window keywords (university, college): after removing some keywords (e.g., college students, university cities, etc.), the accuracy is generally high, the coverage is relatively high

After analysis of the sample performed "open window with keywords (university, college)", the sample performed this operation was: 543, among them, the samples of college students are: 365, samples not college students were: 178, the ratio of the number of samples of the college student to the total number of samples for performing the operation is: 67.2%.

6) Open window other keywords (e.g., specify platform names): even with accuracy, coverage is relatively low.

7) The specific social platform process has the following keywords (ban, then): no obvious difference

After analyzing the operation of the sample execution "use a specific social platform with a specific keyword (ban )", the sample execution of the operation is: 912, among them, the samples of college students are: the 401 samples, not college students, are: 511, the ratio of the number of samples of the college students to the total number of samples for performing the operation is: 44%.

8) The open window has the following keywords: the operation, course, school time, experiment report, higher degree of differentiation and lower coverage rate.

After analyzing the operation of executing the "window with specific keywords (job, course, school, experiment report)" on the sample, the sample executing the operation is: 106, among them, the samples of college students are: the 69 samples, not college students, were: the ratio of the number of samples, which are college students, to the total number of samples for performing the operation is 37: 65.1%.

S608, training the data algorithm model by using the characteristic values of the operation characteristics of the collected samples.

By training the data algorithm model (initial model), all the features (or part of the features) are mixed, and each feature weight value (M ₁ ,…,M _n ). The analysis results show that: accessing 2 features of edu.cn and dialing software, in terms of accuracy and coverageThe effect of the rate is greatest, with a weighting value that is greater relative to the weighting values of the other features.

S610, predicting the test sample by using the trained data algorithm model.

After training is completed and weighting values for each feature are obtained, the data of the other 50% of users of the known users are selected for testing. The prediction result is 1, which is regarded as college students, and the prediction result is 0, which is regarded as non-college students.

Through detection, the accuracy rate of identifying college students by the model is high, and the service of accurately identifying college students can be provided.

And S612, carrying out IP diffusion according to the IP address of the user of the college student in the prediction result, identifying more college student users, and further optimizing the data algorithm model.

This step is an optional step. Considering the network IP of each university, there is a range of IP, and if a user a under an IP is considered a college student, all IP users under a certain range can be considered college students. Thus, more college students can be identified by means of IP diffusion. Further, through multidimensional learning, the iteration is continued until the optimal algorithm weighting value (M ₁ ,…,M _n )。

It should be noted that, for simplicity of description, the foregoing method embodiments are all described as a series of acts, but it should be understood by those skilled in the art that the present invention is not limited by the order of acts described, as some steps may be performed in other orders or concurrently in accordance with the present invention. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily required for the present invention.

From the description of the above embodiments, it will be clear to a person skilled in the art that the method according to the above embodiments may be implemented by means of software plus the necessary general hardware platform, but of course also by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) comprising several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the method of the various embodiments of the present invention.

According to another aspect of the embodiment of the present invention, there is also provided an object recognition apparatus for implementing the above object recognition method, as shown in fig. 7, the apparatus including:

(1) A first obtaining unit 702 configured to obtain a target operation feature value of a specified operation feature of a specified object, where the specified operation feature is operation information identifying a specified operation performed by the specified object;

(2) A first determining unit 704, configured to determine, according to a target operation feature value and a target model, a target probability that a specified object belongs to a target group, where the target model is a model obtained by training an initial model using a sample operation feature value of a sample object;

(3) The second determining unit 706 is configured to determine an object identification result according to the determined target probability, where the object identification result is used to indicate whether the specified object belongs to the target group.

Alternatively, the above-described object recognition apparatus may be, but is not limited to, a process of recognizing whether or not an object belongs to a target group. For example, the method is applied to service development and popularization, and for example, by taking service development as an example, the identification characteristics of a target packet are analyzed by outputting objects belonging to the target packet, so as to form a "portrait" (common characteristics) of the target packet. Thus, the preference and the demand of the object of the target group are more accurately positioned, and the service research and development aiming at the target object is performed.

