CN111324509A - Method and device for identifying application addiction - Google Patents
Method and device for identifying application addiction Download PDFInfo
- Publication number
- CN111324509A CN111324509A CN202010098424.1A CN202010098424A CN111324509A CN 111324509 A CN111324509 A CN 111324509A CN 202010098424 A CN202010098424 A CN 202010098424A CN 111324509 A CN111324509 A CN 111324509A
- Authority
- CN
- China
- Prior art keywords
- application
- addiction
- target user
- user
- data
- 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
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3055—Monitoring arrangements for monitoring the status of the computing system or of the computing system component, e.g. monitoring if the computing system is on, off, available, not available
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/302—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system component is a software system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/29—Graphical models, e.g. Bayesian networks
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Probability & Statistics with Applications (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method and a device for identifying application addiction, which comprise the following steps: constructing an application addiction recognition model based on the Bayesian network; acquiring the use behavior data of a target user to the application; obtaining an addiction risk index of the target user to the application according to the using behavior data and the application addiction recognition model; and judging whether the target user is addicted to the application or not according to the addicted risk index. According to the application addiction risk index identification method, the Bayesian network technology is applied to application addiction identification, and the prior probability and the posterior probability are used for obtaining the addiction risk index with higher reliability, so that the accuracy of application addiction identification is improved.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to a method and an apparatus for identifying addiction to an application.
Background
With the widespread use of smart devices (such as smart watches, learning tablets, etc.) and the increasing comprehensiveness of the functions (such as various application functions of telephones, WeChat, photos, music, education, games, etc.), children can enjoy various benefits brought by the smart devices, but on the other hand, because the children have limited self-control, some applications, such as network games, videos, social software, are easy to allow the children to be addicted therein, which has certain influence on the mental development of the children and normal learning.
At present, the application enthusiasm recognition is mainly focused on the field of games, the accumulated use duration, the age, the user types (such as working crowds, students and the like), the recharge amount and the like are mainly considered, the judgment process is too simple and insufficient, and for example, if the continuous game exceeds 3 hours, the user is considered to be enthusiasm for the game application.
Disclosure of Invention
The invention aims to provide a method and a device for identifying an application enthusiasm, which are used for solving the problem that the identification of the application enthusiasm of a user in the prior art is too simple and violent.
The technical scheme provided by the invention is as follows:
a method of identifying an application addiction, comprising: constructing an application addiction recognition model based on the Bayesian network; acquiring the use behavior data of a target user to the application; obtaining an addiction risk index of the target user to the application according to the using behavior data and the application addiction recognition model; and judging whether the target user is addicted to the application or not according to the addicted risk index.
Further, the nodes of the bayesian network comprise at least two of the following features: the method has the advantages of long accumulated use time, more application starting times, use of sensitive time, use of sensitive applications, large power consumption ratio of the applications and whether the applications are management and control applications.
Further, the establishing of the application addiction recognition model based on the Bayesian network comprises the following steps: acquiring application use behavior data of a plurality of users as sample data; extracting feature sample data corresponding to the Bayesian network node from the sample data; carrying out normalization processing on each type of feature sample data; and determining the structure and parameters of the Bayesian network according to the feature sample data normalized by each dimension.
Further, the obtaining of the addiction risk index of the target user to the application according to the usage behavior data and the application addiction recognition model includes: acquiring node characteristics of the target user according to the using behavior data; calculating the posterior probability of the target user indulging the application according to the generated node characteristics and the prior probability in the application indulging recognition model; and taking the posterior probability of the target user indulging the application as the indulging risk index of the target user to the application.
Further, the sample data also comprises attribute information of an application user; and obtaining the prior knowledge of the application addiction recognition model according to the attribute information of the application user.
Further, the sample data also comprises feedback and complaint information of the user to the application; the obtaining of the prior knowledge of the application addiction recognition model according to the attribute information of the application user comprises: and obtaining the prior knowledge of the application addiction recognition model according to the attribute information of the application user and the feedback and complaint information of the user to the application.
Further, the using the posterior probability of the target user indulging the application as the indulging risk index of the target user to the application includes: adjusting the posterior probability of the target user addicting the application according to the prior knowledge of the application addict recognition model; and taking the adjusted posterior probability as an indulging risk index of the target user to the application.
The present invention also provides a device for identifying application addiction, comprising: the model construction module is used for constructing an application addiction recognition model based on the Bayesian network; the data acquisition module is used for acquiring the use behavior data of the target user to the application; the addiction risk calculation module is used for obtaining an addiction risk index of the target user to the application according to the use behavior data and the application addiction recognition model; and the judging module is used for judging whether the target user is addicted to the application according to the addict risk index.
Further, the model building module comprises: the system comprises a sample acquisition unit, a data processing unit and a data processing unit, wherein the sample acquisition unit is used for acquiring application use behavior data of a plurality of users as sample data; the data preprocessing unit is used for extracting feature sample data corresponding to the Bayesian network node from the sample data; carrying out normalization processing on each type of feature sample data; and the model construction unit is used for determining the structure and the parameters of the Bayesian network according to the feature sample data normalized by each dimension.
Further, the addiction risk calculation module is further configured to obtain node features of the target user according to the usage behavior data; calculating the posterior probability of the target user indulging the application according to the generated node characteristics and the prior probability in the application indulging recognition model; and taking the posterior probability of the target user indulging the application as the indulging risk index of the target user to the application.
The method and the device for identifying the application addiction provided by the invention can bring the following beneficial effects:
1. according to the application addiction risk index identification method, the Bayesian network technology is applied to application addiction identification, and the prior probability and the posterior probability are used for obtaining the addiction risk index with higher reliability, so that the accuracy of application addiction identification is improved.
2. According to the invention, the information related to application use is obtained from multiple channels such as customer service, e-commerce comments, APP feedback data and the like, and the multidimensional characteristics and the prior knowledge are introduced into the Bayesian network, so that the accuracy of application addiction identification is further improved.
