Network space addiction behavior index system construction method and device, equipment and medium

Abstract

The invention belongs to the field of network addiction detection, and provides a network space addiction behavior index system construction method and device, equipment and a medium, and the method comprises the steps: obtaining an index and an addiction level of a network space addiction behavior, and determining a visual angle; performing two-point anchor point selection processing on the visual angle to obtain an anchor point set; according to the anchor point set, a similarity graph, a diversity graph and a consistency graph of the anchor points are constructed, and the similarity graph, the diversity graph and the consistency graph are used for representing the incidence relation of the view angles; performing fusion calculation on the diversity graph and the consistency graph of the anchor points to obtain a weighted similarity matrix; carrying out eigendecomposition on the weighted similarity matrix to obtain low-dimensional embedded representation, and carrying out supervised training on the low-dimensional embedded representation through a loss function to obtain a network space addiction behavior index system; and network addiction detection is carried out through the network space addiction behavior index system. According to the invention, the accuracy of user network addiction level identification is improved.

Description

Technical Field

[0001] This invention relates to the field of internet addiction detection technology, and in particular to a method, apparatus, equipment and medium for constructing an index system for internet addiction behavior. Background Technology

[0002] With the rapid development of digital technology and the internet, internet users' online behavior has gradually become a focus of social attention. Individual user behavior online, including online time, frequency, and types of content accessed, has become an important indicator for measuring the degree of internet addiction. However, due to the openness and convenience of the internet, users are prone to internet addiction, seriously affecting their physical and mental health, daily life, and social skills. Therefore, how to effectively assess internet users' internet addiction behavior and take corresponding intervention measures according to different levels of addiction has become an urgent social problem to be solved.

[0003] Currently, existing methods for assessing internet addiction primarily rely on single-perspective data analysis, such as online time, access to games or videos, etc. While these methods can provide some assessment results, the multi-dimensional nature of internet addiction behavior means that analyzing only one dimension can lead to incomplete results or even misjudgments. Most existing assessment systems do not consider physiological and psychological factors, and neglect the diversity and complexity of data. Summary of the Invention

[0004] Aimed at at least in solving one of the technical problems existing in the prior art, the present invention provides a method, apparatus, device and medium for constructing an index system for internet addiction behavior, thereby improving the accuracy of identifying the level of user internet addiction.

[0005] One aspect of the present invention provides a method for constructing an indicator system for cyberspace addiction behavior, comprising:

[0006] To obtain indicators and levels of internet addiction behavior, including both qualitative and quantitative data.

[0007] Qualitative and quantitative data are preprocessed to obtain perspectives, which are used to characterize indicators of different levels of addiction.

[0008] Perform two-point anchor point selection processing on the viewpoint to obtain the anchor point set;

[0009] Based on the set of anchor points, a similarity graph, a diversity graph, and a consistency graph of the anchor points are constructed, wherein the similarity graph, the diversity graph, and the consistency graph are used to represent the relationship between viewpoints;

[0010] The diversity map and consistency map of the anchor points are fused and calculated to obtain a weighted similarity matrix;

[0011] The weighted similarity matrix is ​​decomposed into features to obtain a low-dimensional embedding representation. The low-dimensional embedding representation is then trained under supervision using a loss function to obtain a classifier and classifier weights, thereby obtaining a network space addiction behavior index system.

[0012] By obtaining the perspective of the target users, and using the cyberspace addiction behavior indicator system to analyze the perspective of the target users, the detection results of cyberspace addiction behavior are obtained.

[0013] According to the method for constructing the indicator system for cyberspace addiction behavior, the qualitative and quantitative data are preprocessed to obtain perspectives, including:

[0014] The qualitative data is standardized, the quantitative data is converted using one-hot encoding, and missing and outlier values ​​in both the qualitative and quantitative data are processed to obtain a perspective.

[0015] According to the method for constructing the indicator system for cyberspace addiction behavior, a two-point anchor point selection process is performed on the perspective to obtain a set of anchor points, including:

[0016] Based on the dataset, the dataset is divided into multiple first subsets by global anchor points, and the dataset includes multi-dimensional behavioral data, physiological data and psychological data.

[0017] The first subset is further divided using local anchor points to obtain the second subset;

[0018] The global anchor points and local anchor points are merged to obtain the anchor point set.