In the related art, the object recognition depends on the personal information registered by the user, and the registration of the personal information is subjective and has poor accuracy of the recognition result. In the application, the target operation characteristic value of the specified operation characteristic of the specified object is obtained, wherein the specified operation characteristic is used for identifying the operation information of the specified operation executed by the specified object; according to the target operation characteristic value and the target model, the probability that the appointed object belongs to the target group is determined, and whether the appointed object belongs to the target group is determined according to the determined probability, so that the aim of automatically identifying the object can be fulfilled.

Alternatively, the specified object may be a user of the target software, and the operation behavior of the specified object is acquired through the terminal in which the target software is installed. The designation operation may be any operation performed by a designated user. The specified operation feature may be operation information for identifying a specified operation performed by the specified object, and the operation information may be any information related to the specified operation.

optionally, specifying the operational characteristics may further include:

1) Operation information (sixth feature) specifying a joining operation of the target group by the user.

2) Operation information specifying an access operation performed by the user on the network (seventh feature).

3) Operation information (eighth feature) specifying an opening operation or a closing operation performed by a user on the target device;

4) Operation information (ninth feature) specifying an opening operation or a closing operation performed by a user on a target game or a target video.

The above features may refer to the above embodiments, and are not described herein.

Alternatively, the target model may be a mathematical algorithm model for determining the probability that the sample output is a particular value, and a data algorithm model for classifying the sample. In the target model, a target operation characteristic value (may be plural) for a specified object is input, the definition field of the input may be [ - ≡, the output is a packet to which the specified object belongs, typically discrete, i.e., only a limited number of output values. The output result is: the probability that the object belongs to the target group is specified.

Alternatively, in the case where it is determined that the specified object belongs to the target packet, the target IP address range including the IP address used by the specified object may be determined based on the IP address of the specified object if the IP addresses used by the users in the target packet have an association (e.g., the IP addresses are the same, the IP addresses are similar, or the IP addresses are consecutive) according to the characteristics of the users in the target packet. An object using an IP address contained in the target IP address range also belongs to the target packet.

As an alternative embodiment, the apparatus further comprises:

and the sending unit is used for sending the target information to the processing equipment to instruct the processing equipment to configure the service corresponding to the target information for the specified object after the object identification result is determined according to the determined target probability and in the case that the specified object belongs to the target group as the object identification result, wherein the target information comprises the identification of the specified object and the identification of the target group.

As an alternative, the apparatus further includes:

(1) A third determining unit configured to determine, after determining the object recognition result according to the determined target probability, a target IP address range in a case where the object recognition result is that the specified object belongs to the target packet, wherein the target IP address range includes a first IP address used by the specified object;

(2) A fourth determination unit configured to determine that the object using the second IP address belongs to the target packet, wherein the second IP address is included in the IP address range.

As an alternative, the apparatus further includes:

(1) A second acquisition unit configured to acquire, before acquiring a target operation feature value of a specified operation feature of a specified object, a first operation feature value of a first specified operation feature of a first sample object belonging to a target group, and a second operation feature value of a second specified operation feature of a second sample object not belonging to the target group, wherein the first specified operation feature is used for identifying operation information of a specified operation performed by the first sample object, and the second specified operation feature is used for identifying operation information of a specified operation performed by the second sample object;

(2) And the training unit is used for training the initial model by using the first operation characteristic value and the second operation characteristic value to obtain a target model.

Optionally, the training unit comprises:

the training module is used for training the initial model by using the first operation characteristic value, the second operation characteristic value and the initial weight of the appointed operation characteristic to obtain a target model, wherein the initial weight is calculated according to the number of the sample objects which are subjected to the appointed operation in the first sample object and the number of the sample objects which are subjected to the appointed operation in the second sample object.