Drawings
The above features, technical features, advantages and implementations of an application addiction recognition method and device will be further described in the following detailed description of preferred embodiments in a clearly understandable manner, in conjunction with the accompanying drawings.
FIG. 1 is a flow diagram of one embodiment of a method of identifying application addiction, in accordance with the present invention;
FIG. 2 is a flow diagram of another embodiment of a method of identifying application addiction, in accordance with the present invention;
FIG. 3 is a schematic diagram of a Bayesian network of FIG. 2;
FIG. 4 is a flow diagram of another embodiment of a method of identifying application addiction, in accordance with the present invention;
FIG. 5 is a flowchart of step S100 in FIG. 4;
FIG. 6 is a schematic diagram of an embodiment of an application addiction recognition device according to the present invention;
FIG. 7 is a schematic diagram of one configuration of the mold building block of FIG. 6;
the reference numbers illustrate:
100. the model construction method comprises a model construction module 200, a data acquisition module 300, an addiction risk calculation module 400, a judgment module 110, a sample acquisition unit 120, a data preprocessing unit 130 and a model construction unit.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following description will be made with reference to the accompanying drawings. It is obvious that the drawings in the following description are only some examples of the invention, and that for a person skilled in the art, other drawings and embodiments can be derived from them without inventive effort.
For the sake of simplicity, the drawings only schematically show the parts relevant to the present invention, and they do not represent the actual structure as a product. In addition, in order to make the drawings concise and understandable, components having the same structure or function in some of the drawings are only schematically illustrated or only labeled. In this document, "one" means not only "only one" but also a case of "more than one".
In one embodiment of the present invention, as shown in fig. 1, a method for identifying application addiction includes:
step S100, constructing an application addiction recognition model based on the Bayesian network;
step S200, acquiring the use behavior data of the target user to the application;
step S300, obtaining an addiction risk index of the target user to the application according to the using behavior data and the application addiction recognition model;
step S400, judging whether the target user is addicted to the application according to the addict risk index.
In particular, a bayesian network is a probabilistic graphical model adapted to express and analyze events of uncertainty and probability, applied to decisions that are conditionally dependent on a variety of control factors, and can reason from incomplete, inaccurate knowledge or information.
A bayesian network is a directed acyclic graph that consists of nodes representing variables and directed edges connecting these nodes. The nodes represent random variables, the directed edges among the nodes represent the mutual correlation (from the father node to the son node), and the relation strength is expressed by conditional probability. For any random variable, its joint probability distribution can be multiplied by the respective local conditional probability distributions.
With the prior knowledge, the topological structure of the Bayesian network and the conditional probability distribution parameters of the nodes can be established through given sample data. On the premise of the determined node topological structure and conditional probability distribution, the posterior probability can be calculated for new data by using the network, so that the purposes of diagnosis, prediction and the like are achieved.
Bayesian network techniques are applied to the identification of application addiction. In the construction of the application addiction recognition model, firstly, a plurality of characteristics of the addiction application of the user are determined according to the prior knowledge or expert knowledge, and the characteristics are used as nodes of the Bayesian network. For example, the accumulated usage time is long, the number of times of application start is large, the usage time is sensitive, the sensitive application is used, and the like. Each feature reflects the user's enthusiasm for the application from a different dimension. The sensitive time is generally the time of work/class, and for students, the time of learning is the sensitive time period, such as working day 08: 00-11: 30, 13: 50-17: 00. Sensitive applications refer to applications that are prone to being enthusiastic, such as social entertainment, gaming, and the like.
Application usage behavior data of a plurality of users is collected as sample data. According to the sample data, determining the correlation between the nodes, namely determining a father node, a child node and the conditional probability distribution parameters of each node, and thus establishing an application addiction recognition model based on the Bayesian network structure.
And acquiring the use behavior data of the target user to the application in a period of time, which is equivalent to acquiring new data. And judging which features have occurred according to the new data, and calculating the probability that the user is addicted to the application, namely the posterior probability, according to the occurred features and the prior probability (including the conditional probability distribution parameters of each node) in the application addicted recognition model. And taking the posterior probability as an indulging risk index of the target user to the application, and judging whether the target user indulges the application according to the index. For example, the value is compared with a preset application addiction threshold, and if the value exceeds the preset application addiction threshold, it indicates that the target user is indulging the application.
For example, it is assumed that the node y indicates that the corresponding application is addicted, and has two values, 1 ═ True and 0 ═ false, and the node β indicates that the accumulated usage time is long, and has two values, 1 ═ True and 0 ═ false, the value of β is determined according to the average daily accumulated usage time, and when the average daily accumulated usage time is not less than 1 hour, it is determined that the user-sensitive time is addicted, that is, an event β ═ 1 occurs.
Assume that, based on historical statistics, the prior probability P (y-1) is 20%, P (β -1) is 24%, and the conditional probability P (β -1 | y-1) is 80%.
The data of the application B1 used by the user a1 is that the average daily cumulative use time is 70 minutes, and from the average daily cumulative use time, the user a1 has an addiction to the sensitive time, that is, β is 1 event.
According to this feature, the risk index of addiction of the user a1 to the application B1 is calculated as P (y 1| β ═ 1) ═ P (β ═ 1| y ═ 1)/P (β ═ 1) — 20% × 80%/24% > -67%.
And judging whether the user A1 is addicted to the application B1 according to the addicted risk index. Assuming that the application addiction threshold is 80%, user A1 has not reached addict application B1; however, it can be seen that, in the case of the occurrence of addiction in sensitive time, the probability of addiction to the application of the user is greatly improved, and is increased from 20% to 67%. If user A1 experiences an addiction of a sensitive duration lasting several days, the posterior probability will continue to rise above the application addiction threshold.
If the user is considered to be addicted to the application beyond 1 hour, as judged from the cumulative usage time alone, it is judged to be too rough. But the Bayesian network is used for calculating the posterior probability, and the enthusiasm of the user to the application is judged according to the posterior probability, so that the reliability is obviously higher.
