Internet access perception experience optimization method based on user perception

By employing machine learning to analyze perception indicators and construct a user perception portrait model, the method effectively addresses the challenge of user experience evaluation in 5G networks, enabling targeted network optimization and improved customer satisfaction.

CN119095069AActive Publication Date: 2024-12-06XIAN KAILINSTE NETWORK TECHNOLOGY CO LTD

Patent Information

Application Number
CN202311746200.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-12-19
Publication Date
2024-12-06
Estimated Expiration
2043-12-19

AI Technical Summary

Technical Problem

Traditional network quality monitoring methods fail to accurately reflect user experience, leading to ineffective detection and resolution of perception problems in 5G mobile communication networks.

Method used

A method utilizing machine learning algorithms to analyze various perception indicators, construct a user perception portrait model, and optimize network performance based on user feedback, identifying low-scoring users and their permanent cells for targeted network optimization.

Benefits of technology

Enables accurate evaluation of user perception and service levels, allowing operators to identify and address user dissatisfaction, thereby improving network quality and customer experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119095069A_ABST
    Figure CN119095069A_ABST
Patent Text Reader

Abstract

The invention provides an internet access perception experience optimization method based on user perception, which belongs to the technical field of network communication and comprises the following steps: acquiring various perceptibility indexes influencing the perceptibility of a mobile internet access user; obtaining a perceptual index value when the perceptual degree is not satisfactory as a perceptual inflection point and a penalty value; determining a score corresponding to the perceptibility index value according to the perception inflection point and the penalty value; obtaining the index weight of each type of perceptibility index influencing the user perceptibility; calculating an overall perception score of the user in the Internet surfing service by using the score and the index weight corresponding to the perceptibility index value; determining a service proportion of each internet service of the user; wherein the service proportion is the ratio of the single service duration to the total network use duration; and multiplying the service proportions of all services by the overall perception score in the internet service and then adding to obtain a user perception portrait model. The method can find users unsatisfied with the mobile network data service and improve the perception degree and satisfaction degree of the mobile network users.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of network communications, and in particular relates to a method for optimizing Internet access perception experience based on user perception. Background Art

[0002] After years of large-scale construction, China's 5G mobile communication network has become very large in overall scale, and 5G traffic has maintained rapid growth. However, at the same time, the challenges faced by customer perception and customer awareness are becoming increasingly greater. This year, operators have also clearly proposed to reasonably grasp the balanced and coordinated development of maintaining competitive advantages, ensuring customer perception and guaranteeing investment returns.

[0003] With the arrival of 5G services, bottlenecks in traditional KPI performance optimization and analysis methods and business support capabilities have emerged, and users are unable to evaluate the service level, support capabilities, and usage perception of operators from an experience perspective. Currently, operators' network quality monitoring is shifting from network indicator-centered to user experience-centered; operators' service support is shifting from passive response to complaints to active care for users.

[0004] At the user perception level, traditional indicators cannot reflect user experience, and network monitoring cannot effectively detect perception problems, so perception problems cannot be accurately defined and located. Summary of the invention

[0005] In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a method for optimizing Internet access perception experience based on user perception.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A method for optimizing Internet access perception experience based on user perception, comprising:

[0008] Obtain various perception indicators that affect mobile Internet users' perception;

[0009] Collect historical data of various perception indicators, input the historical data into the clustering algorithm, output the perception indicator value when the perception is unsatisfactory as the perception inflection point and penalty value; determine the score corresponding to the perception indicator value according to the perception inflection point and penalty value;

[0010] Input historical data into the self-organizing map algorithm, output the weights of various perception indicators that affect user perception; use the scores and weights corresponding to the perception indicator values ​​to calculate the overall perception score of users in Internet access services;

[0011] Determine the service proportion of each Internet service of the user; wherein the service proportion is the ratio of the duration of a single service to the total network usage duration;

[0012] The business proportion of all businesses is multiplied by the overall perception score of online services and then added together to obtain a user perception portrait model;

[0013] Use the user perception portrait model to score user perception and find low-scoring users; and optimize the Internet perception experience for low-scoring users.

[0014] Furthermore, the network optimization for the cell where the low-scoring user is located includes:

[0015] Based on the calculated overall perception score of users in Internet access services, identify users with scores of 60 or below 50 as low-score users;

[0016] Obtain the traffic and duration data of the cell where the low-score user uses the network; find the cell with traffic and duration higher than the threshold as the low-score user's permanent cell;

[0017] The network perception of the resident communities is scored, and the network is optimized for the resident communities with low scores.