[0019] According to the method for constructing the indicator system for cyberspace addiction behavior, the method further includes:

[0020] Based on each perspective, the dataset is divided into multiple sub-regions using a spatial partitioning method, resulting in multiple first subsets:

[0021]

[0022] in, It is the perspective after division The The first subset;

[0023] The global anchor point is determined using the maximum-minimum distance method in each first subset. for:

[0024]

[0025] in, Represents data points in the first subset and The distance between them;

[0026] With global anchor Using the first subset as the initial seed point, the first subset is partitioned into multiple second subsets using Voronoi partitioning. ,in Indicates proximity to the first A set of data points for a global anchor point;

[0027] K-means clustering is applied to the second subset, and the center point of each K-means cluster is used as the local anchor point.

[0028] According to the method for constructing an indicator system for internet addiction behavior, a similarity graph, a diversity graph, and a consistency graph of the anchor points are constructed based on the anchor point set. These graphs are used to characterize the relationships between viewpoints, including:

[0029] The similarity graph of anchor points is determined using Gaussian kernel similarity, where Gaussian kernel similarity... for:

[0030]

[0031] in, Represents data points in the anchor set With data points Euclidean distance, These are parameters used to control the scale of similarity;

[0032] An adaptive neighbor graph learning method is used to normalize the similarity graph of anchor points, resulting in the normalized anchor point result. for:

[0033]

[0034] in, Data points The set of neighbors;

[0035] Based on the anchor point similarity diagram and normalization results Determine the diversity and consistency plots of the anchor points:

[0036]

[0037]

[0038] in, For diversity diagrams, It's a perspective Other perspectives Measurement of differences between them; For consistency graph, It is the average similarity value across all perspectives;

[0039] Calculate the consistency graph Laplace matrix And, for consistency graphs Apply Laplace manifold constraints:

[0040]

[0041] in, The degree matrix is ​​calculated using the following formula:

[0042]

[0043] The Laplace manifold constraint is as follows:

[0044]

[0045] in, and For consistency graph Representation of different data points in low-dimensional space.

[0046] According to the method for constructing the indicator system for cyberspace addiction behavior, the diversity graph and consistency graph of anchor points are fused and calculated to obtain a weighted similarity matrix, including:

[0047] By organizing the diversity and consistency graphs from each perspective into a tensor representation, a third-order tensor is obtained. :

[0048]

[0049] For third-order tensors Applying tensor nuclear norm constraints The optimized third-order tensor is obtained. :

[0050]

[0051] in, These are the singular values ​​after tensor pattern decomposition, where the optimization objective of the tensor nuclear norm constraint is:

[0052]

[0053] in, Controlling the low-rank property of tensors; Used to control the reconstruction error for each viewpoint; For hyperparameters;

[0054] Based on the optimized third-order tensor Expand the tensor using mode-1. Transform into a matrix Then, the multi-view weighted similarity matrix is ​​calculated.

[0055]

[0056]

[0057] in, These are weight parameters. Based on the optimized third-order tensor The calculated similarity matrix It is the similarity matrix of viewpoint v. It is the perspective weight, where , This represents the Frobenius norm.

[0058] According to the method for constructing the index system of cyberspace addiction behavior, the weighted similarity matrix is ​​decomposed to obtain a low-dimensional embedding representation. This low-dimensional embedding representation is then trained under supervision using a loss function to obtain a classifier and classifier weights, thereby obtaining the index system of cyberspace addiction behavior, including:

[0059] Calculate the weighted similarity matrix of multiple views Laplace matrix :

[0060]

[0061] in, This is a degree matrix, representing the total similarity of each data point, and its calculation formula is:

[0062]

[0063] Eigenvalue decomposition is performed on the Laplacian matrix to obtain the representation of data points in a low-dimensional space. The eigenvalue decomposition formula is as follows:

[0064]

[0065] in, It is a feature vector, used to represent the projection of data points into a low-dimensional space; These are the corresponding feature values, used to measure the degree of dispersion of data points;

[0066] Take the eigenvectors corresponding to the first t smallest eigenvalues ​​as the data representation after dimensionality reduction. Based on expert knowledge and labeled data, the data representation after dimensionality reduction is improved by minimizing the cross-entropy loss function. Use a classifier to perform classification, and obtain the classifier and classifier weights.

[0067] Another aspect of the present invention provides a device for constructing an indicator system for cyberspace addiction behavior, comprising:

[0068] The first module is used to obtain indicators and addiction levels of cyberspace addiction behavior, where the indicators include qualitative and quantitative data.