According to the method and the device, the Bayesian network technology is applied to the application addiction recognition, the prior probability is utilized to obtain the posterior probability, and the addiction risk index with higher reliability is obtained according to the posterior probability, so that the accuracy of the application addiction recognition is improved.
In another embodiment of the present invention, as shown in fig. 2 and 3, a method for identifying application addiction includes:
step S100, constructing an application addiction recognition model based on the Bayesian network;
wherein, step S100 includes:
step S110, acquiring application use behavior data of a plurality of users as sample data;
step S120, extracting feature sample data corresponding to the Bayesian network node from the sample data;
step S130, normalizing each type of feature sample data;
step S140 determines the structure and parameters of the bayesian network according to the feature sample data normalized by each dimension.
Step S200, acquiring the use behavior data of the target user to the application;
step S310, acquiring node characteristics of the target user according to the using behavior data;
step S320, calculating the posterior probability of the target user indulging the application according to the generated node characteristics and the prior probability in the application indulging recognition model;
step S330, taking the posterior probability of the target user indulging the application as an indulging risk index of the target user to the application;
step S400, judging whether the target user is addicted to the application according to the addict risk index.
Specifically, a bayesian network is established, and first, network nodes are determined. The network node comprises the characteristics of long accumulated use time, more application starting times, sensitive time use, sensitive application use and the like in multiple dimensions.
Acquiring application use behavior data of a plurality of users as sample data. And acquiring feature sample data of corresponding dimensions from the sample data.
Application usage behavior data includes, but is not limited to: application attributes (e.g., system application or third party application, learning tool/life convenience/game/social entertainment/music story, etc.), usage time of the application (e.g., weekday/weekday, work period/lesson period, etc.), number of times of launch of the application (e.g., total number of application clicks during a day), usage duration of the application (e.g., cumulative usage duration, usage duration for a time period), application power consumption (e.g., power consumption for the entire day or time period of the application), application disable period, etc.
And performing more fine division and normalization processing on the feature sample data of each dimension, and unifying the data of different dimensions to the same dimension and range. By carrying out equalization processing on the multidimensional data source, the equalization of the sample weight of each dimension in the whole model is ensured. The discrimination of the model for various situations is enhanced by finer partitioning.
Optionally, finely dividing each type of feature sample data according to a preset rule of the corresponding dimension and converting the feature sample data into an addiction value of the corresponding dimension; wherein the range of addict values of each dimension is uniform. For example, the range of the addiction value is limited to 0-1, and the average daily cumulative use time of each user is converted into a corresponding addiction value according to table 1 (the first column is the average daily cumulative use time range, and the second column is the addiction value); converting the use time period into a corresponding addiction value according to the table 2; converting the average daily starting times into corresponding addiction values according to the table 3; the different application types are converted to corresponding addiction values according to table 4. The following table is merely an example of segmentation, and other segmentation methods are not limited to this.
| (0,5]Minute (min) | 0.1 |
| (5,10]Minute (min) | 0.2 |
| (10,15]Minute (min) | 0.3 |
| (15,20]Minute (min) | 0.4 |
| (20,25]Minute (min) | 0.5 |
| (25,30]Minute (min) | 0.6 |
| (30,35]Minute (min) | 0.65 |
| (35,40]Minute (min) | 0.7 |
| (40,45]Minute (min) | 0.75 |
| (45,50]Minute (min) | 0.8 |
| (50,55]Minute (min) | 0.9 |
| (55,60]Minute (min) | 0.95 |
| (>60) Minute (min) | 1.0 |
TABLE 1
| The number of enthusiasm at working day of 08: 00-11: 30 | 1 |
| 13: 50-17: 00 enthusiasm value in working day | 1 |
| Other time periods | 0 |
TABLE 2
TABLE 3
| Learning tool | 0.4 |
| Life convenience | 0.5 |
| Intelligence game | 0.9 |
| Social entertainment | 0.9 |
| System applications | 0.3 |
| Music story | 0.6 |
TABLE 4
And determining the relation of each node in the network and the conditional probability distribution parameter of each node according to the feature sample data normalized by each dimension. And when the relation of each node is determined, determining the network structure, and when the conditional probability distribution parameter of each node is determined, determining the network parameter. The Bayesian network with determined structure and parameters is used as an application addiction recognition model.
It is assumed that the bayesian network shown in fig. 3 is obtained according to the feature sample data of each dimension. Wherein, the node y represents the application indulgence, and the probability is represented by P (y).
The node β indicates that the cumulative use time is long P2 indicates that the sensitive time is indulged, and the probability that the user is indulged in the application due to the long cumulative use time is indicated by P (β | y).
Node θ represents a usage sensitive application. P3 denotes the addiction of the sensitive application, and the probability of the user being addicted to the application for using the sensitive application is denoted by P (θ | y).
The node α represents the sensitive time usage, P1 represents the probability of the sensitive time being in the state of being long in the sensitive time period, and the probability is represented by P (α | β, y) because the combination of the accumulated usage time and the sensitive time usage may cause the sensitive time to be in the state of being long in the sensitive time period and the sensitive application may cause the sensitive time to be in the state of being in the sensitive time period.
Through sample data analysis, hidden associations exist between the accumulated use time and the starting times and between the sensitive application types and the starting times, the accumulated use time is long, the sensitive application types possibly cause the starting times to be large, P4 indicates a sensitive number addiction, namely the addiction with the large starting times, and the probability of the addiction is represented by P (gamma | β, theta).
As shown in fig. 3, the node γ has a parent-child relationship with the node β and the node θ, respectively, and the node α has a parent-child relationship with the node β and the node y, respectively.
And acquiring application use behavior data of the target user in a period of time, such as a day, a week or a month, and acquiring the use behavior data of each application. Analyzing the usage behavior data of the target user for each application, for example, the usage behavior data of the application a in units of days, and statistically acquiring average daily cumulative usage time, average distribution of daily usage time periods, average daily startup times, and the like.
