An internet access quality monitoring system and method for real-time perception of user experience

Abstract

This invention belongs to the field of internet user access experience technology, specifically relating to an internet access quality monitoring system and method for real-time perception of user experience. This invention constructs a diffusion intensity index by using the adjacency matrix change rate and the experience quality index change rate. This index can identify whether network quality anomalies are propagating in the spatial structure, achieving an improvement from single-point anomaly monitoring to group trend monitoring. The network performance sampling frequency is dynamically adjusted according to the diffusion intensity, improving sampling accuracy during anomaly propagation and reducing sampling load in stable conditions. This reduces system resource consumption while ensuring monitoring accuracy. By integrating individual network performance, spatial adjacency relationships, time window fluctuation statistics, and diffusion trend judgment, real-time dynamic perception and adaptive monitoring and control of user experience quality are achieved, resulting in significant improvements in anomaly identification accuracy, group diffusion perception capability, and system resource utilization efficiency.

Description

Technical Field

[0001] This invention belongs to the field of Internet user access experience technology, specifically relating to an Internet access quality monitoring system and method for real-time perception of user experience. Background Technology

[0002] With the rapid development of mobile internet, cloud computing, online video, online gaming, and remote work, users' demands for internet access quality are constantly increasing. Network performance, including latency, jitter, packet loss rate, and throughput, directly impacts user experience in real-world business scenarios. Therefore, continuous monitoring of internet access quality and accurate evaluation of user experience have become crucial technical requirements for operators and network service providers.

[0003] Existing methods for monitoring internet access quality mainly fall into two categories: active network probing and passive statistical analysis. Active probing typically involves periodically sending test packets to obtain metrics such as network latency and packet loss rate. Passive statistical analysis, on the other hand, assesses network operational status by analyzing traffic data reported by network devices or terminal devices. While these methods can reflect network performance to some extent, they mostly focus on performance metrics at the network device or link level, making it difficult to accurately reflect the user experience quality during actual business operations.

[0004] Existing monitoring solutions often focus on single-user or single-point data analysis, lacking the ability to model the spatiotemporal relationships between users. In real-world network environments, multiple users in similar geographical locations or within the same time period may share the same access resources or network paths. When network anomalies occur, they often exhibit group-wide fluctuations. Current technologies do not fully consider the spatiotemporal adjacency relationships between users, making it difficult to promptly identify whether experience anomalies have a spreading trend, and easily leading to delayed responses to group-wide network quality degradation. Regarding data acquisition mechanisms, traditional monitoring systems typically use fixed sampling frequencies for data collection. When the network is operating smoothly, fixed high-frequency acquisition wastes system resources; while when network anomalies spread rapidly, fixed low-frequency acquisition may fail to capture key changes in a timely manner, affecting the real-time performance and accuracy of monitoring. Summary of the Invention

[0005] The purpose of this invention is to provide an Internet access quality monitoring system and method for real-time perception of user experience. By introducing spatiotemporal adjacency structure modeling, multidimensional offset analysis, adaptive weight calculation and diffusion intensity identification mechanism, it can achieve dynamic, refined and intelligent monitoring of user experience quality.

[0006] The specific technical solution adopted by this invention is as follows: A method for real-time sensing of user experience and monitoring of internet access quality includes: Acquire multi-dimensional network performance data during business operations performed by the target user on the terminal device; Obtain the geographic location information and timestamp information of multiple users within a time window, establish user adjacency relationships that meet preset adjacency conditions, and construct a user adjacency matrix based on the user adjacency relationships; Based on the user adjacency matrix and combined with multidimensional network performance data, the spatiotemporal aggregation value corresponding to the target user is obtained; Multiple dimensional offsets are obtained based on the target user's multidimensional network performance data and spatiotemporal aggregation values; Multidimensional network performance data and multiple dimensional offsets are interval-mapped to obtain corresponding dimensional scores, and a multidimensional score vector is constructed. Obtain the covariance matrix of the multidimensional rating vector within a preset statistical period, obtain the weight coefficients of the corresponding dimensions based on the covariance matrix, and generate the experience quality index of the target user. The system obtains the rate of change of the user adjacency matrix and the rate of change of the experience quality index within a continuous time window. Based on the rate of change of the adjacency matrix and the rate of change of the experience quality index, it obtains the diffusion intensity index. Based on the experience quality index and the diffusion intensity index, it adjusts the collection frequency of multidimensional network performance data and generates monitoring result data.

[0007] In a preferred embodiment, multi-dimensional network performance data is obtained during the service operations performed by the target user on the terminal device, including: Pre-set operation trigger conditions in the terminal device. When the target user triggers the preset operation trigger conditions, it is marked as the start time of the business operation. When the end of a business operation is detected, it is marked as the end time of the business operation; Obtain the time window based on the start and end times of the business operation; Acquire multidimensional network performance data of the target user's terminal device operation within the time window. The multidimensional network performance data includes average round-trip latency, latency jitter, packet loss ratio, and throughput.

[0008] In a preferred embodiment, the geographical location information and timestamp information of multiple users within a time window are obtained, and user adjacency relationships that meet preset adjacency conditions are established. A user adjacency matrix is ​​then constructed based on these relationships, including: Within a time window, acquire the geographic location information and corresponding timestamp information reported by multiple user terminal devices, where the geographic location information includes latitude and longitude coordinate data; Get the spatial distance and time difference between any two users; Obtain a preset user adjacency threshold, which includes a preset spatial distance threshold and a preset time threshold. Determine whether the spatial distance and time difference between any two users meet the preset user adjacency threshold. If the spatial distance is less than the preset spatial distance threshold and the time difference is less than the preset time threshold, it is determined that there is an adjacency relationship between the two users; otherwise, it is determined that there is no adjacency relationship. Construct a user adjacency matrix based on the adjacency relationships between any two users. When any two users are adjacent, the corresponding row and column elements in the matrix are 1, and when any two users are not adjacent, the corresponding row and column elements in the matrix are 0.

[0009] In a preferred embodiment, based on the user adjacency matrix and combined with multidimensional network performance data, the spatiotemporal aggregation value corresponding to the target user is obtained, including: Obtain the row number of the target user in the user adjacency matrix; In the user adjacency matrix, find the column index of the row corresponding to the target user where the element has a value of 1, and obtain the set of neighboring users that have an adjacency relationship with the target user; Multiple network performance dimensions are extracted based on multidimensional network performance data, including round-trip time, latency jitter, packet loss, and throughput. Based on each network performance dimension, extract the network performance values ​​corresponding to the adjacent user sets within the same time window; Obtain the spatial distance between each neighboring user in the neighboring user set and the target user, and obtain the corresponding spatial weight based on the corresponding spatial distance; By grouping adjacent users together on the same network performance dimension and combining them with corresponding spatial weights, the spatiotemporal aggregation value of the target user on the corresponding network performance dimension is obtained.