[0069] The second module is used to preprocess qualitative and quantitative data to obtain perspectives, which are used to characterize indicators of different levels of addiction.

[0070] The third module is used to perform two-point anchor point selection processing on the viewpoint to obtain a set of anchor points;

[0071] The fourth module is used to construct a similarity graph, a diversity graph, and a consistency graph of the anchor points based on the set of anchor points, wherein the similarity graph, the diversity graph, and the consistency graph are used to represent the relationship between viewpoints;

[0072] The fifth module is used to fuse the diversity map and consistency map of the anchor points to obtain a weighted similarity matrix;

[0073] The sixth module is used to perform feature decomposition on the weighted similarity matrix to obtain a low-dimensional embedding representation. The low-dimensional embedding representation is then trained under supervision using a loss function to obtain a classifier and classifier weights, thereby obtaining a network space addiction behavior index system.

[0074] The seventh module is used to obtain the perspective of the target user. It uses the cyberspace addiction behavior index system to analyze the perspective of the target user and obtain the detection results of cyberspace addiction behavior.

[0075] Another aspect of the present invention provides an electronic device, including a processor and a memory;

[0076] The memory is used to store programs;

[0077] The processor executes the program to implement the method as described above.

[0078] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the methods described above.

[0079] The beneficial effects of this invention are as follows: by combining online behavioral data, physiological data, psychological data and other multi-dimensional information, it comprehensively assesses users' internet addiction behavior, solves the problem that traditional internet addiction assessment methods rely on data from only a single perspective and improves the accuracy of identifying users' internet addiction levels; by combining global anchor point selection and local anchor point optimization, it ensures that the selected anchor points are both representative and guarantee data diversity, effectively reducing computational complexity and improving computational efficiency in large-scale datasets. Attached Figure Description

[0080] Figure 1 This is a schematic diagram illustrating the construction process of the cyberspace addiction behavior indicator system according to an embodiment of the present invention.

[0081] Figure 2 This is a schematic diagram of the two-point anchor point selection process according to an embodiment of the present invention.

[0082] Figure 3 This is a schematic diagram of a device for constructing an index system for cyberspace addiction behavior according to an embodiment of the present invention. Detailed Implementation

[0083] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings. Throughout the description, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions. In the following description, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" can be used interchangeably. Terms such as "first," "second," etc., are used only to distinguish technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the sequential relationship of the indicated technical features. In the following description, the consecutive reference numerals for method steps are for ease of review and understanding. Adjusting the implementation order of steps, in conjunction with the overall technical solution of the present invention and the logical relationship between the various steps, will not affect the technical effect achieved by the technical solution of the present invention. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0084] refer to Figure 1 , Figure 1 The method for constructing an indicator system for internet addiction behavior includes, but is not limited to, steps S100 to S600:

[0085] S100 obtains indicators and addiction levels of cyberspace addiction behavior, including qualitative and quantitative data.

[0086] In some embodiments, the metrics are referenced as shown in Table 1 below:

[0087] Table 1. 22 Survey Indicators

[0088] In some embodiments, the perspective is different levels of indicator data.

[0089] In some embodiments, addiction levels include addictive, mild, moderate, high, severe, and extreme.

[0090] S200 preprocesses qualitative and quantitative data to obtain perspectives, which are used to characterize indicators of different levels of addiction.

[0091] In some embodiments, qualitative data is standardized, quantitative data is converted using one-hot encoding, and missing and outlier values ​​in the qualitative and quantitative data are processed to obtain a perspective.

[0092] In some embodiments, quantitative data (such as internet usage time, blink rate, heart rate, etc.) are standardized to eliminate differences in units of measurement and ensure that different data types have the same scale so that they are not affected by different units of measurement during subsequent model training. The specific standardization formula is as follows:

[0093]

[0094] in, This represents the raw quantitative data. This represents the mean of the quantitative indicator. This represents the standard deviation of the quantitative indicator.

[0095] In some embodiments, qualitative data (such as gender, eye gaze, head posture, etc.) are transformed using one-hot encoding to convert categorical data into numerical data for subsequent input into the model for training. For one-hot encoding, if a qualitative indicator has k categories, the encoding transformation formula for that indicator is:

[0096]

[0097] Among them, the category to which the qualitative indicator belongs is 1, and the rest are 0.