And then analyzing whether the node characteristics occur from each node dimension, for example, judging whether the node characteristics are used in sensitive time according to the distribution of the average use time period per day, judging whether the application starting times are large according to the average starting times per day, and the like.
The sampling data of each node dimension can be converted into corresponding addiction values, that is, the addiction values of the node β, the node theta, the node α and the node gamma are respectively calculated according to the application behavior data of the target user, and then whether the corresponding node characteristics occur or not is respectively judged according to the addiction values of the nodes.
For example, if the time period information of the application used by the user is not collected, it cannot be judged whether the user uses the application in a sensitive time period, that is, whether the characteristic of the node α occurs or not, and the bayesian network has the advantage that reasonable reasoning can be carried out on the incomplete information based on the prior probability and the occurred information.
For example, when the features of the nodes β, theta, gamma and α all have occurred, the probability of indulging the application is P (y | β, theta, gamma, α) ═ P (y) 2P 3P 1P 4 ═ P (y) P (β | y) | P (θ | y) × P (α | β, y) × P (γ | β, theta).
And taking the posterior probability as an indulging risk index of the target user to the application, and judging whether the target user indulges the application according to the index.
Fig. 3 is only one example. Nodes may be reduced or increased on the network structure described above. Adding a node is equivalent to introducing new features, such as whether an application is a management application, a large power consumption ratio of the application, and the like. The management and control application comprises management and control application recommended by an official party and management and control application customized by parents. Governing applications are often prone to people becoming addicted, typically setting an application disable period. If the user uses the application during the suggested application disable period, it is indicative of a high user enthusiasm for the application. If a user is indulged in an application, the power consumption duty of the application is usually large. Therefore, the new features are also beneficial to applying enthusiasm recognition, and the addition of the new features can further improve the accuracy of recognition.
In the embodiment, by introducing the multi-dimensional node characteristics (such as long accumulated use time, multiple application starting times, sensitive time use, sensitive application use and the like) into the bayesian network and acquiring the application use behavior data of the multi-dimensional characteristics, the defect that the traditional method cannot accurately identify the addiction due to insufficient characteristics is effectively overcome, and the accuracy of identifying the application addiction is further improved.
In another embodiment of the present invention, as shown in fig. 4 and 5, a method for identifying an application addiction includes:
step S100, constructing an application addiction recognition model based on the Bayesian network;
wherein, step S100 includes:
step S110, acquiring application use behavior data of a plurality of users, attribute information of application users and/or feedback and complaint information of the users to the application as sample data;
step S120, extracting feature sample data corresponding to the Bayesian network node from the application use behavior data;
step S130, normalizing each type of feature sample data;
step S140, determining the structure and parameters of the Bayesian network according to feature sample data normalized by each dimension;
and S150, obtaining the prior knowledge of the application addiction recognition model according to the attribute information of the application user and/or the feedback and complaint information of the user to the application.
Step S200, acquiring the use behavior data of the target user to the application;
step S310, acquiring node characteristics of the target user according to the using behavior data;
step S320, calculating the posterior probability of the target user indulging the application according to the generated node characteristics and the prior probability in the application indulging recognition model;
step S331, adjusting the posterior probability of the target user indulging the application according to the prior knowledge of the application indulging recognition model;
step S332, taking the adjusted posterior probability as an indulging risk index of the target user to the application;
step S400, judging whether the target user is addicted to the application according to the addict risk index.
Specifically, the difference from the previous embodiment is:
when the application addiction recognition model is constructed, the acquired sample data also comprises attribute information of an application user and/or feedback and complaint information of the user to the application. Obtaining the prior knowledge of the application addiction recognition model according to the attribute information of the application user and/or the feedback and complaint information of the user to the application; adjusting the posterior probability of the target user addicting the application according to the prior knowledge of the application addict recognition model; and taking the adjusted posterior probability as an indulging risk index of the target user to the application.
The attribute information of the application user includes the age, sex, region, etc. of the user. The feedback of the user to the application and the source of the complaint information are not limited to e-commerce comments, customer service feedback/complaint information and the like.
The judgment criteria of the application addiction by the users with different age groups/different sexes/different regions are different, for example, the middle school students are addicted to play games for more than 1 hour, and the preschool children are likely to be addicted to use continuously for more than 30 minutes. It is desirable to reflect this difference in the indulgence risk index.
The present embodiment embodies the above difference by introducing a priori knowledge λ. And adjusting the calculated posterior probability of the addiction application of the target user according to the prior knowledge lambda so as to obtain a more accurate addiction risk index. For example, considering different user types, the a priori knowledge λ is: for junior middle school and above, λ ═ 1; below junior school, λ is 1.5. Assuming that the posterior probability calculated based on the bayesian network is P, if the target user is a pupil, the addiction risk index is adjusted to 1.5P.
The priori knowledge lambda can be obtained by statistical data analysis, for example, the same application enthrallment distribution applied to different age groups, different sexes and different regions can be obtained; if λ of one age group is set to 1, λ of other age groups can be obtained according to the applied addiction distribution. Similarly, the prior knowledge λ may also be corrected based on the review feedback.
The embodiment is applied to enthusiasm use recognition of APP software on a child watch. And obtaining application use behavior data of the user through APP feedback data, and obtaining user feedback information through the parent complaint information received by customer service and the E-commerce comment information of the application.
The application use behavior data fed back by the child watches are collected, the indulging risk index of each user to each application is dynamically analyzed according to the application use behavior data of each user, and the related suggestions are fed back to parents according to the indulging risk indexes, so that the parents can better control the applications on the child watches, for example, which applications are set to be forbidden, forbidden time periods and the like.
In the embodiment, the prior knowledge is introduced, the posterior probability calculated by the Bayesian network is corrected, and the accuracy of identifying the application addiction is further improved.