[0010] In a preferred embodiment, multiple dimensional offsets are obtained based on the target user's multidimensional network performance data and spatiotemporal aggregation values, including: Obtain the network performance values ​​and corresponding spatiotemporal aggregation values ​​for each network performance dimension of the target user; Based on the network performance value and spatiotemporal aggregation value corresponding to each network performance dimension of the target user, obtain the dimension offset under the same network performance dimension.

[0011] In a preferred embodiment, multidimensional network performance data and multiple dimensional offsets are interval-mapped to obtain corresponding dimensional scores, and a multidimensional score vector is constructed, including: Obtain the dimension mapping table, which includes multiple network performance value ranges, multiple dimension offset ranges, and a preset score value corresponding to the intersection of each network performance value range and each dimension offset range. Based on the network performance value and corresponding dimension offset for each network performance dimension of the target user, and combined with the dimension mapping table, the corresponding preset score value is obtained and marked as the dimension score of the corresponding network performance dimension. Obtain the preset dimension order and arrange the dimension scores corresponding to each network performance dimension in sequence to generate a multi-dimensional score vector for the target user. The multi-dimensional score vector includes round-trip time dimension score, latency jitter dimension score, packet loss dimension score, and throughput dimension score.

[0012] In a preferred embodiment, the covariance matrix of the multidimensional rating vectors within a preset statistical period is obtained. Based on the covariance matrix, the weight coefficients of the corresponding dimensions are obtained to generate the target user's experience quality index, including: Obtain a preset statistical period, wherein the preset statistical period includes multiple consecutive time windows, and there are no fewer than two consecutive time windows; Within a preset statistical period, multidimensional scoring vectors corresponding to multiple time windows are obtained in chronological order. Based on the score of each dimension in the multidimensional scoring vector, obtain multiple score values ​​corresponding to the corresponding dimension within a preset statistical period to form a score sequence for the corresponding dimension. Obtain the corresponding covariance matrix based on the score sequence for each dimension, and extract the variance value corresponding to the score for each dimension from the covariance matrix; Based on the proportion of the variance value corresponding to each dimension score to the total variance of all dimension scores, obtain the weight coefficient corresponding to each dimension score. Obtain the multi-dimensional rating vector corresponding to the current time window, and obtain the experience quality index of the target user within the current time window based on the rating of each dimension and the corresponding weight coefficient.

[0013] In a preferred embodiment, the rate of change of the user adjacency matrix and the rate of change of the experience quality index within a continuous time window are obtained. A diffusion intensity index is then obtained based on these rates. The collection frequency of multidimensional network performance data is adjusted according to the experience quality index and the diffusion intensity index, and monitoring result data is generated, including: Obtain the user adjacency matrix corresponding to two consecutive time windows; The two user adjacency matrices are compared element by element to obtain the number of changes in the values ​​of the elements in the matrix, and the rate of change of the user adjacency matrix is ​​obtained based on the number of changes in the values ​​of the elements in the matrix. Obtain the experience quality index corresponding to two consecutive time windows; The difference in experience quality is obtained based on the experience quality index corresponding to two consecutive time windows, and the rate of change of the experience quality index is generated. The diffusion intensity index is obtained based on the rate of change of the user adjacency matrix and the rate of change of the experience quality index. Obtain a preset diffusion intensity threshold and determine whether the diffusion intensity index exceeds the preset diffusion intensity threshold; When the diffusion intensity index exceeds the preset diffusion intensity threshold, the collection frequency of multidimensional network performance data is increased. When the diffusion intensity index does not exceed the preset diffusion intensity threshold, the collection frequency of multidimensional network performance data is maintained. Multidimensional network performance data is obtained based on the adjusted acquisition frequency, and monitoring results are generated by combining the experience quality index and diffusion intensity index of the corresponding time window.

[0014] The present invention also provides an Internet access quality monitoring system for real-time perception of user experience, used in the aforementioned Internet access quality monitoring method for real-time perception of user experience, comprising: The network performance module is used to acquire multi-dimensional network performance data during the business operations performed by the target user on the terminal device. The adjacency matrix module is used to obtain the geographical location information and timestamp information of multiple users within a time window, establish user adjacency relationships that meet preset adjacency conditions, and construct a user adjacency matrix based on the user adjacency relationships. The spatiotemporal aggregation module obtains the spatiotemporal aggregation value corresponding to the target user based on the user adjacency matrix and combined with multidimensional network performance data. The dimension offset module is used to obtain multiple dimension offsets based on the target user's multidimensional network performance data and spatiotemporal aggregation values. The dimension scoring module is used to perform interval mapping on multidimensional network performance data and multiple dimension offsets to obtain the corresponding dimension scores and construct a multidimensional scoring vector. The experience quality module is used to obtain the covariance matrix of the multidimensional rating vector within a preset statistical period, obtain the weight coefficients of the corresponding dimensions based on the covariance matrix, and generate the experience quality index of the target user. The monitoring results module is used to obtain the rate of change of the user adjacency matrix and the rate of change of the experience quality index within a continuous time window, obtain the diffusion intensity index based on the rate of change of the adjacency matrix and the rate of change of the experience quality index, adjust the collection frequency of multidimensional network performance data based on the experience quality index and the diffusion intensity index, and generate monitoring results data.

[0015] And, an internet access quality monitoring terminal for real-time perception of user experience, comprising: One or more processors; A storage device on which one or more programs are stored; When one or more programs are executed by one or more processors, the one or more processors can realize a method for monitoring the quality of Internet access by sensing the user experience in real time.

[0016] The technical effects achieved by this invention are as follows: This invention triggers network performance data collection through business operations, ensuring that monitoring results align with real-world user experiences and improving the accuracy of experience assessment. By constructing a user adjacency matrix and calculating spatiotemporal aggregation indicators, it enables comparative analysis of individual network performance against the average level of the group, identifying regional network quality anomalies. Utilizing the difference between individuals and the group to construct dimensional offsets highlights individual or localized anomalies, improving problem localization accuracy. Dynamically determining weighting coefficients based on the fluctuations in scores for each dimension within a statistical period increases the proportion of network performance dimensions with significant impact on experience in the experience quality index, thereby enhancing the sensitivity and real-time nature of experience assessment. Constructing a diffusion intensity indicator using the adjacency matrix change rate and the experience quality index change rate identifies whether network quality anomalies are propagating in the spatial structure, achieving an improvement from single-point anomaly monitoring to group trend monitoring. Dynamically adjusting the network performance data collection frequency based on diffusion intensity improves collection accuracy during anomaly propagation and reduces collection load in stable conditions, thus reducing system resource consumption while ensuring monitoring accuracy. This invention integrates individual network performance, spatial adjacency, time window fluctuation statistics, and diffusion trend judgment to achieve real-time dynamic perception and adaptive monitoring and control of user experience quality. Compared with traditional schemes based on fixed indicators or single-point detection, it has significant improvements in anomaly identification accuracy, group diffusion perception capability, and system resource utilization efficiency. Attached Figure Description

[0017] Figure 1 This is a flowchart of the method provided by the present invention; Figure 2 This is a system module diagram provided by the present invention. Detailed Implementation

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0019] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0020] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that mutually excludes other embodiments.