[0098] In some embodiments, missing and outlier values ​​in the data are identified and processed to ensure data integrity and quality, thereby improving the effectiveness of model training. Specifically, for missing values, the mean is used for imputation; for outliers, box plots are used to mark data points exceeding the boundaries as outliers by calculating the quartiles and upper and lower boundaries of the data, and then processed (such as by replacement or deletion).

[0099] S300 performs two-point anchor point selection processing on the viewpoint to obtain a set of anchor points.

[0100] In some embodiments, reference Figure 2 The flowchart shown below illustrates the selection process for two anchor points, which includes, but is not limited to, steps S310 to S330:

[0101] S310, Based on the dataset, the dataset is divided into multiple first subsets by global anchor points, where the dataset includes multi-dimensional behavioral data, physiological data and psychological data;

[0102] In some embodiments, the data in each viewpoint is divided into multiple subsets by a global anchor point selection method to ensure the representativeness and diversity of anchor points, providing high-quality input for subsequent local anchor point optimization and graph construction.

[0103] In some embodiments, the dataset is divided into multiple sub-regions using a spatial partitioning method according to each perspective, resulting in multiple first subsets:

[0104]

[0105] in, It is the perspective after division The The first subset;

[0106] The global anchor point is determined using the maximum-minimum distance method in each first subset. for:

[0107]

[0108] in, Represents data points in the first subset and The distance between them.

[0109] S320, the first subset is further divided using local anchor points to obtain the second subset.

[0110] In some embodiments, based on the selected global anchor points, the anchor point selection is further optimized within each local subset, making the anchor points more consistent with the local structure of the data and improving the quality of the final representative anchor point set.

[0111] In some embodiments, with global anchor points Using the first subset as the initial seed point, the first subset is partitioned into multiple second subsets using Voronoi partitioning. ,in Indicates proximity to the first The set of data points for the first global anchor point; K-means clustering is applied to the second subset, and the center point of each K-means cluster is used as the local anchor point to make the selected anchor points more representative.

[0112] S330 merges global anchor points and local anchor points to obtain an anchor point set.

[0113] It should be noted that the dataset for the cyberspace addiction behavior indicator system mainly reflects the following two aspects:

[0114] (1) Large sample size: In the application scenarios of this invention, the user sample size can reach thousands to hundreds of thousands, and each user contains multi-dimensional behavioral data, physiological data, psychological data, and other multi-perspective information. Traditional methods based on full sample similarity calculation or clustering typically have a time complexity of O(n). The computational cost is extremely high when the sample size is large.

[0115] (2) High feature dimension and diverse perspectives: The original feature dimension of each sample can exceed 100 dimensions and includes 4 to 6 independent perspectives (behavior, psychology, physiology, real performance, calibration data, etc.), which leads to an exponential increase in the complexity of multi-view similarity calculation and graph construction.

[0116] This invention, through its proposed anchor selection strategy—namely, first employing global anchor selection to ensure the representativeness of the sample space, and then optimizing local anchors within local subsets—can compress the original sample size from n to the anchor size m (where m ≪ n). The computational complexity is effectively reduced to .

[0117] S400, Based on the set of anchor points, construct a similarity graph, a diversity graph, and a consistency graph of the anchor points, wherein the similarity graph, the diversity graph, and the consistency graph are used to represent the relationship between viewpoints.

[0118] In some embodiments, Gaussian kernel similarity is used to determine the similarity map of anchor points, wherein Gaussian kernel similarity... for:

[0119]

[0120] In this embodiment of the invention, the Gaussian kernel similarity between data points is calculated, and only the similarity value of the data points in their k nearest neighbors is retained, while other values ​​are set to zero, in order to reduce noise.

[0121] in, Represents data points in the anchor set With data points Euclidean distance, These are parameters used to control the scale of similarity;

[0122] An adaptive neighbor graph learning method is used to normalize the similarity graph of anchor points, resulting in the normalized anchor point result. for:

[0123]

[0124] in, Data points The set of neighbors;

[0125] Based on the anchor point similarity diagram and normalization results Determine the diversity and consistency plots of the anchor points:

[0126]

[0127]

[0128] in, For diversity diagrams, It's a perspective Other perspectives Measurement of differences between them; For consistency graph, It is the average similarity value across all perspectives;

[0129] Calculate the consistency graph Laplace matrix And, for consistency graphs Apply Laplace manifold constraints:

[0130]

[0131] in, The degree matrix is ​​calculated using the following formula:

[0132]

[0133] Wherein, the Laplace manifold constraint F is:

[0134]

[0135] in, and For consistency graph The representation of different data points in low-dimensional space, and the dimensional representation obtained by Laplace manifold constraints. , equivalent to , It is the representation of data points in a low-dimensional space. It is a Laplace regularization term used to ensure that the local structure of data points remains unchanged during dimensionality reduction.