In one embodiment of the present invention, as shown in fig. 6, a recognition apparatus for an application addiction, includes:
the model building module 100 is used for building an application addiction recognition model based on the Bayesian network;
a data obtaining module 200, configured to obtain usage behavior data of an application by a target user;
the addiction risk calculation module 300 is configured to obtain an addiction risk index of the target user to the application according to the usage behavior data and the application addiction recognition model;
and the judging module 400 is used for judging whether the target user is indulged in the application according to the indulgence risk index.
In particular, a bayesian network is a probabilistic graphical model adapted to express and analyze events of uncertainty and probability, applied to decisions that are conditionally dependent on a variety of control factors, and can reason from incomplete, inaccurate knowledge or information.
A bayesian network is a directed acyclic graph that consists of nodes representing variables and directed edges connecting these nodes. The nodes represent random variables, the directed edges among the nodes represent the mutual correlation (from the father node to the son node), and the relation strength is expressed by conditional probability. For any random variable, its joint probability distribution can be multiplied by the respective local conditional probability distributions.
With the prior knowledge, the topological structure of the Bayesian network and the conditional probability distribution parameters of the nodes can be established through given sample data. On the premise of the determined node topological structure and conditional probability distribution, the posterior probability can be calculated for new data by using the network, so that the purposes of diagnosis, prediction and the like are achieved.
Bayesian network techniques are applied to the identification of application addiction. In the construction of the application addiction recognition model, firstly, a plurality of characteristics of the addiction application of the user are determined according to the prior knowledge or expert knowledge, and the characteristics are used as nodes of the Bayesian network. For example, the accumulated usage time is long, the number of times of application start is large, the usage time is sensitive, the sensitive application is used, and the like. Each feature reflects the user's enthusiasm for the application from a different dimension. The sensitive time is generally the time of work/class, and for students, the time of learning is the sensitive time period, such as working day 08: 00-11: 30, 13: 50-17: 00. Sensitive applications refer to applications that are prone to being enthusiastic, such as social entertainment, gaming, and the like.
Application usage behavior data of a plurality of users is collected as sample data. According to the sample data, determining the correlation between the nodes, namely determining a father node, a child node and the conditional probability distribution parameters of each node, and thus establishing an application addiction recognition model based on the Bayesian network structure.
And acquiring the use behavior data of the target user to the application in a period of time, which is equivalent to acquiring new data. And judging which features have occurred according to the new data, and calculating the probability that the user is addicted to the application, namely the posterior probability, according to the occurred features and the prior probability (including the conditional probability distribution parameters of each node) in the application addicted recognition model. And taking the posterior probability as an indulging risk index of the target user to the application, and judging whether the target user indulges the application according to the index. For example, the value is compared with a preset application addiction threshold, and if the value exceeds the preset application addiction threshold, it indicates that the target user is indulging the application.
According to the method and the device, the Bayesian network technology is applied to the application addiction recognition, the prior probability is utilized to obtain the posterior probability, and the addiction risk index with higher reliability is obtained according to the posterior probability, so that the accuracy of the application addiction recognition is improved.
In another embodiment of the present invention, as shown in fig. 6 and 7, a device for recognizing an application addiction, includes:
the model building module 100 is used for building an application addiction recognition model based on the Bayesian network;
wherein, the model building module 100 includes:
a sample acquiring unit 110, configured to acquire application usage behavior data of a plurality of users as sample data;
a data preprocessing unit 120, configured to extract feature sample data corresponding to the bayesian network node from the sample data; carrying out normalization processing on each type of feature sample data;
and the model building unit 130 is configured to determine a structure and parameters of the bayesian network according to the feature sample data normalized by each dimension.
A data obtaining module 200, configured to obtain usage behavior data of an application by a target user;
the addiction risk calculation module 300 is configured to obtain node features that have occurred to the target user according to the usage behavior data; calculating the posterior probability of the target user indulging the application according to the generated node characteristics and the prior probability in the application indulging recognition model; taking the posterior probability of the target user indulging the application as an indulging risk index of the target user to the application;
and the judging module 400 is used for judging whether the target user is indulged in the application according to the indulgence risk index.
Specifically, a bayesian network is established, and first, network nodes are determined. The network node comprises the characteristics of long accumulated use time, more application starting times, sensitive time use, sensitive application use and the like in multiple dimensions.
Acquiring application use behavior data of a plurality of users as sample data. And acquiring feature sample data of corresponding dimensions from the sample data.
Application usage behavior data includes, but is not limited to: application attributes (e.g., system application or third party application, learning tool/life convenience/game/social entertainment/music story, etc.), usage time of the application (e.g., weekday/weekday, work period/lesson period, etc.), number of times of launch of the application (e.g., total number of application clicks during a day), usage duration of the application (e.g., cumulative usage duration, usage duration for a time period), application power consumption (e.g., power consumption for the entire day or time period of the application), application disable period, etc.
And performing more fine division and normalization processing on the feature sample data of each dimension, and unifying the data of different dimensions to the same dimension and range. By carrying out equalization processing on the multidimensional data source, the equalization of the sample weight of each dimension in the whole model is ensured. The discrimination of the model for various situations is enhanced by finer partitioning.
Optionally, finely dividing each type of feature sample data according to a preset rule of the corresponding dimension and converting the feature sample data into an addiction value of the corresponding dimension; wherein the range of addict values of each dimension is uniform.
And determining the relation of each node in the network and the conditional probability distribution parameter of each node according to the feature sample data normalized by each dimension. And when the relation of each node is determined, determining the network structure, and when the conditional probability distribution parameter of each node is determined, determining the network parameter. The Bayesian network with determined structure and parameters is used as an application addiction recognition model.
It is assumed that the bayesian network shown in fig. 3 is obtained according to the feature sample data of each dimension. Wherein, the node y represents the application indulgence, and the probability is represented by P (y).
The node β indicates that the cumulative use time is long P2 indicates that the sensitive time is indulged, and the probability that the user is indulged in the application due to the long cumulative use time is indicated by P (β | y).