According to a further aspect of embodiments of the present invention there is also provided a storage medium having stored therein a computer program, wherein the computer program is arranged to perform the steps of any of the method embodiments described above when run.

Alternatively, in the present embodiment, the above-described storage medium may be configured to store a computer program for performing the steps of:

s1, acquiring a target operation characteristic value of a specified operation characteristic of a specified object, wherein the specified operation characteristic is used for identifying operation information of a specified operation executed by the specified object;

s2, determining target probability that the specified object belongs to a target group according to the target operation characteristic value and a target model, wherein the target model is a model obtained by training an initial model by using the sample operation characteristic value of the sample object;

and S3, determining an object recognition result according to the determined target probability, wherein the object recognition result is used for indicating whether the specified object belongs to the target group.

s1, after an object recognition result is determined according to the determined object probability, sending target information to processing equipment to instruct the processing equipment to configure a service corresponding to the target information for the specified object under the condition that the specified object belongs to a target group as the object recognition result, wherein the target information comprises the identification of the specified object and the identification of the target group.

s1, after an object recognition result is determined according to the determined target probability, determining a target IP address range under the condition that the object recognition result is that a specified object belongs to a target packet, wherein the target IP address range comprises a first IP address used by the specified object;

s1, determining that an object using a second IP address belongs to a target packet, wherein the second IP address is contained in an IP address range.

s1, before acquiring a target operation characteristic value of a specified operation characteristic of a specified object, acquiring a first operation characteristic value of a first specified operation characteristic of a first sample object belonging to a target group and a second operation characteristic value of a second specified operation characteristic of a second sample object not belonging to the target group, wherein the first specified operation characteristic is used for identifying operation information of a specified operation executed by the first sample object, and the second specified operation characteristic is used for identifying operation information of the specified operation executed by the second sample object;

Alternatively, in this embodiment, it will be understood by those skilled in the art that all or part of the steps in the methods of the above embodiments may be performed by a program for instructing a terminal device to execute the steps, where the program may be stored in a computer readable storage medium, and the storage medium may include: flash disk, read-Only Memory (ROM), random-access Memory (Random Access Memory, RAM), magnetic or optical disk, and the like.

According to still another aspect of the embodiment of the present invention, there is also provided an electronic device for implementing the above object recognition method, as shown in fig. 8, the electronic device including: a processor 802, a memory 804, a transmission device 806, and the like. The memory has stored therein a computer program, the processor being arranged to perform the steps of any of the method embodiments described above by means of the computer program.

Alternatively, in this embodiment, the electronic apparatus may be located in at least one network device of a plurality of network devices of the computer network.

Alternatively, in the present embodiment, the above-described processor may be configured to execute the following steps by a computer program:

Alternatively, it will be understood by those skilled in the art that the structure shown in fig. 8 is only schematic, and the electronic device may also be a terminal device such as a smart phone (e.g. an Android phone, an iOS phone, etc.), a tablet computer, a palm computer, and a mobile internet device (Mobile Internet Devices, abbreviated as MID), a PAD, etc. Fig. 8 is not limited to the structure of the electronic device. For example, the electronic device may also include more or fewer components (e.g., network interfaces, etc.) than shown in FIG. 8, or have a different configuration than shown in FIG. 8.

The memory 804 may be used to store software programs and modules, such as program instructions/modules corresponding to the object recognition method and apparatus in the embodiment of the present invention, and the processor 802 executes the software programs and modules stored in the memory 804, thereby performing various functional applications and data processing, that is, implementing the object recognition method. The memory 804 may include high-speed random access memory, but may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid state memory. In some examples, the memory 804 may further include memory remotely located relative to the processor 802, which may be connected to the terminal via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The transmission device 806 is used to receive or transmit data via a network. Specific examples of the network described above may include wired networks and wireless networks. In one example, the transmission device 806 includes a network adapter (Network Interface Controller, simply referred to as NIC) that can connect to other network devices and routers via a network cable to communicate with the internet or a local area network. In one example, the transmission device 806 is a Radio Frequency (RF) module, which is configured to communicate with the internet wirelessly.