[0018] Further, the method of finding out the cell with the highest traffic and duration as the permanent cell of the low-score user includes:

[0019] The traffic,duration data of the cells are input into the k++ clustering algorithm,,and the cells with traffic duration higher than the threshold are output,by the k++ clustering algorithm as the permanent cells of low-score users.

[0020] Furthermore, the network perception of the resident cell is scored, and the calculation formula is:

[0021]

[0022] Wherein, n is the average perception score of users in the resident cell.

[0023] Furthermore, it also includes:

[0024] Obtain network B-domain data and O-domain data of low-scoring users;

[0025] According to the network B domain data and O domain data, the user's area is associated with the location in the map, and a user trajectory evaluation model is constructed to observe the permanent campus of low-scoring users from the map.

[0026] Furthermore, the user perception portrait model is:

[0027]

[0028] Furthermore, it also includes:

[0029] According to the business proportion and the overall perception score of Internet services, the user perception portrait model is built using the tensorflow framework.

[0030] Furthermore, the historical data of various perception indicators are collected and input into a clustering algorithm, and the clustering algorithm is a Kmeans algorithm.

[0031] Furthermore, the perception inflection point is a perception value at which the perception starts to rise or fall, and the penalty value is a perception value when the perception is between 0 and 100 points.

[0032] Furthermore, the perception index includes:

[0033] Page response success rate, page response time, page display success rate, instant messaging success rate, instant messaging server side rtt, instant messaging terminal side rtt, instant messaging TCP link establishment delay, video service pause frequency per minute, video pause delay ratio, video initial playback success rate, video initial cache delay, Game service response success rate, Game service response average delay, game freeze rate.

[0034] The present invention provides a method for optimizing Internet access experience based on user perception, which has the following beneficial effects:

[0035] The present invention first obtains various perception indicators that affect the perception of mobile Internet users, and then determines the user's perception score in a single Internet service by analyzing the values ​​of each indicator and the degree of influence of each indicator on the perception. Secondly, according to the market share of different Internet services, the user's total perception score model, that is, the user perception portrait model, is obtained; this model takes into account the scenarios of each service when the user uses the network, and can fully evaluate the user's Internet perception and output the perception score, and evaluate the operator's service level, support capabilities, and usage perception from the perspective of user experience; so that network service providers can accurately find users who are dissatisfied with the service, and then optimize the user's Internet perception experience in a targeted manner, solving the problem in the prior art that the user's perception analysis of the operator's mobile network products and the user's perception experience improvement experience cannot be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiment of the present invention and its design scheme, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0037] Figure 1 A schematic diagram of calculating the user perception score of a data service of the present invention;

[0038] Figure 2 The figure is a schematic diagram of the data processing and analysis flow of dissatisfied users resident in a cell according to the present invention. DETAILED DESCRIPTION

[0039] In order to enable those skilled in the art to better understand the technical solution of the present invention and implement it, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention.

[0040] Example:

[0041] The present invention provides a method for optimizing Internet access perception experience based on user perception, specifically as follows Figure 1 As shown, it includes: obtaining various perception indicators that affect the perception of mobile Internet users; using machine learning to determine the indicator weights of various perception indicators that affect user perception, the inflection point and penalty value of the perception indicator when the perception is unsatisfactory, and determining the score corresponding to the perception indicator value according to the perception inflection point and the penalty value; using the score corresponding to the perception indicator value and the indicator weight to calculate the user's overall perception score in Internet services; determining the business proportion of each Internet service of the user, the business proportion is the ratio of the duration of a single business to the total network usage time; constructing a user perception portrait model according to the business proportion and the overall perception score in the Internet service, using the user perception portrait model to score user perception, and finding low-scoring users; finding the permanent residential area of ​​low-scoring users, scoring the user's permanent residential area, and finding the low-scoring permanent residential area for network optimization.

[0042] The following are the specific implementation details of the present invention:

[0043] 1.1 Overall idea

[0044] A method and device for determining root causes of user perception based on data service dissatisfaction, comprising:

[0045] According to the network B-domain data (economic data, user number) and O-domain data (user KQI, signaling data, cell engineering parameters), based on the user dimension and cell dimension, a user trajectory evaluation model is established, and the correlation analysis between dissatisfied users and permanent cells is carried out to find the permanent cells that affect the network quality.