[0136] S500 performs a fusion calculation on the diversity map and consistency map of the anchor points to obtain a weighted similarity matrix.

[0137] In some embodiments, the diversity and consistency graphs for each viewpoint are organized into tensor representations to preserve multi-view information and are jointly processed during optimization. Specifically, a third-order tensor is used. To indicate:

[0138]

[0139] In some embodiments, the constructed multi-view tensor By imposing tensor nuclear norm constraints, low-rank information in the data can be extracted, and redundancy can be removed. The tensor nuclear norm is defined as follows:

[0140]

[0141] in, These are the singular values ​​resulting from tensor pattern decomposition.

[0142] The specific optimization goal is:

[0143]

[0144] Among them, the first item Controlling the low-rank property of the tensor ensures that the tensor after data fusion does not overfit noise information; the second term Control the reconstruction error of each viewpoint data so that the optimized tensor still conforms to the structure of the original data; It is a hyperparameter used to balance the relationship between data fitting and noise reduction.

[0145] In some embodiments, a multi-view weighted similarity matrix is ​​calculated based on the optimized tensor. Specifically, using mode-1 expansion to expand the tensor Transform into a matrix Then, perform similarity calculations based on tensor expansion:

[0146]

[0147]

[0148] in, It is a similarity matrix calculated based on optimized tensors. It is the similarity matrix of viewpoint v. These are perspective weights, reflecting the importance of a perspective within the overall data, and are determined through optimization.

[0149]

[0150] Among them, among them, These are weight parameters. Based on the optimized third-order tensor The calculated similarity matrix It is the similarity matrix of viewpoint v. It is the perspective weight, where , Describing the Frobenius norm

[0151] S600 performs feature decomposition on the weighted similarity matrix to obtain a low-dimensional embedding representation. The low-dimensional embedding representation is then trained under supervision using a loss function to obtain a classifier and classifier weights, thereby obtaining a network space addiction behavior index system.

[0152] In some embodiments, a multi-view weighted similarity matrix is ​​calculated. Laplace matrix This is to facilitate subsequent eigenvalue decomposition. Laplace matrix The specific calculation formula is as follows:

[0153]

[0154] in, This is a degree matrix, representing the total similarity of each data point. Its calculation formula is as follows:

[0155]

[0156] in, Let be the value of the i-th row and i-th column of the degree matrix.

[0157] In some embodiments, eigenvalue decomposition is performed on the Laplacian matrix to obtain the representation of data points in a low-dimensional space. The eigenvalue decomposition formula is as follows:

[0158]

[0159] in, It is an eigenvector, used to represent the projection of data points into a low-dimensional space. These are the corresponding feature values, used to measure the degree of dispersion of data points.

[0160] After obtaining the eigenvalues, the eigenvectors corresponding to the first t smallest eigenvalues ​​are taken as the data representation after dimensionality reduction. These feature vectors constitute the reduced-dimensional data representation space, where similar data points are closer together in the new subspace.

[0161] In some embodiments, based on expert knowledge and manually labeled data, the dimensionality-reduced data representation is optimized by minimizing the cross-entropy loss function. By using a classifier for classification, we can ultimately obtain the weights of the trained classifier, which are the weights of various aspects of the internet addiction behavior data, i.e., the indicator system.

[0162] The S700 acquires the perspective of the target user and analyzes the perspective of the target user using a network addiction behavior indicator system to obtain the detection results of network addiction behavior.

[0163] In some embodiments, a multi-perspective data sample from 500 internet users was selected. In addition to basic data, each user's data included the following four categories:

[0164] 1. Online behavior data: including seven quantitative indicators such as total online time, nighttime online time, online frequency, proportion of game content, proportion of video content, number of devices used, and game spending;

[0165] 2. Physiological data: including three quantitative indicators: average heart rate, blood pressure, and blood oxygen saturation;

[0166] 3. Psychological data: including four indicators: eye gaze pattern, head posture, arousal level, and blink frequency. The first three are qualitative indicators (after coding), and the last one is a quantitative indicator.