[0021] Furthermore, the present invention will be described in detail with reference to the schematic diagrams. When describing the embodiments of the present invention in detail, the schematic diagrams are merely examples for ease of explanation and should not limit the scope of protection of the present invention.

[0022] Please see the appendix Figure 1 As shown, a method for real-time perception of user experience and monitoring of Internet access quality is provided, including: S1. Obtain multi-dimensional network performance data during the business operations performed by the target user on the terminal device; S2. Obtain the geographical location information and timestamp information of multiple users within the time window, establish user adjacency relationships that meet the preset adjacency conditions, and construct a user adjacency matrix based on the user adjacency relationships; S3. Based on the user adjacency matrix and combined with multidimensional network performance data, obtain the spatiotemporal aggregation value corresponding to the target user; S4. Obtain multiple dimensional offsets based on the target user's multidimensional network performance data and spatiotemporal aggregation values; S5. Map the multidimensional network performance data and multiple dimension offsets to intervals to obtain the corresponding dimension scores and construct a multidimensional score vector. S6. Obtain the covariance matrix of the multidimensional rating vector within the preset statistical period, obtain the weight coefficients of the corresponding dimensions based on the covariance matrix, and generate the experience quality index of the target user. S7. Obtain the rate of change of the user adjacency matrix and the rate of change of the experience quality index within a continuous time window. Obtain the diffusion intensity index based on the rate of change of the adjacency matrix and the rate of change of the experience quality index. Adjust the collection frequency of multidimensional network performance data based on the experience quality index and the diffusion intensity index, and generate monitoring result data.

[0023] As described in steps S1 to S7 above, a business time window is determined on the terminal device through a business operation triggering mechanism. Within this time window, network performance data such as average round-trip latency, latency jitter, packet loss ratio, and throughput are collected. By using real business operations as the collection trigger condition, the collected network performance data is ensured to be consistent with the actual user experience scenario, thereby avoiding isolated network detection detached from the business scenario. Within the same time window, the geographical location information and timestamp information of multiple users are obtained. Spatial distance thresholds and time thresholds are used to determine whether there is an adjacency relationship between users, and a user adjacency matrix is ​​constructed. Based on the user adjacency matrix, the set of neighboring users that are adjacent to the target user is extracted, and the data of neighboring users at the same time is obtained. The network performance values ​​within the window are weighted by introducing spatial weights to the network performance of adjacent users, generating a spatiotemporal aggregate value for the target user. The multidimensional network performance data of the target user is then compared with the corresponding spatiotemporal aggregate value to obtain multiple dimensional offsets. These multidimensional network performance data and offsets are input into a preset dimensional mapping table. Preset scores are obtained based on corresponding cross-intervals, forming dimensional scores for each network performance dimension. A multidimensional score vector is constructed, and through interval mapping, unified quantification of network indicators with different dimensions is achieved. Within a preset statistical period, multiple time windows corresponding to the multidimensional score vectors are obtained. The covariance matrix is ​​calculated based on the fluctuation of each dimension's score within the statistical period, and the variance value of each dimension is extracted. The weighting coefficients are determined based on the proportion of each dimension's variance to the total variance, increasing the weight of dimensions with greater fluctuations in the experience quality index calculation. Then, the multi-dimensional scoring vectors within the current time window are weighted to generate the target user's experience quality index. The rate of change in adjacency relationships is calculated by comparing changes in the user's adjacency matrix within a continuous time window, and the rate of change in the experience quality index is also calculated within the continuous time window. A diffusion intensity index is generated based on the combination of the adjacency matrix change rate and the experience quality index change rate. This diffusion intensity index reflects whether user experience anomalies are spreading within the user group structure. When the diffusion intensity exceeds a preset threshold, the network performance sampling frequency is increased; otherwise, the original sampling frequency is maintained. Network performance sampling is triggered by business operations. This system ensures that monitoring results align with real-world user experiences, improving the accuracy of experience assessments. By constructing a user adjacency matrix and calculating spatiotemporal aggregation values, it enables comparative analysis of individual network performance against the average level of the group, identifying regional network quality anomalies. Utilizing the difference between individuals and the group to construct a dimensional offset highlights individual or localized anomalies, improving problem localization accuracy. The system dynamically determines weighting coefficients based on the fluctuations in scores for each dimension within the statistical period, increasing the proportion of network performance dimensions with significant impact on experience in the experience quality index, thereby enhancing the sensitivity and real-time nature of experience assessments. Finally, by constructing a diffusion intensity index using the rate of change of the adjacency matrix and the rate of change of the experience quality index, it can identify whether network quality anomalies are propagating within the spatial structure.This approach enhances monitoring capabilities from single-point anomaly detection to group trend monitoring. It dynamically adjusts network performance sampling frequency based on diffusion intensity, improving sampling accuracy during anomaly propagation and reducing sampling load in stable conditions. This ensures monitoring accuracy while reducing system resource consumption. By integrating individual network performance, spatial adjacency relationships, time window fluctuation statistics, and diffusion trend judgment, it achieves real-time dynamic perception and adaptive monitoring and control of user experience quality. Compared to traditional solutions based on fixed indicators or single-point detection, it significantly improves anomaly identification accuracy, group diffusion perception capability, and system resource utilization efficiency.

[0024] In a preferred embodiment, multi-dimensional network performance data is obtained during the service operations performed by the target user on the terminal device, including: S101. Preset operation trigger conditions in the terminal device. When the target user is detected to have triggered the preset operation trigger conditions, mark it as the start time of the business operation. S102. When the end of a business operation is detected, mark it as the end time of the business operation; S103. Obtain the time window based on the start time and end time of the business operation; S104. Obtain multi-dimensional network performance data of the target user's terminal device operation within the time window. The multi-dimensional network performance data includes the average round-trip delay, delay jitter value, packet loss ratio and throughput.