[0046] Analyze the perception of dissatisfied users. The specific process is as follows: ① Through correlation algorithm analysis, select and analyze user KQI perception indicators to complete the establishment of the indicator model set; ② Combine the service duration and traffic proportion of different services used by users to establish the service weight. Combine the degree of influence of different service stages on resident users to form a perception factor, realize customer perception quantitative scoring, and form a user perception portrait model based on perception; (user perception quantitative scoring) ③ Output the overall perception score of daily users and the perception score of 4 types of data services, traffic and duration and other indicators; ④ Combine machine learning through multiple rounds of data accumulation to output the top five resident communities every day, the perception score of the resident community where the daily user is located, and the perception score of 4 types of services, solve the problem of low-scoring communities on the network side, improve network quality, and improve customer perception.

[0047] 1.2 Introduction to the analysis process and analysis algorithm of the resident community

[0048] According to the network B-domain data (classified data, user number attribution dictionary table data) and O-domain data (cell engineering parameter data, cell performance data, user signaling data), based on user dimension and network dimension, a user trajectory evaluation model is established, and the correlation analysis between dissatisfied users and permanent cells is carried out to evaluate the user's permanent cell situation.

[0049] Data processing and analysis process of dissatisfied users in residential areas:

[0050] First, various types of data are obtained. Based on the two dimensions of users and networks, more than 30 variables are selected to establish the model indicator set. The threshold model and time of each variable can be adaptively adjusted according to the local network conditions.

[0051] Secondly, data cleaning and data association are carried out, and input data is mined through multiple rounds of machine learning. A user trajectory evaluation model and an analysis algorithm for permanent communities are established to conduct association analysis between users and permanent communities.

[0052] Thirdly, based on the combination of long-term historical data and analysis algorithms, users and their permanent residential areas are accurately identified, and labels are identified for users' permanent residential areas (workplaces, residences).

[0053] Introduction to Resident Cell Analysis Algorithm:

[0054] ① Analyze and select indicators related to users and resident communities through correlation algorithms to complete the establishment of the model indicator set;

[0055] ② Combine data cleaning and data association, rely on multiple rounds of machine learning to mine the input data, and establish a user trajectory evaluation model and a resident community analysis algorithm.

[0056] ③ According to the analysis algorithm, conduct correlation analysis between users and their permanent communities.

[0057] ④ Combine long-term historical data and analysis algorithms to accurately identify users and residential communities.

[0058] The specific algorithm model and algorithm steps are as follows:

[0059] The method used by k-means to converge local extreme values ​​is the k++ clustering algorithm. k-means++ initializes cluster centers by moving cluster centers away from each other. This time, the k++ clustering algorithm is used to train the model of the resident community.

[0060] Specific algorithm steps:

[0061] The method used by k-means to converge local extreme values ​​is the k++ clustering algorithm. k-means++ initializes cluster centers by moving cluster centers away from each other. This time, the k++ clustering algorithm is used to train the model of the resident community.

[0062] Specific algorithm steps:

[0063] 1) Randomly select a sample data as the first cluster center C1;

[0064] 2) Calculate each sample x i To cluster center C j The minimum distance;

[0065]

[0066] 3) Select the sample point with the maximum distance as the cluster center;

[0067] 4) Repeat (2) and (3) until the number of clusters k is reached;

[0068] 5) Use these k cluster centers as the initialized cluster centers to run the k-means algorithm;

[0069] Through model training of 50,000 users, a k++ clustering algorithm that conforms to the users' permanent work and residential communities is obtained.

[0070] Calculation method of user resident cell:

[0071] 1. Calculate the msisdn, cell_id, traffic, and duration data from 9:00 to 12:00 and from 22:00 to 5:00 every day.

[0072] 2. Summarize the data of the previous month every January:

[0073] Summarize the msisdn, cell, and traffic from 9:00 to 12:00 to obtain the top 5 cells with the highest traffic at the workplace.

[0074] 3. Summarize the msisdn, cell, and duration from 9:00 to 12:00 to obtain the top 5 cells with the longest working time.

[0075] 4. Summarize the msisdn, community, and traffic from 22:00 to 5:00 in the morning to obtain the top 5 residential communities with the most traffic.

[0076] 5. Summarize the msisdn, community, and duration from 22:00 to 5:00 in the morning to obtain the top 5 communities with the longest residence time.

[0077] List of algorithms:

[0078]

[0079] 1.3 Introduction to the Construction of User Perception Portrait Model

[0080] Taking the customer's perception characteristics as the research object, through 8 specific tasks in 3 stages including data acquisition, modeling research and portrait establishment, we can achieve quantitative scoring of customer perception and finally form a user perception portrait model based on the community.