[0167] 4. Actual performance data: including three qualitative indicators: task completion, sleep quality, and exercise.

[0168] Some users also underwent offline psychological assessments, obtaining qualitative labels such as "self-control," "social dependence," and "emotional stability," which were used for category labeling in supervised learning of the model.

[0169] After standardization and encoding, the data is processed according to steps 3 to 5 to complete anchor point selection, multi-view map construction and fusion, resulting in a multi-view weighted similarity matrix. .

[0170] After performing eigenvalue decomposition, the embedding vector representation of each user in the low-dimensional space is obtained. .

[0171] The supervised classifier training process then begins:

[0172] (1) Input data: low-dimensional vectors for each user As input samples, the addiction levels were labeled by experts. As a supervisory label.

[0173] (2) Model Structure: In this embodiment, a three-layer fully connected neural network (DNN) is used as the supervised classifier. The input layer has a dimension of 32, the number of hidden layer nodes is 64 and 16 respectively, and the number of output layer nodes is 6, corresponding to the six levels of addiction. The activation function is ReLU, and the output layer uses the Softmax function for multi-class prediction.

[0174] (3) Loss function and optimization algorithm: Cross-entropy loss function is adopted.

[0175]

[0176] in, For real labels, The model is used to predict probabilities. During training, the Adam optimization algorithm is used with a learning rate of 0.001, a batch size of 64, a maximum number of iterations of 200, and an early stopping strategy to prevent overfitting.

[0177] After training, the resulting model weight vector This represents the weighted results of the internet addiction behavior indicator system. Some example weights are shown in Table 2 below:

[0178] Table 2. Indicator System for Internet Addiction Behavior

[0179] As can be seen from the obtained weights, online time, gaming ratio, psychological excitement level, and offline psychological assessment results have the most significant impact on the level of internet addiction. This weighting system constitutes the "Internet Addiction Behavior Index System" proposed in this invention, which can provide quantitative assessment basis for user management systems or relevant departments, and can design targeted strategies based on different weight priorities in subsequent interventions.

[0180] Figure 3 This is a schematic diagram of a device for constructing an index system for cyberspace addiction behavior according to an embodiment of the present invention. The device includes a first module 310, a second module 320, a third module 330, a fourth module 340, a fifth module 350, a sixth module 360, and a seventh module 370.

[0181] The system comprises seven modules: First, it acquires indicators and levels of internet addiction behavior, including qualitative and quantitative data. Second, it preprocesses the qualitative and quantitative data to obtain perspectives, which are used to characterize different levels of addiction. Third, it performs two-point anchor selection on the perspectives to obtain an anchor set. Fourth, it constructs similarity, diversity, and consistency maps of the anchors based on the anchor set, representing the relationships between perspectives. Fifth, it fuses the diversity and consistency maps to obtain a weighted similarity matrix. Sixth, it performs feature decomposition on the weighted similarity matrix to obtain a low-dimensional embedding representation, which is then trained under supervision using a loss function to obtain a classifier and classifier weights, thus deriving the internet addiction behavior indicator system. Seventh, it acquires the perspective of the target user, analyzes this perspective using the internet addiction behavior indicator system, and obtains the internet addiction behavior detection results.

[0182] Exemplarily, with the cooperation of the first, second, third, fourth, fifth, sixth, and seventh modules in the device, the embodiment device can implement any of the aforementioned methods for constructing a network space addiction behavior indicator system. This involves obtaining indicators and addiction levels of network space addiction behavior, where the indicators include qualitative and quantitative data; preprocessing the qualitative and quantitative data to obtain perspectives, where perspectives are used to characterize indicators of different addiction levels; performing two-point anchor selection processing on the perspectives to obtain an anchor set; constructing a similarity graph, diversity graph, and consistency graph of the anchors based on the anchor set, where the similarity graph, diversity graph, and consistency graph are used to characterize the correlation between perspectives; fusing the diversity graph and consistency graph of the anchors to obtain a weighted similarity matrix; performing feature decomposition on the weighted similarity matrix to obtain a low-dimensional embedding representation; supervising training the low-dimensional embedding representation using a loss function to obtain a classifier and classifier weights, thereby obtaining a network space addiction behavior indicator system; obtaining the perspective of the target user; and analyzing the target user's perspective using the network space addiction behavior indicator system to obtain network space addiction behavior detection results. The beneficial effects of this invention are as follows: by combining online behavioral data, physiological data, psychological data and other multi-dimensional information, it comprehensively assesses users' internet addiction behavior, solves the problem that traditional internet addiction assessment methods rely on data from only a single perspective and improves the accuracy of identifying users' internet addiction levels; by combining global anchor point selection and local anchor point optimization, it ensures that the selected anchor points are both representative and guarantee data diversity, effectively reducing computational complexity and improving computational efficiency in large-scale datasets.