[0025] As described in steps S101 to S104 above, operation trigger conditions are preset in the terminal device. Examples include: an application initiating a network connection request, video playback starting, a page load request being sent, file upload or download starting, or a game entering a battle phase. When the trigger conditions are met, this moment is marked as the start of the business operation. By listening to application layer events or network layer session establishment behavior, the user's proactive business initiation behavior is used as the starting point for network measurement, avoiding idle measurements detached from actual business scenarios. Operation end conditions are also preset in the terminal device. Examples include: detecting application layer business completion indicators (such as page loading completion, video playback ending, file transfer completion confirmation), and normal network connection closure (such as TCP connection termination). The end of a service operation is marked when any of the following conditions are met: completion of FIN message interaction, no data interaction within a preset duration (idle timeout determination), user active exit or service switching. Through application layer state monitoring or transport layer session state monitoring, the entire service lifecycle is captured, forming a measurement cycle that perfectly corresponds to the real service. A time window is determined based on the start and end times of the service operation, driven by user behavior rather than a fixed clock. Within the time window, multi-dimensional network performance data is statistically analyzed, including average round-trip time (calculated by averaging the round-trip time between requests and responses using transport layer protocols), latency jitter (calculated by the change in arrival time difference between adjacent data packets, reflecting transmission stability), packet loss ratio (obtained by the ratio of the number of lost packets to the total number of sent packets within the statistical time window), and throughput (obtained by dividing the total amount of successfully received data within the statistical time window by the time window length). A dynamic time window, strictly aligned with the service lifecycle, is constructed using real service operations as the trigger source, and multi-dimensional network performance statistical calculations are performed within this time window to obtain network quality data with a true semantic basis for the service.

[0026] In a preferred implementation, the geographical location information and timestamp information of multiple users within a time window are obtained, and user adjacency relationships that meet preset adjacency conditions are established. A user adjacency matrix is ​​then constructed based on these user adjacency relationships, including: S201. Obtain the geographic location information and corresponding timestamp information reported by multiple user terminal devices within the time window, wherein the geographic location information includes latitude and longitude coordinate data; S202. Obtain the spatial distance and time difference between any two users; S203. Obtain a preset user adjacency threshold, wherein the preset user adjacency threshold includes a preset spatial distance threshold and a preset time threshold, and determine whether the spatial distance and time difference between any two users meet the preset user adjacency threshold. When the spatial distance is less than the preset spatial distance threshold and the time difference is less than the preset time threshold, it is determined that there is an adjacency relationship between the two users; otherwise, it is determined that there is no adjacency relationship. S204. Construct a user adjacency matrix based on the adjacency relationships between any two users. When any two users are adjacent, the corresponding row and column elements in the matrix are 1, and when any two users are not adjacent, the corresponding row and column elements in the matrix are 0.

[0027] As described in steps S201 to S204 above, within a time window, geographic location information and corresponding timestamp information reported by multiple user terminal devices are received. The geographic location information includes latitude and longitude coordinates, used to characterize the user's specific location in space. The timestamp information identifies the specific time of data collection, ensuring the alignment of different user data across time. For any two users, spatial distance is obtained based on their latitude and longitude coordinates (latitude and longitude coordinates are obtained through positioning information provided by the terminal devices (such as GPS positioning, base station positioning, or Wi-Fi positioning), and the spherical distance between the two points is calculated to obtain the spatial distance value between the two users). Spatial distance quantifies the physical proximity of two users. Simultaneously, by comparing the timestamp information of the two user data, the time difference between the two users is obtained, used to quantify the temporal proximity of user actions. The spatial distance value reflects spatial proximity, and the time difference value reflects temporal synchronization; both constitute the basic parameters for user adjacency determination. A preset user adjacency threshold is included, including a preset spatial distance threshold. The system employs a distance threshold and a preset time threshold. The thresholds can be obtained through statistical analysis of historical user distribution data, pre-set according to business scenario requirements, or dynamically adjusted based on network coverage, user density, or business type. The spatial distance between any two users is compared to the preset spatial distance threshold, and the time difference is also compared to the preset time threshold. If both conditions are met, a connection is established between the two users; otherwise, no connection is established. This ensures that the connection is not only spatially close but also temporally synchronized, thus improving the accuracy of the connection determination. After determining the connection between all user pairs, a user adjacency matrix is ​​constructed. The rows and columns of the matrix correspond to different users. If user i and user j are adjacent, the element in the i-th row and j-th column of the matrix is ​​set to 1; otherwise, the corresponding element is set to 0. This matrix is ​​used to systematically describe the spatiotemporal adjacency status between user groups. The adjacency matrix can be updated dynamically with the time window, reflecting changes in user distribution and transforming scattered adjacency relationships into structured matrix data for easy retrieval and calculation.

[0028] In a preferred implementation, based on the user adjacency matrix and combined with multidimensional network performance data, the spatiotemporal aggregation value corresponding to the target user is obtained, including: S301. Obtain the row number of the target user in the user adjacency matrix; S302. In the user adjacency matrix, find the column index of the row corresponding to the target user where the element has a value of 1, and obtain the set of neighboring users that have an adjacency relationship with the target user. S303. Extract multiple network performance dimensions based on multidimensional network performance data. These multiple network performance dimensions include round-trip time, latency jitter, packet loss, and throughput. S304. Based on each network performance dimension, extract the network performance values ​​corresponding to the adjacent user sets within the same time window; S305. Obtain the spatial distance between each neighboring user in the neighboring user set and the target user, and obtain the corresponding spatial weight based on the corresponding spatial distance; S306. Gather the network performance values ​​of adjacent users on the same network performance dimension, and combine them with the corresponding spatial weights to obtain the spatiotemporal aggregation value of the target user on the corresponding network performance dimension.

[0029] As described in steps S301 to S306 above, the row number corresponding to the target user is determined in the constructed user adjacency matrix. Each row and column in the adjacency matrix uniquely corresponds to a user. Therefore, the target user's position in the matrix can be quickly located through the mapping relationship between the user identifier and the matrix index. This mapping relationship can be generated synchronously during the matrix construction stage, for example, by establishing an index table associated with the user ID and the matrix index, avoiding repeated traversal of all users and improving processing efficiency. In the matrix row corresponding to the target user, the column number with an element value of 1 is searched. Since the elements in the matrix... A value of 1 indicates the existence of an adjacency relationship. Therefore, all users corresponding to columns with a value of 1 constitute the set of adjacent users that are adjacent to the target user. This set represents the group of users that meet the adjacency condition with the target user in both spatial and temporal dimensions within the current time window. This allows for the rapid extraction of the set of adjacent users related to the target user. Based on the collected multidimensional network performance data, multiple network performance dimensions are extracted, including round-trip latency, latency jitter, packet loss, and throughput. These dimensions can be obtained through the terminal-side network protocol statistics module or system monitoring module, avoiding evaluation based on a single indicator. To mitigate the bias caused by price discrepancies, after determining the set of neighboring users, the corresponding values ​​of these neighboring users in each network performance dimension within the same time window are extracted. The time window is kept consistent with the adjacency matrix construction stage to ensure data synchronization in the time dimension and that all aggregated indicator data originates from the same time interval, thereby avoiding distortion caused by cross-time window statistics. The spatial distance between each neighboring user and the target user in the set of neighboring users is obtained (by obtaining latitude and longitude coordinates and calculating the spherical distance between two points based on latitude and longitude data). After obtaining the spatial distances between all neighboring users and the target user, corresponding spatial weights are generated based on the distance values ​​(which can be obtained by calculating the ratio of each corresponding spatial distance to the total spatial distance). The values ​​of each neighboring user in the same indicator dimension are combined with their spatial weights, and the weighted results of all neighboring users are summarized and normalized based on the number of neighboring users. Finally, the spatiotemporal aggregated value of the target user in that network performance dimension is obtained, which not only reflects the target user's own state but also comprehensively reflects the network performance of its surrounding neighboring users, expanding individual network indicators into group-aware indicators.