[0081] The overall work is divided into three phases and seven specific tasks:

[0082] Phase 1, data acquisition: ① Use various available data sources to acquire and preprocess input data. Obtain raw data without user awareness through surveys; ② Ensure the accuracy of research input data through data cleaning; ③ Also consider other dimensions, such as user-specific data, application-related content, and even data from outside the operator (such as planned events, weather forecasts, etc.).

[0083] Phase II, modeling research: ① Through the algorithm analysis of relevant resident communities, screen out various perception indicators related to user perception, complete the establishment of the model indicator set, take the user's resident community as the research object, through the specific work of data acquisition, modeling research, and portrait establishment, realize the quantitative scoring of customer perception, and finally form a user perception portrait based on perception, and finally focus on analyzing the user's problem community (permanent community).

[0084] Combined with machine learning and multiple rounds of data accumulation, the service weight is established based on the proportion of service time users spend in different communities, and the precise residential community of each user is output.

[0085] Phase 3: Portrait creation: Through algorithm standardization, a complete set of quantifiable computing models is formed; through model calibration, the portrait of the user’s permanent community is finally completed.

[0086] In the discovery phase of AI, machine learning is used to mine the input data and extract relevant knowledge models at different levels: cell level (containing characteristics of conditions on a cell-by-cell basis), cell cluster level (characteristics of cell groups built based on cell similarities), and user level (characteristics of conditions experienced by individual users). The overall goal of machine learning is to build computer models that adapt and learn from their experiences.

[0087] Specific machine learning capabilities of interest include:

[0088] 1) Classification Algorithm: It is the process of finding a model or function that describes and distinguishes data classes or concepts. The obtained model (i.e., classifier) ​​is then used to determine the class to which the object belongs. An object is an entity to be classified, and it is usually represented by a tuple including a set of attribute values. The classification process assumes that the possible classes are predefined in advance. Then, the classifier model is usually obtained from a supervised learning algorithm that analyzes a set of training tuples associated with known classes.

[0089] The community algorithm construction module is used to build a multivariate linear regression model algorithm using the tensorflow framework during the algorithm construction process to obtain the weight of the impact of each KPI indicator on the resident community. At the same time, it selects and compares multiple machine learning algorithms, selects the CART decision tree algorithm with the best evaluation effect to train the ternary classification model, and identifies each community by calculating the information gain of each node.

[0090] 2) Clustering algorithms: Group a set of objects so that objects in the same cluster are similar to each other and different from objects in other clusters. The clustering mechanism does not rely on a set of predefined classes, but these classes are obtained through an unsupervised learning process.

[0091] 1.4 Introduction to User Perception Portrait Analysis Model and Perception Scoring Algorithm:

[0092] A user perception portrait analysis model based on user perception, the analysis includes:

[0093] Permanent cell indicator analysis is used to analyze the network-side causes of current number portability, analyze the network indicators of the user's permanent cell, analyze the correlation of network-side factors affecting user perception, solve problem cells, and improve network indicators;

[0094] The customer perception portrait unit is used to take the perception characteristics of unaware customers as the research object, and through 4 stages and 9 specific tasks including data acquisition, modeling research, and portrait establishment, it realizes the quantitative scoring of customer perception and finally forms a user perception portrait based on perception.

[0095] Customer perception portrait establishment process:

[0096] Phase 1 Data Acquisition: A total of 15,000 perceived user data were cleaned. Analysis found that the main factor affecting user perception on the network side was the perception or non-perception of network usage, accounting for 39%, and Internet access issues were the main problem.

[0097] Phase 2 Modeling Research: Lay the foundation for the perception quantification algorithm by studying the indicator set, weights, factors and thresholds.

[0098] Initial indicator set unit: The signaling data is used to correlate the data of unaware users, and 24 relevant service perception indicators of unaware users are screened. Machine learning is performed and 15 indicators are finally determined, as shown in the following table:

[0099] Table 1 Perception index

[0100]

[0101] Threshold unit: The initial threshold of the portrait is obtained by taking the signaling data of the surveyed unaware users, and obtaining the values ​​of the sub-item indicators of all unaware users through machine learning operations, which are preliminarily determined as the threshold of the perception portrait model (full score penalty and 0 score penalty).

[0102] Perception factor unit: The perception impact factor is based on the description of the specific problem phenomenon provided by the unaware user, combined with the degree of impact on user perception at different stages of the business, and the probability of the problem occurring, to determine the weight of each business sub-item indicator. The self-organizing map algorithm is used to obtain the indicator weight of each perception indicator.