[0183] This invention also provides an electronic device, which includes a processor and a memory;

[0184] The memory stores the program;

[0185] The processor executes a program to perform the aforementioned method for constructing an index system for cyberspace addiction behavior; the electronic device has the function of carrying and running the software system for constructing an index system for cyberspace addiction behavior provided in the embodiments of the present invention, such as a personal computer, minicomputer, mainframe, workstation, network or distributed computing environment, standalone or integrated computer platform, or communicating with charged particle tools or other imaging devices, etc.

[0186] This invention also provides a computer-readable storage medium storing a program that is executed by a processor to implement the method for constructing an index system for cyberspace addiction behavior as described above.

[0187] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented in the embodiments of this invention. Alternative embodiments are contemplated, in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0188] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the aforementioned method for constructing an indicator system for internet addiction behavior.

[0189] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, considering the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed in the embodiments of the invention, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0190] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0191] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can include, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0192] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0193] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0194] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0195] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0196] The above is a detailed description of the preferred embodiments of the present invention, but the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method for constructing an indicator system for internet addiction behavior, characterized in that, include: To obtain indicators and levels of internet addiction behavior, including both qualitative and quantitative data. Qualitative and quantitative data are preprocessed to obtain perspectives, which are used to characterize indicators of different levels of addiction. Perform two-point anchor point selection processing on the viewpoint to obtain the anchor point set; Based on the set of anchor points, a similarity graph, a diversity graph, and a consistency graph of the anchor points are constructed, wherein the similarity graph, the diversity graph, and the consistency graph are used to represent the relationship between viewpoints; The diversity map and consistency map of the anchor points are fused and calculated to obtain a weighted similarity matrix; The weighted similarity matrix is ​​decomposed into features to obtain a low-dimensional embedding representation. The low-dimensional embedding representation is then trained under supervision using a loss function to obtain a classifier and classifier weights, thereby obtaining a network space addiction behavior index system. By obtaining the perspective of the target users, and using the cyberspace addiction behavior indicator system to analyze the perspective of the target users, the detection results of cyberspace addiction behavior are obtained.

2. The method for constructing an indicator system for cyberspace addiction behavior according to claim 1, characterized in that, The preprocessing of qualitative and quantitative data to obtain perspectives includes: The qualitative data is standardized, the quantitative data is converted using one-hot encoding, and missing and outlier values ​​in both the qualitative and quantitative data are processed to obtain a perspective.

3. The method for constructing an indicator system for cyberspace addiction behavior according to claim 1, characterized in that, The two-point anchor point selection process performed on the viewpoint yields a set of anchor points, including: Based on the dataset, the dataset is divided into multiple first subsets by global anchor points, and the dataset includes multi-dimensional behavioral data, physiological data and psychological data. The first subset is further divided using local anchor points to obtain the second subset; The global anchor points and local anchor points are merged to obtain the anchor point set.

4. The method for constructing an indicator system for cyberspace addiction behavior according to claim 3, characterized in that, The method further includes: Based on each perspective, the dataset is divided into multiple sub-regions using a spatial partitioning method, resulting in multiple first subsets: ； in, It is the perspective after division The The first subset; The global anchor point is determined using the maximum-minimum distance method in each first subset. for: ； in, Represents data points in the first subset and The distance between them; With global anchor Using the first subset as the initial seed point, the first subset is partitioned into multiple second subsets using Voronoi partitioning. ,in Indicates proximity to the first A set of data points for a global anchor point; K-means clustering is applied to the second subset, and the center point of each K-means cluster is used as the local anchor point.