[0030] In a preferred implementation, multiple dimensional offsets are obtained based on the target user's multidimensional network performance data and spatiotemporal aggregation values, including: S401. Obtain the network performance value and corresponding spatiotemporal aggregation value for each network performance dimension of the target user; S402. Based on the network performance value and the corresponding spatiotemporal aggregation value corresponding to each network performance dimension of the target user, obtain the dimension offset under the same network performance dimension.

[0031] As described in steps S401 to S402 above, the raw network performance values ​​of the target user across various network performance dimensions are obtained (derived from multi-dimensional network performance data collected from the target user within a time window, including round-trip latency, latency jitter, packet loss ratio, and throughput), along with the corresponding spatiotemporal aggregation values. For each network performance dimension, the network performance value corresponding to the target user is compared and analyzed with the spatiotemporal aggregation value corresponding to that dimension to obtain the dimension offset within the same dimension (the dimension offset is obtained by calculating the difference between the network performance value corresponding to the target user and the corresponding spatiotemporal aggregation value). The result (i.e., dimensional offset) is used to characterize the degree of deviation of the target user's current network performance status from the average level of its neighborhood. When the difference between the target user's indicator value and the spatiotemporal aggregation value is small, it indicates that its network status is basically consistent with the overall status of surrounding users. When the difference is large, it indicates that the target user has a significant deviation in this indicator dimension, which may indicate individual anomalies or local network fluctuations. If the overall neighborhood status is poor but the individual offset is small, it indicates a regional problem. If the overall neighborhood status is normal but the individual offset is large, it indicates an individual anomaly. This enables refined identification of individual network anomalies and improves the accuracy of experience quality assessment.

[0032] In a preferred implementation, the multidimensional network performance data and multiple dimensional offsets are interval-mapped to obtain corresponding dimensional scores, and a multidimensional score vector is constructed, including: S501. Obtain the dimension mapping table, wherein the dimension mapping table includes multiple network performance value ranges, multiple dimension offset ranges, and a preset score value corresponding to the intersection of each network performance value range and each dimension offset range. S502. Based on the network performance value and corresponding dimension offset of each network performance dimension of the target user, and in conjunction with the dimension mapping table, obtain the corresponding preset score value and mark it as the dimension score of the corresponding network performance dimension. S503. Obtain the preset dimension order and arrange the dimension scores corresponding to each network performance dimension in sequence to generate a multi-dimensional score vector for the target user. The multi-dimensional score vector includes round-trip time dimension score, latency jitter dimension score, packet loss dimension score, and throughput dimension score.

[0033] As described in steps S501 to S503 above, a dimension mapping table is obtained. This table is a two-dimensional mapping structure, with its horizontal axis consisting of multiple network performance value ranges and its vertical axis consisting of multiple dimension offset ranges. The cells corresponding to the intersection of the horizontal and vertical axes contain preset dimension score values. The dimension mapping table can be obtained through historical data statistical analysis (by statistically analyzing a large amount of historical network performance data and user experience feedback data, different network performance ranges and user experience levels corresponding to different offset degrees are classified and graded to generate score ranges), expert experience modeling (based on network optimization experience and business service standards, different network performance levels and offset degrees are classified to form a rule mapping table), or business level standard mapping (which can combine the operator's or system's preset service level standards to divide network performance into excellent, good, medium, and poor ranges, and combine the offset degree to form a two-dimensional score). The system, after generating the mapping table, can store it permanently. During real-time calculation, the corresponding score can be directly retrieved from the table. For each network performance dimension of the target user, the corresponding network performance value and the corresponding dimension offset are obtained. Based on the interval position of the two in the dimension mapping table, the intersection cell of the horizontal and vertical axes is determined. The preset score value is read from the cell and marked as the dimension score of the network performance dimension. The two-dimensional continuous value is mapped to the discrete score value to achieve standardized scoring processing. The preset dimension order is obtained. The preset dimension order can be set according to the importance of business or the degree of impact on network performance. According to the preset order, the dimension scores corresponding to each dimension are arranged in sequence to form a multi-dimensional score vector of the target user. The multi-dimensional score vector is a structured score set that can completely represent the comprehensive status of the target user in multiple network performance dimensions.

[0034] In a preferred implementation, the covariance matrix of the multidimensional rating vectors within a preset statistical period is obtained, and the weight coefficients of the corresponding dimensions are obtained based on the covariance matrix to generate the experience quality index of the target user, including: S601. Obtain a preset statistical period, wherein the preset statistical period includes multiple consecutive time windows, and there are no fewer than two consecutive time windows; S602. Within a preset statistical period, obtain multidimensional scoring vectors corresponding to multiple time windows in chronological order. S603. Based on the score of each dimension in the multidimensional scoring vector, obtain multiple score values ​​corresponding to the corresponding dimension score within a preset statistical period to form a score sequence for the corresponding dimension. S604. Obtain the corresponding covariance matrix based on the score sequence of each dimension, and extract the variance value corresponding to the score of each dimension from the covariance matrix; S605. Obtain the weight coefficient corresponding to each dimension score based on the proportion of the variance value corresponding to each dimension score to the total variance of all dimension scores. S606. Obtain the multi-dimensional rating vector corresponding to the current time window, and obtain the experience quality index of the target user within the current time window based on the rating of each dimension and the corresponding weight coefficient.

[0035] As described in steps S601 to S606 above, a preset statistical period is obtained, including multiple consecutive time windows, with no fewer than two consecutive time windows. The statistical period is used to reflect the overall fluctuation of the network status over a period of time. By setting a statistical period, the random errors caused by relying solely on data from a single time window can be avoided. Within the preset statistical period, multi-dimensional scoring vectors corresponding to multiple time windows are obtained in chronological order. Each time window corresponds to a multi-dimensional scoring vector, which contains multiple network performance dimension scores, such as round-trip latency score, latency jitter score, packet loss score, and throughput score, arranged in chronological order. This can create a multidimensional scoring dataset that varies over time. For each dimension in the multidimensional scoring vector, a score is assigned. Within the statistical period, the score values ​​for that dimension are extracted across all time windows, forming a scoring sequence for that dimension. For example, round-trip delay forms a time series. Each scoring sequence reflects the fluctuation of that dimension within the statistical period. The scoring sequences for each dimension are aligned chronologically to form a two-dimensional data table. Each column corresponds to a network performance dimension, and each row corresponds to a time window. For each dimension's scoring sequence, its mean within the statistical period is calculated. For any two dimension scoring sequences, the mean is calculated at the same time window position. The deviations of each dimension from its respective mean are calculated, and their joint trend is statistically analyzed. By statistically analyzing the joint volatility of each dimension across all time windows, the covariance values ​​between each dimension are obtained. The covariance values ​​between all pairs of dimensions are arranged in dimensional order to form a covariance matrix. The values ​​on the diagonal of the covariance matrix represent the variance of each dimension's score, used to measure the intensity of its volatility within the statistical period. The weighting coefficient for each dimension is determined based on the proportion of its variance to the total variance of all dimensions. Specifically, dimensions with larger variances indicate more significant volatility and a greater impact on user experience changes, while those with smaller variances... Dimensions are relatively stable and have a relatively small impact on the overall experience. Therefore, by calculating the variance ratio of each dimension, the weight coefficient of each dimension can be obtained. The weights are then normalized to keep the sum of all weights consistent, achieving adaptive weight adjustment. The multi-dimensional score vector corresponding to the current time window is obtained. The score of each dimension is combined with the corresponding weight coefficient, and the weighted results of each dimension are summarized to obtain the experience quality index of the target user in the current time window. This experience quality index is a comprehensive evaluation value that reflects the overall network experience level of the target user in the current time window. Multi-dimensional scores are compressed into a single comprehensive indicator, which is convenient for management and display.

[0036] In a preferred implementation, the rate of change of the user adjacency matrix and the rate of change of the experience quality index within a continuous time window are obtained. A diffusion intensity index is obtained based on the rate of change of the adjacency matrix and the rate of change of the experience quality index. The collection frequency of multidimensional network performance data is adjusted based on the experience quality index and the diffusion intensity index, and monitoring result data is generated, including: S701. Obtain the user adjacency matrix corresponding to two consecutive time windows; S702. Compare the two user adjacency matrices element by element to obtain the number of changes in the values ​​of the elements in the matrices, and obtain the rate of change of the user adjacency matrix based on the number of changes in the values ​​of the elements in the matrices. S703: Obtain the experience quality index corresponding to two consecutive time windows; S704. Obtain the difference in experience quality based on the experience quality index corresponding to two consecutive time windows, and generate the rate of change of the experience quality index. S705. Obtain the diffusion intensity index based on the rate of change of the user adjacency matrix and the rate of change of the experience quality index. S706. Obtain a preset diffusion intensity threshold and determine whether the diffusion intensity index exceeds the preset diffusion intensity threshold. When the diffusion intensity index exceeds the preset diffusion intensity threshold, the collection frequency of multidimensional network performance data is increased. When the diffusion intensity index does not exceed the preset diffusion intensity threshold, the collection frequency of multidimensional network performance data is maintained. S707. Obtain multi-dimensional network performance data based on the adjusted acquisition frequency, and generate monitoring result data by combining the experience quality index and diffusion intensity index of the corresponding time window.

[0037] As described in steps S701 to S707 above, user adjacency matrices corresponding to two consecutive time windows are obtained. These two adjacency matrices reflect the spatiotemporal adjacency structure between users within different time windows. Elements at corresponding positions in the two matrices are compared, and the number of positions where element values ​​change is counted. The proportion of the number of elements that have changed to the total number of elements in the matrix is ​​calculated to obtain the adjacency matrix change rate. The adjacency matrix change rate is used to quantify the degree of change in the user group structure over time. The experience quality index corresponding to two consecutive time windows is obtained, and the difference between the experience quality index of the later time window and the previous time window is calculated. The difference is proportionally calculated with the experience quality index of the previous time window to obtain the experience quality index change rate, which reflects the magnitude of change in the target user experience status. Based on the adjacency matrix change rate and the experience quality index change rate, a diffusion intensity index is generated (this can be obtained by multiplying the user adjacency matrix change rate and the experience quality index change rate, or by weighting and summing the user adjacency matrix change rate and the experience quality index change rate according to preset weighting coefficients). The combination of these two methods can reflect whether changes in individual experience are accompanied by changes in group structure, thereby identifying whether abnormal experience has a spreading trend. A preset diffusion intensity threshold is obtained (determined based on historical diffusion data statistical analysis, preset based on business risk level, or dynamically adjusted according to system operation). The diffusion intensity index is compared with this threshold. When the diffusion intensity index exceeds the threshold, it indicates that the current experience change may have a spreading trend, and the collection frequency of multidimensional network performance data is increased. When the threshold is not exceeded, the original collection frequency is maintained, realizing dynamic allocation of monitoring resources. This improves data collection accuracy when the risk increases and saves system resources when the risk is low, improving overall monitoring efficiency. New multidimensional network performance data is obtained based on the adjusted collection frequency. Simultaneously, the experience quality index of the current time window, the diffusion intensity index of the current time window, and the collection frequency adjustment results are integrated and processed to generate monitoring result data, including the current experience level, change trend indicator, whether there is a diffusion risk warning, and suggested handling strategies. This enables real-time monitoring of changes in user experience quality, dynamic identification of group diffusion trends, and improved network quality assurance capabilities.

[0038] Please see the appendix Figure 2 As shown, the present invention also provides an Internet access quality monitoring system for real-time perception of user experience, used in the aforementioned Internet access quality monitoring method for real-time perception of user experience, comprising: The network performance module is used to acquire multi-dimensional network performance data during the business operations performed by the target user on the terminal device. The adjacency matrix module is used to obtain the geographical location information and timestamp information of multiple users within a time window, establish user adjacency relationships that meet preset adjacency conditions, and construct a user adjacency matrix based on the user adjacency relationships. The spatiotemporal aggregation module obtains the spatiotemporal aggregation value corresponding to the target user based on the user adjacency matrix and combined with multidimensional network performance data. The dimension offset module is used to obtain multiple dimension offsets based on the target user's multidimensional network performance data and spatiotemporal aggregation values. The dimension scoring module is used to perform interval mapping on multidimensional network performance data and multiple dimension offsets to obtain the corresponding dimension scores and construct a multidimensional scoring vector. The experience quality module is used to obtain the covariance matrix of the multidimensional rating vector within a preset statistical period, obtain the weight coefficients of the corresponding dimensions based on the covariance matrix, and generate the experience quality index of the target user. The monitoring results module is used to obtain the rate of change of the user adjacency matrix and the rate of change of the experience quality index within a continuous time window, obtain the diffusion intensity index based on the rate of change of the adjacency matrix and the rate of change of the experience quality index, adjust the collection frequency of multidimensional network performance data based on the experience quality index and the diffusion intensity index, and generate monitoring results data.

[0039] The above-mentioned network performance module collects multi-dimensional network performance data during actual business operations of the target user on the terminal device. This data reflects the network status under the user's real usage scenario and is the basic input for experience perception. While acquiring individual network performance data, the adjacency matrix module collects the geographical location information and timestamp information of multiple users within the corresponding time window. The system establishes adjacency relationships between users according to preset adjacency conditions (such as spatial distance threshold and time overlap threshold) and constructs a user adjacency matrix accordingly. This adjacency matrix depicts the spatiotemporal relationship structure between users within the same time period. Subsequently, the spatiotemporal aggregation module extracts the set of users that have an adjacency relationship with the target user based on the constructed user adjacency matrix and performs analysis on the set. The system statistically processes the corresponding multidimensional network performance data to obtain the spatiotemporal aggregate value for the target user. Next, the dimension offset module calculates the difference between the target user's multidimensional network performance value and the corresponding spatiotemporal aggregate value, resulting in multiple dimension offsets. These offsets reflect the performance difference of the target user relative to its spatiotemporally adjacent group, achieving relative positioning between the individual and the group. After obtaining the original indicator data and dimension offsets, the dimension scoring module maps the two types of data to intervals according to a preset dimension mapping table, obtaining the dimension scores corresponding to each network performance dimension. It then constructs a multidimensional scoring vector according to a preset order. Through this mapping process, the original indicators are standardized into a unified scoring system, laying the foundation for comprehensive calculation. Subsequently... The experience quality module acquires multi-dimensional rating vectors corresponding to multiple consecutive time windows within a preset statistical period and constructs a covariance matrix. It calculates the variance of each dimension's rating within the statistical period and determines weighting coefficients based on the proportion of each dimension's variance to the total variance, forming an adaptive weighting mechanism. Finally, it weights and sums the multi-dimensional rating vector of the current time window with its corresponding weighting coefficients to obtain the target user's experience quality index. This index comprehensively reflects the user's current network experience level. After obtaining the experience quality index for consecutive time windows, the monitoring results module further performs dynamic change analysis. The system first calculates the rate of change of the user adjacency matrix between two consecutive time windows to measure the degree of change in the group structure, and simultaneously calculates the experience quality... The exponential rate of change measures the magnitude of changes in individual experience. The two rates of change are then combined to obtain the diffusion intensity index. This index is used to determine whether changes in individual experience are accompanied by changes in group structure, thereby identifying whether there is a group diffusion trend. When the diffusion intensity index exceeds a preset threshold, the system determines that there may be a risk of abnormal experience diffusion, thus increasing the collection frequency of multidimensional network performance data to enhance monitoring accuracy. When the diffusion intensity index does not exceed the threshold, the original collection frequency is maintained to reduce system resource consumption. Finally, the system combines the experience quality index, diffusion intensity index, and collection frequency status of the current time window to generate monitoring result data, achieving real-time perception of user experience quality and group diffusion trends.

[0040] And, an internet access quality monitoring terminal for real-time perception of user experience, comprising: One or more processors; A storage device on which one or more programs are stored; When one or more programs are executed by one or more processors, the one or more processors can realize a method for monitoring the quality of Internet access by sensing the user experience in real time.

[0041] The above description is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained in this invention are implemented according to conventional methods in the art unless otherwise specified or limited.

Claims

1. A method for real-time sensing of user experience and monitoring of Internet access quality, characterized in that, include: Acquire multi-dimensional network performance data during business operations performed by the target user on the terminal device; Obtain the geographic location information and timestamp information of multiple users within a time window, establish user adjacency relationships that meet preset adjacency conditions, and construct a user adjacency matrix based on the user adjacency relationships; Based on the user adjacency matrix and combined with multidimensional network performance data, the spatiotemporal aggregation value corresponding to the target user is obtained; Multiple dimensional offsets are obtained based on the target user's multidimensional network performance data and spatiotemporal aggregation values; Multidimensional network performance data and multiple dimensional offsets are interval-mapped to obtain corresponding dimensional scores, and a multidimensional score vector is constructed. Obtain the covariance matrix of the multidimensional rating vector within a preset statistical period, obtain the weight coefficients of the corresponding dimensions based on the covariance matrix, and generate the experience quality index of the target user. The system obtains the rate of change of the user adjacency matrix and the rate of change of the experience quality index within a continuous time window. Based on the rate of change of the adjacency matrix and the rate of change of the experience quality index, it obtains the diffusion intensity index. Based on the experience quality index and the diffusion intensity index, it adjusts the collection frequency of multidimensional network performance data and generates monitoring result data.

2. The Internet access quality monitoring method for real-time perception of user experience according to claim 1, characterized in that, Obtain multi-dimensional network performance data during the business operations performed by the target user on the terminal device, including: Pre-set operation trigger conditions in the terminal device. When the target user triggers the preset operation trigger conditions, it is marked as the start time of the business operation. When the end of a business operation is detected, it is marked as the end time of the business operation; Obtain the time window based on the start and end times of the business operation; Acquire multidimensional network performance data of the target user's terminal device operation within the time window. The multidimensional network performance data includes average round-trip latency, latency jitter, packet loss ratio, and throughput.

3. The Internet access quality monitoring method for real-time perception of user experience according to claim 1, characterized in that, Obtain the geographic location and timestamp information of multiple users within a time window, establish user adjacency relationships that meet preset adjacency conditions, and construct a user adjacency matrix based on the user adjacency relationships, including: Within a time window, acquire the geographic location information and corresponding timestamp information reported by multiple user terminal devices, where the geographic location information includes latitude and longitude coordinate data; Get the spatial distance and time difference between any two users; Obtain a preset user adjacency threshold, which includes a preset spatial distance threshold and a preset time threshold. Determine whether the spatial distance and time difference between any two users meet the preset user adjacency threshold. If the spatial distance is less than the preset spatial distance threshold and the time difference is less than the preset time threshold, it is determined that there is an adjacency relationship between the two users; otherwise, it is determined that there is no adjacency relationship. Construct a user adjacency matrix based on the adjacency relationships between any two users. When any two users are adjacent, the corresponding row and column elements in the matrix are 1, and when any two users are not adjacent, the corresponding row and column elements in the matrix are 0.

4. The Internet access quality monitoring method for real-time perception of user experience according to claim 1, characterized in that, Based on the user adjacency matrix and combined with multidimensional network performance data, the spatiotemporal aggregation value corresponding to the target user is obtained, including: Obtain the row number of the target user in the user adjacency matrix; In the user adjacency matrix, find the column index of the row corresponding to the target user where the element has a value of 1, and obtain the set of neighboring users that have an adjacency relationship with the target user; Multiple network performance dimensions are extracted based on multidimensional network performance data, including round-trip time, latency jitter, packet loss, and throughput. Based on each network performance dimension, extract the network performance values ​​corresponding to the adjacent user sets within the same time window; Obtain the spatial distance between each neighboring user in the neighboring user set and the target user, and obtain the corresponding spatial weight based on the corresponding spatial distance; By grouping adjacent users together on the same network performance dimension and combining them with corresponding spatial weights, the spatiotemporal aggregation value of the target user on the corresponding network performance dimension is obtained.

5. The Internet access quality monitoring method for real-time perception of user experience according to claim 1, characterized in that, Based on the target user's multidimensional network performance data and spatiotemporal aggregation values, multiple dimensional offsets are obtained, including: Obtain the network performance values ​​and corresponding spatiotemporal aggregation values ​​for each network performance dimension of the target user; Based on the network performance value and spatiotemporal aggregation value corresponding to each network performance dimension of the target user, obtain the dimension offset under the same network performance dimension.

6. The Internet access quality monitoring method for real-time perception of user experience according to claim 5, characterized in that, Multidimensional network performance data and multiple dimensional offsets are interval-mapped to obtain corresponding dimensional scores, and a multidimensional score vector is constructed, including: Obtain the dimension mapping table, which includes multiple network performance value ranges, multiple dimension offset ranges, and a preset score value corresponding to the intersection of each network performance value range and each dimension offset range. Based on the network performance value and corresponding dimension offset for each network performance dimension of the target user, and combined with the dimension mapping table, the corresponding preset score value is obtained and marked as the dimension score of the corresponding network performance dimension. Obtain the preset dimension order and arrange the dimension scores corresponding to each network performance dimension in sequence to generate a multi-dimensional score vector for the target user. The multi-dimensional score vector includes round-trip time dimension score, latency jitter dimension score, packet loss dimension score, and throughput dimension score.

7. The Internet access quality monitoring method for real-time perception of user experience according to claim 1, characterized in that, Obtain the covariance matrix of the multidimensional rating vectors within a preset statistical period, extract the weight coefficients of the corresponding dimensions based on the covariance matrix, and generate the experience quality index for the target user, including: Obtain a preset statistical period, wherein the preset statistical period includes multiple consecutive time windows, and there are no fewer than two consecutive time windows; Within a preset statistical period, multidimensional scoring vectors corresponding to multiple time windows are obtained in chronological order. Based on the score of each dimension in the multidimensional scoring vector, obtain multiple score values ​​corresponding to the corresponding dimension within a preset statistical period to form a score sequence for the corresponding dimension. Obtain the corresponding covariance matrix based on the score sequence for each dimension, and extract the variance value corresponding to the score for each dimension from the covariance matrix; Based on the proportion of the variance value corresponding to each dimension score to the total variance of all dimension scores, obtain the weight coefficient corresponding to each dimension score. Obtain the multi-dimensional rating vector corresponding to the current time window, and obtain the experience quality index of the target user within the current time window based on the rating of each dimension and the corresponding weight coefficient.

8. The method for real-time sensing of user experience and monitoring of Internet access quality according to claim 1, characterized in that, Obtain the rate of change of the user adjacency matrix and the rate of change of the experience quality index within a continuous time window. Based on these rates, obtain the diffusion intensity index. Adjust the collection frequency of multidimensional network performance data according to the experience quality index and the diffusion intensity index, and generate monitoring result data, including: Obtain the user adjacency matrix corresponding to two consecutive time windows; The two user adjacency matrices are compared element by element to obtain the number of changes in the values ​​of the elements in the matrix, and the rate of change of the user adjacency matrix is ​​obtained based on the number of changes in the values ​​of the elements in the matrix. Obtain the experience quality index corresponding to two consecutive time windows; The difference in experience quality is obtained based on the experience quality index corresponding to two consecutive time windows, and the rate of change of the experience quality index is generated. The diffusion intensity index is obtained based on the rate of change of the user adjacency matrix and the rate of change of the experience quality index. Obtain a preset diffusion intensity threshold and determine whether the diffusion intensity index exceeds the preset diffusion intensity threshold; When the diffusion intensity index exceeds the preset diffusion intensity threshold, the collection frequency of multidimensional network performance data is increased. When the diffusion intensity index does not exceed the preset diffusion intensity threshold, the collection frequency of multidimensional network performance data is maintained. Multidimensional network performance data is obtained based on the adjusted acquisition frequency, and monitoring results are generated by combining the experience quality index and diffusion intensity index of the corresponding time window.

9. A real-time user experience sensing internet access quality monitoring system, applied to the real-time user experience sensing internet access quality monitoring method according to any one of claims 1 to 8, characterized in that, include: The network performance module is used to acquire multi-dimensional network performance data during the business operations performed by the target user on the terminal device. The adjacency matrix module is used to obtain the geographical location information and timestamp information of multiple users within a time window, establish user adjacency relationships that meet preset adjacency conditions, and construct a user adjacency matrix based on the user adjacency relationships. The spatiotemporal aggregation module obtains the spatiotemporal aggregation value corresponding to the target user based on the user adjacency matrix and combined with multidimensional network performance data. The dimension offset module is used to obtain multiple dimension offsets based on the target user's multidimensional network performance data and spatiotemporal aggregation values. The dimension scoring module is used to perform interval mapping on multidimensional network performance data and multiple dimension offsets to obtain the corresponding dimension scores and construct a multidimensional scoring vector. The experience quality module is used to obtain the covariance matrix of the multidimensional rating vector within a preset statistical period, obtain the weight coefficients of the corresponding dimensions based on the covariance matrix, and generate the experience quality index of the target user. The monitoring results module is used to obtain the rate of change of the user adjacency matrix and the rate of change of the experience quality index within a continuous time window, obtain the diffusion intensity index based on the rate of change of the adjacency matrix and the rate of change of the experience quality index, adjust the collection frequency of multidimensional network performance data based on the experience quality index and the diffusion intensity index, and generate monitoring results data.

10. A real-time user experience sensing internet access quality monitoring terminal, characterized in that, include: One or more processors; A storage device on which one or more programs are stored; When one or more programs are executed by one or more processors, the one or more processors implement the Internet access quality monitoring method for real-time perception of user experience as described in any one of claims 1 to 8.

Images