[0103] The self-organizing map (SOM) algorithm is an unsupervised neural network model that finds the weights in a set of data by self-organizing the input data. The SOM algorithm clusters the input data on a two-dimensional grid to achieve data classification and weight calculation.

[0104] Table 2 Penalty values ​​of each perception index

[0105]

[0106] Following are the steps of SOM algorithm to find a set of data weights:

[0107] Initialization: First, choose a suitable grid size and initialize each node in the grid as an independent neuron. Each neuron has a weight vector whose size is the same as the input data.

[0108] Input Data: Then, the input data is given to the SOM algorithm. This data can be in any form like numbers, text, images, etc.

[0109] Calculate the distance: For each input data, calculate its distance to each neuron. Different methods can be used to calculate the distance, such as Euclidean distance or Manhattan distance.

[0110] Select the best matching unit: After calculating the distance, find the neuron that is closest to the input data, which is the best matching unit (BMU).

[0111] Update weights: The input data is then compared with the BMU’s weight vector and the BMU’s weight vector is updated based on the difference between them. The updating method is to use a strategy called Winner-Take-All.

[0112] Repeat step: The input data is repeatedly provided to the SOM algorithm until all the input data has been processed.

[0113] Output: Ultimately, the SOM algorithm generates a two-dimensional grid where each node has a weight vector. These weight vectors can be used to represent patterns in the input data and can be used to classify or cluster the input data.

[0114] Table 3 Weights of each perception index

[0115]

[0116] Service weight unit: refers to the ratio of a single service duration to the total duration, that is, the user's service duration is summarized, and the ratio of a single service duration to the total duration is the user's service weight K; it determines the impact of each service on the user.

[0117] Table 4 Weights of various Internet services

[0118]

[0119] Phase 3: Portrait building unit: Taking into account the impact of business weights, business indicators, and business quality in different time periods on perception, a comprehensive scoring system of hourly aggregation, perception stratification, and business weights was developed to quantify user perception.

[0120] Scoring Algorithm Unit:

[0121]

[0122] Specifically, the user perception scoring algorithm unit further includes:

[0123] Perception factor analysis module: Based on the use of big data algorithm learning from more than 30 data business indicators in the initial indicator set, the key customer perception indicators are determined. There are 15 data business perception factors, and these data business perception indicators have the strongest correlation with customer perception.

[0124] Portrait construction algorithm module: During the algorithm construction process, the tensorflow framework is used to build a multivariate linear regression model algorithm to obtain the influence weight of each perceived KQI. At the same time, a variety of machine learning algorithms are selected and compared, and the CART decision tree algorithm with the best evaluation effect is selected to train the ternary classification model. The perceptual inflection point of each KQI indicator is identified by calculating the information gain of each node.

[0125] Portrait calibration and tuning module: To prevent the model from overfitting, the empirical perception constant Δx is introduced as a reference to effectively improve the robustness of the model. At the same time, non-survey data is introduced to perform bidirectional iterative calibration of the model to improve the accuracy and universality of model evaluation. The threshold and algorithm are calibrated to maximize the matching between the model output and the user's perception, and to continuously improve the accuracy of the perception portrait.

[0126] Accumulate previous survey data, continuously modify model variables, iterate bidirectionally, dynamically adjust perception thresholds, achieve model adaptive optimization, and improve model accuracy;

[0127] Cell perception score calculation:

[0128]

[0129] The cell perception score is an average of the perception scores of resident users in the cell.

[0130] 1.6 Analysis output of users dissatisfied with data services:

[0131] By performing aggregate analysis of various indicators on the permanent cells where the users work and live, users who are dissatisfied with data service perception and potential number portability users are found.

[0132] Innovations of the present invention:

[0133] 1. Judgment method based on the user's permanent residential area. Combining machine learning of various data, through multiple rounds of data accumulation, model output and algorithm output of permanent residential area analysis, the user's permanent residential areas ranked at the top within 3 months are analyzed, including permanent residential area_residence and permanent residential area_workplace.

[0134] 2. User perception portrait model and calculation method based on user perception. Based on survey data analysis, we associate perception KQI big data, various attribute data and performance data, conduct research on perception portrait and attribute portrait, and establish the data source of perception system scoring by establishing the indicator set. At the same time, through algorithm research, we establish the factor W, weight K, threshold T, solidify a set of standard algorithms, and perform penalty tuning and calibration to establish a user perception satisfaction scoring system. We achieve quantitative scoring of customer perception and ultimately form a user perception portrait model based on user perception. Through continuous model correction, we eventually output a method for calculating user perception scores and community perception scores.