Node θ represents a usage sensitive application. P3 denotes the addiction of the sensitive application, and the probability of the user being addicted to the application for using the sensitive application is denoted by P (θ | y).
The node α represents the sensitive time usage, P1 represents the probability of the sensitive time being in the state of being long in the sensitive time period, and the probability is represented by P (α | β, y) because the combination of the accumulated usage time and the sensitive time usage may cause the sensitive time to be in the state of being long in the sensitive time period and the sensitive application may cause the sensitive time to be in the state of being in the sensitive time period.
Through sample data analysis, hidden associations exist between the accumulated use time and the starting times and between the sensitive application types and the starting times, the accumulated use time is long, the sensitive application types possibly cause the starting times to be large, P4 indicates a sensitive number addiction, namely the addiction with the large starting times, and the probability of the addiction is represented by P (gamma | β, theta).
As shown in fig. 3, the node γ has a parent-child relationship with the node β and the node θ, respectively, and the node α has a parent-child relationship with the node β and the node y, respectively.
And acquiring application use behavior data of the target user in a period of time, such as a day, a week or a month, and acquiring the use behavior data of each application. Analyzing the usage behavior data of the target user for each application, for example, the usage behavior data of the application a in units of days, and statistically acquiring average daily cumulative usage time, average distribution of daily usage time periods, average daily startup times, and the like.
And then analyzing whether the node characteristics occur from each node dimension, for example, judging whether the node characteristics are used in sensitive time according to the distribution of the average use time period per day, judging whether the application starting times are large according to the average starting times per day, and the like.
The sampling data of each node dimension can be converted into corresponding addiction values, that is, the addiction values of the node β, the node theta, the node α and the node gamma are respectively calculated according to the application behavior data of the target user, and then whether the corresponding node characteristics occur or not is respectively judged according to the addiction values of the nodes.
For example, if the time period information of the application used by the user is not collected, it cannot be judged whether the user uses the application in a sensitive time period, that is, whether the characteristic of the node α occurs or not, and the bayesian network has the advantage that reasonable reasoning can be carried out on the incomplete information based on the prior probability and the occurred information.
For example, when the features of the nodes β, theta, gamma and α all have occurred, the probability of indulging the application is P (y | β, theta, gamma, α) ═ P (y) 2P 3P 1P 4 ═ P (y) P (β | y) | P (θ | y) × P (α | β, y) × P (γ | β, theta).
And taking the posterior probability as an indulging risk index of the target user to the application, and judging whether the target user indulges the application according to the index.
Fig. 3 is only one example. Nodes may be reduced or increased on the network structure described above. Adding a node is equivalent to introducing new features, such as whether an application is a management application, a large power consumption ratio of the application, and the like. The management and control application comprises management and control application recommended by an official party and management and control application customized by parents. Governing applications are often prone to people becoming addicted, typically setting an application disable period. If the user uses the application during the suggested application disable period, it is indicative of a high user enthusiasm for the application. If a user is indulged in an application, the power consumption duty of the application is usually large. Therefore, the new features are also beneficial to applying enthusiasm recognition, and the addition of the new features can further improve the accuracy of recognition.
In the embodiment, by introducing the multi-dimensional node characteristics (such as long accumulated use time, multiple application starting times, sensitive time use, sensitive application use and the like) into the bayesian network and acquiring the application use behavior data of the multi-dimensional characteristics, the defect that the traditional method cannot accurately identify the addiction due to insufficient characteristics is effectively overcome, and the accuracy of identifying the application addiction is further improved.
In another embodiment of the present invention, as shown in fig. 6 and 7, a device for recognizing an application addiction, includes:
the model building module 100 is used for building an application addiction recognition model based on the Bayesian network;
wherein, the model building module 100 includes:
a sample acquiring unit 110, configured to acquire application usage behavior data of multiple users, attribute information of application users, and/or feedback and complaint information of the users to the application as sample data;
a data preprocessing unit 120, configured to extract feature sample data corresponding to the bayesian network node from the application usage behavior data; carrying out normalization processing on each type of feature sample data;
the model construction unit 130 is configured to determine a structure and parameters of the bayesian network according to feature sample data normalized by each dimension; and obtaining the prior knowledge of the application addiction recognition model according to the attribute information of the application user and/or the feedback and complaint information of the user to the application.
A data obtaining module 200, configured to obtain usage behavior data of an application by a target user;
the addiction risk calculation module 300 is configured to obtain node features that have occurred to the target user according to the usage behavior data; calculating the posterior probability of the target user indulging the application according to the generated node characteristics and the prior probability in the application indulging recognition model; adjusting the posterior probability of the target user addicting the application according to the prior knowledge of the application addict recognition model; taking the adjusted posterior probability as an indulging risk index of the target user to the application;
and the judging module 400 is used for judging whether the target user is indulged in the application according to the indulgence risk index.
Specifically, the difference from the previous embodiment is:
when the application addiction recognition model is constructed, the acquired sample data also comprises attribute information of an application user and/or feedback and complaint information of the user to the application. Obtaining the prior knowledge of the application addiction recognition model according to the attribute information of the application user and/or the feedback and complaint information of the user to the application; adjusting the posterior probability of the target user addicting the application according to the prior knowledge of the application addict recognition model; and taking the adjusted posterior probability as an indulging risk index of the target user to the application.
The attribute information of the application user includes the age, sex, region, etc. of the user. The feedback of the user to the application and the source of the complaint information are not limited to e-commerce comments, customer service feedback/complaint information and the like.
The judgment criteria of the application addiction by the users with different age groups/different sexes/different regions are different, for example, the middle school students are addicted to play games for more than 1 hour, and the preschool children are likely to be addicted to use continuously for more than 30 minutes. It is desirable to reflect this difference in the indulgence risk index.