The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.

The integrated units in the above embodiments may be stored in the above-described computer-readable storage medium if implemented in the form of software functional units and sold or used as separate products. Based on such understanding, the technical solution of the present invention may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, comprising several instructions for causing one or more computer devices (which may be personal computers, servers or network devices, etc.) to perform all or part of the steps of the method described in the embodiments of the present invention.

In the foregoing embodiments of the present application, the descriptions of the embodiments are emphasized, and for a portion of this disclosure that is not described in detail in this embodiment, reference is made to the related descriptions of other embodiments.

In several embodiments provided by the present application, it should be understood that the disclosed client may be implemented in other manners. The above-described embodiments of the apparatus are merely exemplary, and the division of the units, such as the division of the units, is merely a logical function division, and may be implemented in another manner, for example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some interfaces, units or modules, or may be in electrical or other forms.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The foregoing is merely a preferred embodiment of the present invention and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present invention, which are intended to be comprehended within the scope of the present invention.

Claims

1. An object recognition method, comprising:

acquiring a target operation characteristic value of a specified operation characteristic of a specified object, wherein the specified operation characteristic is used for identifying operation information of a specified operation executed by the specified object;

the acquiring the specified operation characteristics of the specified object comprises at least one of the following steps: enumerating a current process of a target device corresponding to the specified object at regular time to determine a first feature of the specified operational features, wherein the first feature is used for identifying whether the specified object uses target software; determining a website accessed by the specified object through a process of entering a browser so as to determine a second characteristic in the specified operation characteristics, wherein the second characteristic is used for identifying whether a target website accessed by the specified object comprises a specific field or is a specific website;

Determining target probability that the specified object belongs to a target group according to the target operation characteristic value and a target model, wherein the target model is a model obtained by training an initial model by using a sample operation characteristic value of a sample object;

the determining, according to the target operation feature value and the target model, the target probability that the specified object belongs to the target group includes: determining the target operation characteristic value according to the first characteristic and/or the second characteristic, and inputting the target operation characteristic value into the target model to determine the target probability that the specified object belongs to a target group;

and determining an object recognition result according to the determined target probability, wherein the object recognition result is used for indicating whether the specified object belongs to the target group.

2. The method of claim 1, wherein after determining the object recognition result according to the determined target probability, the method further comprises:

and sending target information to processing equipment to instruct the processing equipment to configure a service corresponding to the target information for the specified object under the condition that the object identification result is that the specified object belongs to the target group, wherein the target information comprises the identification of the specified object and the identification of the target group.

3. The method of claim 1, wherein after determining an object recognition result based on the determined target probability, the method further comprises:

determining a target IP address range under the condition that the object identification result is that the specified object belongs to the target packet, wherein the target IP address range comprises a first IP address used by the specified object;

determining that an object using a second IP address belongs to the target packet, wherein the second IP address is included in the IP address range.

4. The method of claim 1, wherein prior to obtaining the target operational characteristic value for the specified operational characteristic of the specified object, the method further comprises:

acquiring a first operation characteristic value of a first specified operation characteristic of a first sample object belonging to the target group and a second operation characteristic value of a second specified operation characteristic of a second sample object not belonging to the target group, wherein the first specified operation characteristic is used for identifying operation information of the specified operation performed by the first sample object, and the second specified operation characteristic is used for identifying operation information of the specified operation performed by the second sample object;

And training the initial model by using the first operation characteristic value and the second operation characteristic value to obtain the target model.

5. The method of claim 4, wherein training the initial model using the first operational feature value and the second operational feature value to obtain the target model comprises:

training the initial model by using the first operation characteristic value, the second operation characteristic value and the initial weight of the appointed operation characteristic to obtain the target model, wherein the initial weight is obtained by calculating the number of sample objects which execute the appointed operation in the first sample object and the number of sample objects which execute the appointed operation in the second sample object.