[0135] The embodiments described above are only preferred specific implementation modes of the present invention, and the protection scope of the present invention is not limited thereto. Any simple changes or equivalent replacements of the technical solutions that can be obviously obtained by any technician familiar with the field within the technical scope disclosed in the present invention belong to the protection scope of the present invention.

Claims

1. A method for optimizing Internet access experience based on user perception, characterized in that: include: Obtain various perception indicators that affect mobile Internet users' perception; Collect historical data of various perception indicators, input the historical data into the clustering algorithm, output the perception indicator value when the perception is unsatisfactory as the perception inflection point and penalty value; determine the score corresponding to the perception indicator value according to the perception inflection point and penalty value; Input historical data into the self-organizing map algorithm, output the weights of various perception indicators that affect user perception; use the scores and weights corresponding to the perception indicator values ​​to calculate the overall perception score of users in Internet access services; Determine the service proportion of each Internet service of the user; wherein the service proportion is the ratio of the duration of a single service to the total network usage duration; The business proportion of all businesses is multiplied by the overall perception score of online services and then added together to obtain a user perception portrait model; Use the user perception portrait model to score user perception and find low-scoring users; And optimize the Internet experience for low-scoring users.

2. According to the method for optimizing Internet access experience based on user perception in claim 1, it is characterized in that: The network optimization of the cell where the low-scoring user is located comprises: Based on the calculated overall perception score of users in Internet access services, identify users with scores of 60 or below 50 as low-score users; Obtain the traffic and duration data of the cell where the low-score user uses the network; find the cell with traffic and duration higher than the threshold as the low-score user's permanent cell; The network perception of the resident communities is scored, and the network is optimized for the resident communities with low scores.

3. The method for optimizing Internet access experience based on user perception according to claim 2, characterized in that: The method of finding the cell with the highest traffic and duration as the permanent cell of the low-score user includes: The traffic,duration data of the cells are input into the k++ clustering algorithm,,and the cells with traffic duration higher than the threshold are output,by the k++ clustering algorithm as the permanent cells of low-score users.

4. The method for optimizing Internet access experience based on user perception according to claim 1, characterized in that: The network perception of the resident cell is scored, and the calculation formula is: Wherein, n is the average perception score of users in the resident cell.

5. The method for optimizing Internet access experience based on user perception according to claim 2, characterized in that: Also includes: Obtain network B-domain data and O-domain data of low-scoring users; According to the network B domain data and O domain data, the user's area is associated with the location in the map, and a user trajectory evaluation model is constructed to observe the permanent campus of low-scoring users from the map.

6. The method for optimizing Internet access experience based on user perception according to claim 1, characterized in that: The user perception portrait model is:

7. The method for optimizing Internet access experience based on user perception according to claim 4, characterized in that: Also includes: According to the business proportion and the overall perception score of Internet services, the user perception portrait model is built using the tensorflow framework.

8. The method for optimizing Internet access experience based on user perception according to claim 1, characterized in that: The historical data of various perception indicators are collected and input into a clustering algorithm, and the clustering algorithm is a Kmeans algorithm.

9. The method for optimizing Internet access experience based on user perception according to claim 1, characterized in that: The perception inflection point is the perception value at which the perception starts to rise or fall, and the penalty value is the perception value when the perception is between 0 and 100 points.

10. The method for optimizing Internet access experience based on user perception according to claim 1, characterized in that: The perception indicators include: Page response success rate, page response time, page display success rate, instant messaging success rate, instant messaging server side rtt, instant messaging terminal side rtt, instant messaging TCP link establishment delay, video service pause frequency per minute, video pause delay ratio, video initial playback success rate, video initial cache delay, Game service response success rate, Game service response average delay, game freeze rate.

Citation Information

Patent Citations

  • Method and device for realizing web browsing service perception evaluation based on DPI

    CN110535684A

  • Power grid customer grading method and system based on big data analysis, computer equipment and storage medium

    CN111275485A

  • User perception quantification method and device, computing equipment and computer storage medium

    CN113052412A

  • Grid management optimization method and system based on multi-dimensional portrait evaluation

    CN114547452A

  • User perception evaluation method and device and electronic equipment

    CN115941517A

Cited By

  • Data-driven poor-quality root cause reasoning method and system

    CN120856541A

  • A data-driven quality difference root cause reasoning method and system

    CN120856541B