The present embodiment embodies the above difference by introducing a priori knowledge λ. And adjusting the calculated posterior probability of the addiction application of the target user according to the prior knowledge lambda so as to obtain a more accurate addiction risk index. For example, considering different user types, the a priori knowledge λ is: for junior middle school and above, λ ═ 1; below junior school, λ is 1.5. Assuming that the posterior probability calculated based on the bayesian network is P, if the target user is a pupil, the addiction risk index is adjusted to 1.5P.
The priori knowledge lambda can be obtained by statistical data analysis, for example, the same application enthrallment distribution applied to different age groups, different sexes and different regions can be obtained; if λ of one age group is set to 1, λ of other age groups can be obtained according to the applied addiction distribution. Similarly, the prior knowledge λ may also be corrected based on the review feedback.
The embodiment is applied to enthusiasm use recognition of APP software on a child watch. And obtaining application use behavior data of the user through APP feedback data, and obtaining user feedback information through the parent complaint information received by customer service and the E-commerce comment information of the application.
The application use behavior data fed back by the child watches are collected, the indulging risk index of each user to each application is dynamically analyzed according to the application use behavior data of each user, and the related suggestions are fed back to parents according to the indulging risk indexes, so that the parents can better control the applications on the child watches, for example, which applications are set to be forbidden, forbidden time periods and the like.
In the embodiment, the prior knowledge is introduced, the posterior probability calculated by the Bayesian network is corrected, and the accuracy of identifying the application addiction is further improved.
It is understood that the functional modules (or units or components) adopted by the device for identifying application addiction can be in one-to-one correspondence with the steps of the method for identifying application addiction, which are described in the foregoing embodiments, so as to implement the steps. Accordingly, the specific names and functional divisions of the functional modules of the recognition device with application addiction are selected according to needs, and can be correspondingly combined, split, recombined or changed.
It should be noted that the above embodiments can be freely combined as necessary. The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A method for identifying an application addiction, comprising:
constructing an application addiction recognition model based on the Bayesian network;
acquiring the use behavior data of a target user to the application;
obtaining an addiction risk index of the target user to the application according to the using behavior data and the application addiction recognition model;
and judging whether the target user is addicted to the application or not according to the addicted risk index.
2. The method of identifying an application addiction according to claim 1, wherein:
the nodes of the bayesian network comprise at least two of the following features: the method has the advantages of long accumulated use time, more application starting times, use of sensitive time, use of sensitive applications, large power consumption ratio of the applications and whether the applications are management and control applications.
3. The method for identifying application addiction according to claim 2, wherein the constructing of the application addiction identification model based on the Bayesian network comprises:
acquiring application use behavior data of a plurality of users as sample data;
extracting feature sample data corresponding to the Bayesian network node from the sample data;
carrying out normalization processing on each type of feature sample data;
and determining the structure and parameters of the Bayesian network according to the feature sample data normalized by each dimension.
4. The method for identifying the application addiction according to claim 3, wherein the obtaining the addiction risk index of the target user to the application according to the usage behavior data and the application addiction identification model comprises:
acquiring node characteristics of the target user according to the using behavior data;
calculating the posterior probability of the target user indulging the application according to the generated node characteristics and the prior probability in the application indulging recognition model;
and taking the posterior probability of the target user indulging the application as the indulging risk index of the target user to the application.
5. The method of identifying an application addiction according to claim 4, wherein:
the sample data also comprises attribute information of an application user;
and obtaining the prior knowledge of the application addiction recognition model according to the attribute information of the application user.
6. The method of identifying an application addiction according to claim 5, wherein:
the sample data also comprises feedback and complaint information of the user to the application;
the obtaining of the prior knowledge of the application addiction recognition model according to the attribute information of the application user comprises:
and obtaining the prior knowledge of the application addiction recognition model according to the attribute information of the application user and the feedback and complaint information of the user to the application.
7. The method for identifying application addiction according to claim 6, wherein the step of using the posterior probability of the target user addiction to the application as the addiction risk index of the target user to the application comprises:
adjusting the posterior probability of the target user addicting the application according to the prior knowledge of the application addict recognition model;
and taking the adjusted posterior probability as an indulging risk index of the target user to the application.
8. An apparatus for recognizing an application addiction, comprising:
the model construction module is used for constructing an application addiction recognition model based on the Bayesian network;
the data acquisition module is used for acquiring the use behavior data of the target user to the application;
the addiction risk calculation module is used for obtaining an addiction risk index of the target user to the application according to the use behavior data and the application addiction recognition model;
and the judging module is used for judging whether the target user is addicted to the application according to the addict risk index.
9. The device for identifying the application addiction according to claim 8, wherein the model building module comprises:
the system comprises a sample acquisition unit, a data processing unit and a data processing unit, wherein the sample acquisition unit is used for acquiring application use behavior data of a plurality of users as sample data;
the data preprocessing unit is used for extracting feature sample data corresponding to the Bayesian network node from the sample data; carrying out normalization processing on each type of feature sample data;
and the model construction unit is used for determining the structure and the parameters of the Bayesian network according to the feature sample data normalized by each dimension.