6. The method of any one of claims 1 to 5, wherein the specified operating characteristics include at least one of:

a third feature, wherein the third feature is used for identifying operation information of a downloading operation performed on a first target file by the specified object, and the first target file contains a first specific keyword;

a fourth feature, wherein the fourth feature is used for identifying operation information of an opening operation performed by the specified object on a second target file, and the second target file contains a second specific keyword;

And a fifth feature, wherein the fifth feature is used for identifying operation information of an opening operation performed on a target window by the specified object, and the target window contains a third specific keyword.

7. An object recognition apparatus, comprising:

a first acquisition unit configured to acquire a target operation feature value of a specified operation feature of a specified object, where the specified operation feature is used to identify operation information of a specified operation performed by the specified object;

the first acquisition unit includes at least one of:

a first obtaining subunit, configured to periodically enumerate a current process of a target device corresponding to the specified object, so as to determine a first feature in the specified operation features, where the first feature is used to identify whether the specified object uses target software;

a second obtaining subunit, configured to determine, by a process of entering the browser, a website visited by the specified object, so as to determine a second feature in the specified operation features, where the second feature is used to identify whether a target website visited by the specified object includes a specific field or is a specific website;

the first determining unit is used for determining the target probability that the specified object belongs to a target group according to the target operation characteristic value and a target model, wherein the target model is a model obtained by training an initial model by using the sample operation characteristic value of the sample object;

The first determining unit is configured to determine, according to the target operation feature value and the target model, a target probability that the specified object belongs to a target group by: determining the target operation characteristic value according to the first characteristic and/or the second characteristic, and inputting the target operation characteristic value into the target model to determine the target probability that the specified object belongs to a target group;

and a second determining unit, configured to determine an object recognition result according to the determined target probability, where the object recognition result is used to indicate whether the specified object belongs to the target group.

8. The apparatus of claim 7, wherein the apparatus further comprises:

and the sending unit is used for sending target information to processing equipment to instruct the processing equipment to configure a service corresponding to the target information for the specified object after the object identification result is determined according to the determined target probability, wherein the target information comprises the identification of the specified object and the identification of the target group when the object identification result is that the specified object belongs to the target group.

9. The apparatus of claim 7, wherein the apparatus further comprises:

a third determining unit configured to determine, after determining an object recognition result according to the determined target probability, a target IP address range in a case where the object recognition result is that the specified object belongs to the target packet, wherein the target IP address range includes a first IP address used by the specified object;

a fourth determining unit configured to determine that an object using a second IP address belongs to the target packet, where the second IP address is included in the IP address range.

10. The apparatus of claim 7, wherein the apparatus further comprises:

a second acquisition unit configured to acquire, before acquiring the target operation feature value of the specified operation feature of the specified object, a first operation feature value of a first specified operation feature of a first sample object belonging to the target group and a second operation feature value of a second specified operation feature of a second sample object not belonging to the target group, wherein the first specified operation feature is used for identifying operation information of the specified operation performed by the first sample object, and the second specified operation feature is used for identifying operation information of the specified operation performed by the second sample object;

And the training unit is used for training the initial model by using the first operation characteristic value and the second operation characteristic value to obtain the target model.

11. The apparatus of claim 10, wherein the training unit comprises:

the training module is configured to train the initial model by using the first operation feature value, the second operation feature value and the initial weight of the specified operation feature to obtain the target model, where the initial weight is calculated according to the number of sample objects in the first sample object, in which the specified operation is performed, and the number of sample objects in the second sample object, in which the specified operation is performed.

12. A storage medium having a computer program stored therein, wherein the computer program is arranged to perform the method of any of claims 1 to 6 when run.

13. An electronic device comprising a memory and a processor, characterized in that the memory has stored therein a computer program, the processor being arranged to execute the method according to any of the claims 1 to 6 by means of the computer program.