5. The method for constructing an indicator system for cyberspace addiction behavior according to claim 1, characterized in that, The step involves constructing a similarity graph, a diversity graph, and a consistency graph for the anchor points based on the set of anchor points. These graphs are used to characterize the relationships between viewpoints, including: The similarity graph of anchor points is determined using Gaussian kernel similarity, where Gaussian kernel similarity... for: ； in, Represents data points in the anchor set With data points Euclidean distance, These are parameters used to control the scale of similarity; An adaptive neighbor graph learning method is used to normalize the similarity graph of anchor points, resulting in the normalized anchor point result. for: ； in, Data points The set of neighbors; Based on the anchor point similarity diagram and normalization results Determine the diversity and consistency plots of the anchor points: ； ； in, For diversity diagrams, It's a perspective Other perspectives Measurement of differences between them; For consistency graph, It is the average similarity value across all perspectives; Calculate the consistency graph Laplace matrix And, for consistency graphs Apply Laplace manifold constraints: ； in, The degree matrix is ​​calculated using the following formula: ； The Laplace manifold constraint is as follows: ； in, and For consistency graph Representation of different data points in low-dimensional space.

6. The method for constructing an indicator system for cyberspace addiction behavior according to claim 5, characterized in that, The process of fusing the diversity map and consistency map of the anchor points to obtain a weighted similarity matrix includes: By organizing the diversity and consistency graphs from each perspective into a tensor representation, a third-order tensor is obtained. : ； For third-order tensors Applying tensor nuclear norm constraints The optimized third-order tensor is obtained. : ； in, These are the singular values ​​after tensor pattern decomposition, where the optimization objective of the tensor nuclear norm constraint is: ； in, Controlling the low-rank property of tensors; Used to control the reconstruction error for each viewpoint; For hyperparameters; Based on the optimized third-order tensor Expand the tensor using mode-1. Transform into a matrix Then, the multi-view weighted similarity matrix is ​​calculated. ; ； ； in, These are weight parameters. Based on the optimized third-order tensor The calculated similarity matrix It is the similarity matrix of viewpoint v. It is the perspective weight, where , This represents the Frobenius norm.

7. The method for constructing an indicator system for cyberspace addiction behavior according to claim 6, characterized in that, The weighted similarity matrix is ​​decomposed into features to obtain a low-dimensional embedding representation. This low-dimensional embedding representation is then trained under supervision using a loss function to obtain a classifier and classifier weights, thereby deriving a network space addiction behavior index system, including: Calculate the weighted similarity matrix of multiple views Laplace matrix : ； in, This is a degree matrix, representing the total similarity of each data point, and its calculation formula is: ； Eigenvalue decomposition is performed on the Laplacian matrix to obtain the representation of data points in a low-dimensional space. The eigenvalue decomposition formula is as follows: ； in, It is a feature vector, used to represent the projection of data points into a low-dimensional space; These are the corresponding feature values, used to measure the degree of dispersion of data points; Take the eigenvectors corresponding to the first t smallest eigenvalues ​​as the data representation after dimensionality reduction. Based on expert knowledge and labeled data, the data representation after dimensionality reduction is improved by minimizing the cross-entropy loss function. Use a classifier to perform classification, and obtain the classifier and classifier weights.

8. A device for constructing an indicator system for internet addiction behavior, characterized in that, include: The first module is used to obtain indicators and addiction levels of cyberspace addiction behavior, where the indicators include qualitative and quantitative data. The second module is used to preprocess qualitative and quantitative data to obtain perspectives, which are used to characterize indicators of different levels of addiction. The third module is used to perform two-point anchor point selection processing on the viewpoint to obtain a set of anchor points; The fourth module is used to construct a similarity graph, a diversity graph, and a consistency graph of the anchor points based on the set of anchor points, wherein the similarity graph, the diversity graph, and the consistency graph are used to represent the relationship between viewpoints; The fifth module is used to fuse the diversity map and consistency map of the anchor points to obtain a weighted similarity matrix; The sixth module is used to perform feature decomposition on the weighted similarity matrix to obtain a low-dimensional embedding representation. The low-dimensional embedding representation is then trained under supervision using a loss function to obtain a classifier and classifier weights, thereby obtaining a network space addiction behavior index system. The seventh module is used to obtain the perspective of the target user. It uses the cyberspace addiction behavior index system to analyze the perspective of the target user and obtain the detection results of cyberspace addiction behavior.

9. An electronic device, characterized in that, Including the processor and memory; The memory is used to store programs; The processor executes the program to implement the method for constructing an index system for cyberspace addiction behavior as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The storage medium stores a program, which is executed by a processor to implement the method for constructing an index system for cyberspace addiction behavior as described in any one of claims 1-7.

Images