10. The device for identifying an application addiction according to claim 9, wherein:
the addiction risk calculation module is further used for acquiring node characteristics of the target user according to the using behavior data; calculating the posterior probability of the target user indulging the application according to the generated node characteristics and the prior probability in the application indulging recognition model; and taking the posterior probability of the target user indulging the application as the indulging risk index of the target user to the application.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010098424.1A CN111324509B (en) | 2020-02-18 | 2020-02-18 | Identification method and device for application addiction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010098424.1A CN111324509B (en) | 2020-02-18 | 2020-02-18 | Identification method and device for application addiction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111324509A true CN111324509A (en) | 2020-06-23 |
| CN111324509B CN111324509B (en) | 2023-07-11 |
Family
ID=71163434
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010098424.1A Active CN111324509B (en) | 2020-02-18 | 2020-02-18 | Identification method and device for application addiction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111324509B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112451973A (en) * | 2020-12-17 | 2021-03-09 | 徐岩 | Identity authentication system relating to specific network user |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020099594A1 (en) * | 2000-05-26 | 2002-07-25 | Nicholas Heard | Method and apparatus for determining one or more statistical estimators of customer behavior |
| CN103927675A (en) * | 2014-04-18 | 2014-07-16 | 北京京东尚科信息技术有限公司 | Method and device for judging age brackets of users |
| CN104703195A (en) * | 2015-02-16 | 2015-06-10 | 南京邮电大学 | Method for predicating behavior of mobile self-networking routing node |
| US20160078365A1 (en) * | 2014-03-21 | 2016-03-17 | Philippe Baumard | Autonomous detection of incongruous behaviors |
| US20170063897A1 (en) * | 2015-08-31 | 2017-03-02 | Splunk Inc. | Interface Providing An Interactive Timeline For Evaluating Instances Of Potential Network Compromise |
| CN106708715A (en) * | 2015-07-13 | 2017-05-24 | 阿里巴巴集团控股有限公司 | Automatic test method and device |
| CN109376784A (en) * | 2018-10-29 | 2019-02-22 | 四川大学 | A kind of personality prediction technique and personality prediction meanss |
| CN109858970A (en) * | 2019-02-02 | 2019-06-07 | 中国银行股份有限公司 | A kind of user's behavior prediction method, apparatus and storage medium |
| CN110193199A (en) * | 2019-05-26 | 2019-09-03 | 韶关市启之信息技术有限公司 | A kind of method and device of game interruption-free |
| CN110543904A (en) * | 2019-08-26 | 2019-12-06 | 河北航天信息技术有限公司 | Enterprise risk classification model construction method based on Bayes |
| CN110620760A (en) * | 2019-07-30 | 2019-12-27 | 东南大学 | FlexRay bus fusion intrusion detection method and detection device for SVM (support vector machine) and Bayesian network |
-
2020
- 2020-02-18 CN CN202010098424.1A patent/CN111324509B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020099594A1 (en) * | 2000-05-26 | 2002-07-25 | Nicholas Heard | Method and apparatus for determining one or more statistical estimators of customer behavior |
| US20160078365A1 (en) * | 2014-03-21 | 2016-03-17 | Philippe Baumard | Autonomous detection of incongruous behaviors |
| CN103927675A (en) * | 2014-04-18 | 2014-07-16 | 北京京东尚科信息技术有限公司 | Method and device for judging age brackets of users |
| CN104703195A (en) * | 2015-02-16 | 2015-06-10 | 南京邮电大学 | Method for predicating behavior of mobile self-networking routing node |
| CN106708715A (en) * | 2015-07-13 | 2017-05-24 | 阿里巴巴集团控股有限公司 | Automatic test method and device |
| US20170063897A1 (en) * | 2015-08-31 | 2017-03-02 | Splunk Inc. | Interface Providing An Interactive Timeline For Evaluating Instances Of Potential Network Compromise |
| CN109376784A (en) * | 2018-10-29 | 2019-02-22 | 四川大学 | A kind of personality prediction technique and personality prediction meanss |
| CN109858970A (en) * | 2019-02-02 | 2019-06-07 | 中国银行股份有限公司 | A kind of user's behavior prediction method, apparatus and storage medium |
| CN110193199A (en) * | 2019-05-26 | 2019-09-03 | 韶关市启之信息技术有限公司 | A kind of method and device of game interruption-free |
| CN110620760A (en) * | 2019-07-30 | 2019-12-27 | 东南大学 | FlexRay bus fusion intrusion detection method and detection device for SVM (support vector machine) and Bayesian network |
| CN110543904A (en) * | 2019-08-26 | 2019-12-06 | 河北航天信息技术有限公司 | Enterprise risk classification model construction method based on Bayes |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112451973A (en) * | 2020-12-17 | 2021-03-09 | 徐岩 | Identity authentication system relating to specific network user |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111324509B (en) | 2023-07-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103686237B (en) | Recommend the method and system of video resource | |
| CN112199608B (en) | Social media rumor detection method based on network information propagation graph modeling | |
| US20220172090A1 (en) | Data identification method and apparatus, and device, and readable storage medium | |
| CN104133817A (en) | Online community interaction method and device and online community platform | |
| CN103647800A (en) | Method and system of recommending application resources | |
| CN103198086A (en) | Information processing device, information processing method, and program | |
| CN103647671A (en) | Gur Game based crowd sensing network management method and system | |
| CN107679920A (en) | The put-on method and device of a kind of advertisement | |
| CN111510368B (en) | Family group identification method, device, equipment and computer readable storage medium | |
| CN106528850A (en) | Door access data abnormal detection method based on machine learning clustering algorithm | |
| Gokulkumari et al. | Analyze the political preference of a common man by using data mining and machine learning | |
| CN111324509A (en) | Method and device for identifying application addiction | |
| CN114024737A (en) | Method and device for determining brushing amount of live broadcast room | |
| CN112560105A (en) | Joint modeling method and device for protecting multi-party data privacy | |
| CN109992592B (en) | College poverty and poverty identification method based on flow data of campus consumption card | |
| CN110210884B (en) | Method, device, computer equipment and storage medium for determining user characteristic data | |
| Oliveira et al. | Long-term characterization of political communications on social media | |
| CN113850350B (en) | Urban building land intelligent planning system and method | |
| CN113254788A (en) | Big data based recommendation method and system and readable storage medium | |
| CN112100294A (en) | User relationship analysis method and device for network platform and related equipment | |
| Kelleher et al. | Finding common ground for citizen empowerment in the smart city | |
| Dhiman | Attitude of Parents towards Schooling of their Children | |
| Ankoju et al. | Sampling-based probability construction explains individual differences in risk preference | |
| Li et al. | Data driven user feature construction and requirement generation in mobile applications | |
| CN106228394A (en) | A kind of information mean analysis system based on big data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |