Broadband user perception evaluation method and device and medium

Abstract

The invention provides a broadband user perception evaluation method and device and a medium, and relates to the technical field of computers. The method comprises the following steps: acquiring a plurality of perception indexes of each dimension of a plurality of broadband users from a plurality of dimensions, wherein the perception indexes comprise non-label perception indexes; obtaining an index score of each perception index of each broadband user, including obtaining an index score of each non-label perception index of each broadband user according to a plurality of feature quantiles of perception distribution of a plurality of broadband users to each non-label perception index; and obtaining a dimension score of each dimension of each broadband user according to the index scores, and obtaining a total score of perception of each broadband user according to the dimension scores. According to the method, the multi-perception indexes are obtained from multiple dimensions, the score of each non-label perception index is obtained by adopting the characteristic quantile, the score of multiple dimensions is obtained by integrating the multiple indexes, the score of the overall perception of the user is obtained by integrating the multiple dimensions, and the broadband user perception evaluation result of the system can be obtained.

Description

Technical Field

[0001] This application is at least related to the field of computer technology, and particularly relates to a broadband user perception evaluation method, device and medium. Background Art

[0002] In the face of the increasingly complex broadband networking environment, the data obtained by traditional broadband user perception evaluation methods is difficult to be associated with the true perception of users. There may be a situation where the KPI (Key Performance Indicator) statistically by the network management is very good, but the result of the user satisfaction survey shows that users are not satisfied with the Internet access quality. Summary of the Invention

[0003] In view of the above deficiencies, this application provides a broadband user perception evaluation method, device and medium to solve the following technical problems: how to obtain user perception indicators from multiple dimensions and obtain index scores according to user perception, so as to obtain a more comprehensive broadband evaluation result that can reflect user perception.

[0004] In a first aspect, this application provides a broadband user perception evaluation method, and the method includes:

[0005] Obtain multiple perception indicators of each dimension of multiple broadband users from multiple dimensions, and the perception indicators include non-label perception indicators;

[0006] Obtain the index score of each perception indicator of each broadband user, including obtaining the index score of each non-label perception indicator of each broadband user according to multiple characteristic quantiles of the perception distribution of each non-label perception indicator by multiple broadband users;

[0007] Obtain the dimension score of each dimension of each broadband user according to the index score, and obtain the overall score of the perception of each broadband user according to the dimension score.

[0008] Further, obtaining multiple perception indicators of each dimension of multiple broadband users from multiple dimensions specifically includes:

[0009] Obtain all broadband users of the operator within a preset range;

[0010] Obtain multiple perception indicators of each dimension of each broadband user from multiple dimensions at a preset frequency of each perception indicator within a preset time period;

[0011] Among them, the multiple dimensions include the user side, the network side, the application side and the service side, and the perception indicators further include label perception indicators. The label representation of the label perception indicators is yes or no, and the corresponding index scores are 100 points or 0 points.

[0012] Further, obtain multiple perception metrics for each broadband user from multiple dimensions, specifically including:

[0013] Obtain the service B domain data, operation O domain data, complaint data, order dispatching data, and alarm data of the operator within a preset range. Among them, the O domain data includes SA data collected by setting a service assurance SA board on the broadband access server BRAS or broadband gateway BNG in the metropolitan area network;

[0014] Obtain multiple package adaptation parameters for each broadband user according to the B domain data, and determine whether the mobile hotspot WiFi performance, first part performance of the optical network terminal, and router performance on the user side of each broadband user meet the corresponding package adaptation parameters according to the O domain data, so as to obtain multiple tag - type perception metrics on the user side;

[0015] Obtain the measured values of the second part performance of the optical network terminal and the IPTV performance on the user side of each broadband user, and the measured values of the optical line terminal OLT device performance, switch performance, and broadband access server BAS performance on the network side according to the O domain data, so as to obtain multiple non - tag - type perception metrics on the user side and the network side;

[0016] Obtain the application metrics and poor - quality application metrics for each broadband user according to the SA data, so as to obtain multiple non - tag - type perception metrics on the application side, and obtain the metrics representing the services obtained by each broadband user according to the complaint data, order dispatching data, and alarm data, so as to obtain multiple non - tag - type perception metrics on the service side.

[0017] Further, obtain the metric scores for each non - tag - type perception metric of each broadband user according to multiple characteristic quantiles of the perception distribution of multiple broadband users for each non - tag - type perception metric, specifically including:

[0018] Obtain the perception evaluations of N' broadband users on their own use of broadband. The perception evaluations are divided into M≥3 levels, and obtain the percentage p of broadband users in each perception evaluation level m∈[1,M]; m ;

[0019] Sort the non - tag - type perception metrics with the same ji of all N broadband users according to the numerical size as where n∈[1,N] represents each broadband user, i∈[1,I] represents each dimension, and ji∈[1,Ji] represents each perception metric of each dimension;

[0020] Obtain the percentage p of broadband users in the perception evaluation level m∈[1,M - 1] m At the demarcation position among N broadband users, obtain the values of M - 1 non - tag - type perception metrics located at the corresponding demarcation position after sorting as the characteristic quantiles

[0021] For each characteristic quantile Assign scores between 0 and 100, and use interpolation between characteristic quantiles to obtain each Index score between 0 and 100

[0022] Furthermore, where:

[0023] The perceived evaluation is divided into 4 levels: excellent, good, medium, and poor;

[0024] p m The demarcation position of is n = ceil(p m *N), where ceil means rounding up by 1 digit;

[0025] Less than or equal to / greater than or equal to of The index score of is 0 points, equal to of The index score of is 60 points, greater than or equal to / less than or equal to of The index score of is 100 points.

[0026] Furthermore, obtain the dimension score of each dimension of each broadband user according to the index score, and obtain the overall perceived score of each broadband user according to the dimension score, specifically including:

[0027] Obtain the index weight of each perceived index according to the difference degree of the index scores of the same perceived index of multiple broadband users, obtain the dimension score of each dimension of each broadband user according to each index score and each index weight, obtain the dimension weight of each dimension according to the difference degree of the dimension scores of the same dimension of multiple broadband users, and obtain the overall perceived score of each broadband user according to each dimension score and each dimension weight.

[0028] Furthermore, obtain the index weight of each perceived index according to the difference degree of the index scores of the same perceived index of multiple broadband users, obtain the dimension score of each dimension of each broadband user according to each index score and each index weight, obtain the dimension weight of each dimension according to the difference degree of the dimension scores of the same dimension of multiple broadband users, and obtain the overall perceived score of each broadband user according to each dimension score and each dimension weight, specifically including:

[0029] Calculate the information entropy of the perceived index of each ji ∈ [1, Ji] among all N broadband users where k = 1 / ln(N), Calculate the index weight of the perceived index of each ji ∈ [1, Ji] where dji = 1 - e ji , calculate the dimension score of each broadband user in each dimension

[0030] Calculate the information entropy of each dimension among all N broadband users for each i ∈ [1, I] where k = 1 / ln(N), Calculate the dimension weight of each dimension for each i ∈ [1, I] where d i = 1 - e i , calculate the perceived overall score of each broadband user

[0031] Furthermore, the method further includes:

[0032] Predict the occurrence of network faults and / or locate the location where network faults occur based on the metric scores, dimension scores, and overall scores of multiple broadband users;

[0033] Obtain the metric scores, dimension scores, and overall scores of multiple broadband users during the period of using a certain network service, and evaluate the quality of a certain network service based on the metric scores, dimension scores, and overall scores corresponding to the period;

[0034] Market network services to a certain broadband user based on the metric scores, dimension scores, and overall scores of the certain broadband user.

[0035] In a second aspect, the present application provides a broadband user perception evaluation device, and the device includes:

[0036] An acquisition module, configured to acquire multiple perception metrics of each dimension of multiple broadband users from multiple dimensions, and the perception metrics include non-label perception metrics;

[0037] A metric score module, connected to the acquisition module, configured to obtain the metric scores of each perception metric of each broadband user, including obtaining the metric scores of each non-label perception metric of each broadband user according to multiple characteristic quantiles of the perception distribution of each non-label perception metric by multiple broadband users;

[0038] A comprehensive score module, connected to the metric score module, configured to obtain the dimension scores of each dimension of each broadband user according to the metric scores, and obtain the perceived overall score of each broadband user according to the dimension scores.

[0039] In a third aspect, the present application provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is run by a processor, the broadband user perception evaluation method as described above is implemented.

[0040] The present application provides a broadband user perception evaluation method, device, and medium, which obtain multi-perception indicators from multiple dimensions, obtain characteristic quantiles according to the perception distribution of multiple broadband users, use the characteristic quantiles to obtain the scores of each non-labeled perception indicator, obtain multi-dimensional scores by integrating multiple indicators, and obtain the overall score of the user's overall perception by integrating multiple dimensions. Three-layer scoring results of the overall score, dimension score, and indicator score can be obtained, and the perception provided by the broadband to the user can be evaluated more comprehensively. Based on the indicator scores obtained from the user perception, the broadband user perception evaluation result can reflect the true experience of the user using the broadband. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 is a flowchart of a broadband user perception evaluation method according to an embodiment of the present application;

[0042] Figure 2 is a schematic structural diagram of a broadband user perception evaluation device according to an embodiment of the present application;

[0043] Figure 3 is an overall architecture diagram of a broadband user perception evaluation method according to an embodiment of the present application;

[0044] Figure 4 is an architecture diagram of a broadband user perception evaluation data acquisition method according to an embodiment of the present application;

[0045] Figure 5 is a flowchart of a broadband user perception evaluation score calculation method according to an embodiment of the present application;

[0046] Figure 6 is a schematic diagram of a broadband user perception evaluation indicator scoring method according to an embodiment of the present application;

[0047] Figure 7 is a schematic diagram of another broadband user perception evaluation indicator scoring method according to an embodiment of the present application;

[0048] Figure 8 is an architecture diagram of a broadband user perception evaluation result application method according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] To enable those skilled in the art to better understand the technical solutions of the present application, the embodiments of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0050] It can be understood that the specific embodiments and drawings described herein are only used to explain the present application, rather than limiting the present application.

[0051] It can be understood that, without conflict, the various embodiments in the present application and the various features in the embodiments can be combined with each other.

[0052] It can be understood that for the convenience of description, only the parts related to the present application are shown in the drawings of the present application, while the parts unrelated to the present application are not shown in the drawings.

[0053] It can be understood that each module and unit involved in the embodiments of the present application may correspond to only one entity structure, or may be composed of multiple entity structures. Alternatively, multiple modules and units may also be integrated into one entity structure.

[0054] It can be understood that without conflict, the functions and steps marked in the flowcharts and block diagrams of the present application may occur in an order different from that marked in the drawings.

[0055] It can be understood that in the flowcharts and block diagrams of the present application, the possible architectures, functions, and operations of the systems, devices, equipment, and methods according to the various embodiments of the present application are shown. Among them, each block in the flowchart or block diagram may represent a module, unit, program segment, or code, which contains executable instructions for implementing the specified function. Moreover, each block or combination of blocks in the block diagram and flowchart may be implemented by a hardware-based device for implementing the specified function, or may be implemented by a combination of hardware and computer instructions.

[0056] It can be understood that the modules and units involved in the embodiments of the present application may be implemented in software or in hardware. For example, the modules and units may be located in the processor.

[0057] Embodiment 1:

[0058] As Figure 1 shown, the present application provides a broadband user perception evaluation method, and the method includes:

[0059] S1. Obtain multiple perception indicators of each dimension of multiple broadband users from multiple dimensions, where the perception indicators include non-label perception indicators;

[0060] S2. Obtain the index scores of each perception indicator of each broadband user, including obtaining the index scores of each non-label perception indicator of each broadband user according to multiple characteristic quantiles of the perception distribution of each non-label perception indicator by multiple broadband users;

[0061] S3. Obtain the dimension scores of each dimension of each broadband user according to the index scores, and obtain the overall score of the perception of each broadband user according to the dimension scores.

[0062] In this embodiment, the method obtains multi-sensing indicators from multiple dimensions, obtains characteristic quantiles according to the perception distribution of multi-broadband users, uses the characteristic quantiles to obtain the scores of each non-labeled sensing indicator, obtains the multi-dimensional scores by integrating multiple indicators, and obtains the overall perception score of the user by integrating multiple dimensions. Three-layer scoring results of the overall score, dimension score, and indicator score can be obtained, which more comprehensively evaluates the perception provided by the broadband to the user. Based on the indicator scores obtained from the user perception, the broadband user perception evaluation result can reflect the real experience of the user using the broadband. As Figure 1 The method shown is correspondingly applied to the device as shown in Figure 2 the device shown.

[0063] More specifically, this embodiment provides a method and device for evaluating and diagnosing the broadband user perception portrait, which can solve the following pain points: The home broadband networking scenario is complex, and the types of terminals in the home network are numerous and complex, resulting in large broadband traffic, complex service types, and high concurrency. In the scenario with many family members, the usage time and types of services are also different, resulting in differences in the perception of service usage; The monitoring of service quality is insufficient, and it cannot comprehensively reflect the usage experience of a certain service application, resulting in the inability to know, feel, and control the service situation, lacking real, objective, and complete user experience data, and unable to effectively serve, support, and manage the service; The ability of fault prediction and rapid diagnosis is lacking. Due to the large number of third-party devices in the home network, problems such as mismatched network cable specifications, loose connectors, router aging, and unstable Wi-Fi (mobile hotspot) signals can affect the Internet access experience. However, currently, in terms of fault prediction and rapid diagnosis, it often relies on the maintenance engineer to visit the site to locate the fault based on manual experience, and this passive and inefficient operation and maintenance mode seriously affects the operation and maintenance efficiency and user satisfaction; The user experience is difficult to quantify. Due to reasons such as insufficient statistical granularity and incomplete statistical dimensions, the home broadband network has the problem that traditional network indicators cannot represent the real experience of users. Network indicators such as coverage and interference cannot be directly equated with user experience problems such as "lagging, disconnection, and slow speed" of applications. The operation and maintenance often find that the KPIs counted by the network management are good, but the results of the user satisfaction survey show that users are not satisfied with the Internet access quality; The operation and maintenance workload is large and the cost is high. Due to the large number and complexity of devices in the home network, the maintenance personnel of the operator need to visit frequently to handle faults, resulting in a large operation and maintenance workload and high cost; The perception evaluation seriously relies on manual experience judgment. The traditional perception problem troubleshooting collects the feedback of users on broadband perception through questionnaires and telephone interviews, and records detailed information, including problem descriptions, occurrence times, influence scopes, etc. The perception evaluation takes a long time, is difficult to locate, and deviates from the concept of network self-intelligence.

[0064] The overall architecture of this embodiment is as shown in Figure 3As shown in the figure, it includes: selecting multiple indicators from the home side (user side), network side, application side, and service side respectively to construct a perception evaluation index system; conducting single-index perception evaluation, sub-dimension perception evaluation, and overall perception evaluation for each single user respectively to obtain multi-level evaluation results; pushing the evaluation results to the intelligent home customer service market middle platform, and the evaluation results can be used for fault location, network repair, marketing, etc. The following technical effects are achieved: constructing a broadband user perception portrait diagnosis, and systematically analyzing the relevant data of the user's home side, network side, service side, and business side; providing the real-time processing ability of the perception diagnosis model, normalizing and storing the complex data after real-time processing, ensuring the accuracy of the data dimension timeliness, realizing the multi-dimensional real-time one-key diagnosis and strategy matching of perception problems, and ensuring the accuracy of perception problem location; through the user quality evaluation algorithm and the available data sources, outputting the all-round evaluation portrait and quality difference data of the user, which are used for the home customer professional maintenance personnel and installation and maintenance personnel to carry out complaint handling, precise quality difference rectification, fault handling, and quality closed-loop management, etc.; based on the portrait evaluation method, optimizing the corresponding process system, and realizing the broadband user perception location analysis and the matching of the quality difference closed-loop rectification strategy.

[0065] In one implementation, in S1, obtaining multiple perception indicators of each dimension of multiple broadband users from multiple dimensions specifically includes:

[0066] Obtaining all broadband users of the operator within a preset range;

[0067] Obtaining multiple perception indicators of each dimension of each broadband user from multiple dimensions at the preset frequency of each perception indicator within a preset time period;

[0068] Among them, the multiple dimensions include the user side, network side, application side, and service side, and the perception indicators also include label-type perception indicators. The label representation of the label-type perception indicators is yes or no, and the corresponding indicator scores are 100 points or 0 points.

[0069] In this embodiment, as Figure 3As shown in the figure, the construction plan for the broadband user perception portrait diagnosis includes the following steps: the construction of a perception index system, the construction of a single-index perception evaluation model, the construction of a sub-dimension perception evaluation model, the construction of an overall perception evaluation model, quality difference diagnosis and strategy matching, and the closed-loop of quality difference repair. The factors affecting user perception are diverse and complex, running through the entire end-to-end process of broadband users. In order to accurately evaluate the perception of each link and process of broadband users, a construction method for the broadband user perception portrait diagnosis is proposed, which integrates the multi-dimensional data collaboration of 16+ systems in the O / B domain (the data domain of the business support system, also known as the business domain, abbreviated as BSS; the data domain of the operation support system, also known as the operation domain, abbreviated as OSS), establishes an index lake on the home side, network side, application perception side, and service side, selects more than 40 indicators to establish a set of home broadband service perception evaluation criteria, and conducts application perception evaluation and diagnosis on broadband users across the province.

[0070] More specifically, data is the foundation for model construction, and data problems affect the prediction results of the model to a certain extent. Given the current data situation, different data comes from different platforms, and the data upload specifications are not unified. At the same time, the data on different platforms is severely lacking, and coupled with the problem of mismatched cycles, these scattered data cannot construct a complete broadband user perception index system because there is no unified platform for standardized management. Based on this starting point, a certain communication operator coordinated multiple departments and manufacturers to conduct unified data collection, sorting, and spent a lot of time on the normalized management of perception indicators, and constructed a broadband user perception index system as Figure 4 shown in the figure. Necessary data processing is carried out during the construction of the user perception index system, including: obtaining raw data, obtaining home broadband data of different types and different time dimensions from different platforms and performing standardized access processing; data verification and cleaning, preprocessing each type of raw data after data access to ensure the accuracy and consistency of the data. Specific operations may include removing duplicates, filling in missing values, converting data formats, unit conversion, verifying the integrity and accuracy of the data, etc. These operations help reduce noise and anomalies in the data and provide a reliable basis for subsequent data analysis; data parsing and aggregation. After data cleaning, the data parsing step is responsible for extracting valuable information, which may include indicator calculation, dimension integration, key information extraction, classification, etc. The purpose is to mine useful information from the raw data and form usable and reliable model tables according to different data types, and then through steps such as data association to further integrate and enrich the data set to provide support for subsequent data analysis and model construction; output the broadband perception index model set. After a series of the above data processing, a perception evaluation index model set is formed according to the input requirements of the model indicators.

[0071] In one embodiment, multiple perception metrics of each broadband user are obtained from multiple dimensions, specifically including:

[0072] Obtain the service B domain data, operation O domain data, complaint data, order dispatch data, and alarm data of the operator within a preset range. Among them, the O domain data includes SA data collected by setting a service assurance (SA) board on the broadband access server (BRAS) or broadband network gateway (BNG) in the metropolitan area network;

[0073] Obtain multiple package adaptation parameters of each broadband user according to the B domain data, and determine whether the mobile hotspot WiFi performance, first part performance of the optical network terminal (ONT), and router performance on the user side of each broadband user meet the corresponding package adaptation parameters according to the O domain data, so as to obtain multiple tag - type perception metrics on the user side;

[0074] Obtain the measured values of the second part performance of the ONT and the Internet Protocol Television (IPTV) performance on the user side of each broadband user, and the measured values of the optical line terminal (OLT) device performance, switch performance, and broadband access server (BAS) performance on the network side according to the O domain data, so as to obtain multiple non - tag - type perception metrics on the user side and the network side;

[0075] Obtain the application metrics and poor - quality application metrics of each broadband user according to the SA data, so as to obtain multiple non - tag - type perception metrics on the application side, and obtain the metrics representing the services obtained by each broadband user according to the complaint data, order dispatch data, and alarm data, so as to obtain multiple non - tag - type perception metrics on the service side.

[0076] In this embodiment, since there are many features obtained when building the broadband user perception model, using all features to calculate the metric scores will affect the calculation efficiency. Therefore, principal component analysis (PCA) is used here for feature selection. As Figure 3 shown, finally, a home - side metric model set (28), a network - side metric model set (12), an application - side metric model set (6), and a service - side metric model set (6) are determined among numerous metrics, with a total of 4 dimensions as the final input of the broadband user perception portrait diagnosis model. Among them, the tag - type perception metrics are obtained by associating with the Resource Management System (rms) and the B domain data. The B domain data contains information such as the service activation information of users, and information such as their package requirements can be obtained. The tag - type perception metrics are the metrics with values of 0 or 1 in Table 1 below. 0 and 1 respectively represent two situations of not meeting the requirements or meeting the requirements. The remaining metrics that can have multiple numerical value results and the numerical values represent the magnitudes of the measured values are called non - tag - type perception metrics; Obtain the user application perception - side evaluation dimension data through SA (Service Assurance) coverage. For example,Figure 4 As shown in the figure, SA is a service assurance solution provided by Huawei. By adding a SA (model VSUI-400A) board to the BRAS (Broadband Remote Access Server) / BNG (Broadband Network Gateway) devices in the metropolitan area network, the SA board collects user service traffic metrics, device performance metrics, etc. for analysis. By real-time monitoring, analyzing, and optimizing network performance, it ensures that users obtain a high-quality service experience. From four dimensions of the home side, network side, application awareness side, and service side, a broadband user perception evaluation index model is formulated based on single-index thresholds and linear algorithms, supporting one-key triggering of important indicators, matching repair strategies for quality problems, and supporting front-line rectification and marketing of broadband user perception problems. The complete perception index system is shown in Table 1 below:

[0077] Table 1 Example of Perception Index System

[0078]

[0079]

[0080]

[0081] In one implementation, in S2, according to multiple characteristic quantiles of the perception distribution of each non-labeled perception index by multiple broadband users, the index score of each non-labeled perception index of each broadband user is obtained, specifically including:

[0082] Obtain the perception evaluations of N' broadband users on their own broadband usage. The perception evaluations are divided into M≥3 levels, and obtain the percentage p of broadband users in each perception evaluation level m∈[1,M] m ;

[0083] Sort the non-labeled perception indexes with the same ji of all N broadband users according to the numerical size as where n∈[1,N] represents each broadband user, i∈[1,I] represents each dimension, and Ji∈[1,Ji] represents each perception index of each dimension;

[0084] Obtain the percentage p of broadband users in the perception evaluation level m∈[1,M-1] m At the demarcation position among N broadband users, obtain the values of the M-1 non-labeled perception indexes located at the corresponding demarcation position after sorting as the characteristic quantiles

[0085] For each characteristic quantile Assign scores between 0 and 100, and use interpolation method between characteristic quantiles to obtain each Index score between 0 and 100

[0086] In this embodiment, to measure whether the broadband user perception is accurate, the algorithm idea of the model is crucial. Through continuous polishing, an algorithm for evaluating broadband user perception is sorted out, specifically as Figure 5 shown. Among them, the single-index algorithm model is specifically to determine the final input index set of the broadband perception portrait diagnosis model, which is divided into two types of indicators according to the characteristics of the indicators. One is the label type indicator (i.e., the result = yes / no), and the other is the non-label type indicator (such as delay, success rate, etc.). The score of the label type indicator is based on the two output results, only 100 points and 0 points; the score of the non-label indicator is interpolated and scored, associating with user perception. For example, 100 user perception evaluations are obtained through investigation. Among them, 5% of the users report poor network quality, 5% of the users report excellent network quality, 35% of the users report average network quality, and 55% of the users report good network quality. Then, according to this proportion, the characteristic quantiles of each user perception index are obtained among all users. For each perception index, the worst 5% of the perception indexes among all users are marked with 0 points, and the demarcation point of the upper 35% and 55% of the users is set as a quantile value, such as 60 points. The perception index scores of 35% of the users are between 0 and 60 points, the perception index scores of 55% of the users are between 60 and 100 points, and the optimal 5% of the user index scores are 100 points. The scores of different indexes of the same user may be different, but the overall conforms to this distribution law, so as to reflect the user perception evaluation obtained by investigation on the network indexes.

[0087] In one implementation, where:

[0088] The perception evaluation is divided into 4 levels: excellent, good, medium, and poor;

[0089] p m The demarcation position is n = ceil(p m *N), where ceil means rounding up by 1 bit;

[0090] Less than or equal to / greater than or equal to of The index score is 0 points, equal to of The index score is 60 points, greater than or equal to / less than or equal to of The index score is 100 points.

[0091] In this embodiment, for example, the optimal 5% and the worst 5% of each indicator of all users in 7 days of a week are averaged and used as 100 points and 0 points respectively; the median and average value of each indicator of all users in 7 days of a week are used as a reference for setting the 60-point threshold of a single indicator. Finally, the score of a single indicator is output according to the following algorithm idea: Sort the indicator data set in ascending order; for the characteristic percentile p m , calculate the index position n = (p m * N); determine the percentile value. If the index position is an integer, the percentile is the data point at position n. If it is not an integer, the percentile can be obtained by interpolation or by rounding up or down. The index position represents the point where the data changes drastically; calculate the score according to the percentile. According to the nature of the indicator, it can be divided into indicators that are better when larger and indicators that are better when smaller. For example, the bit error rate of the PON (Passive Optical Network) port is better when smaller, and the work order completion rate is better when larger, etc.

[0092] The calculation of the indicator that is better when larger is as Figure 6 shown, and the formula is as follows:

[0093]

[0094] The calculation of the indicator that is better when smaller is as Figure 7 shown, and the formula is as follows:

[0095]

[0096] In an implementation manner, S3. Obtain the dimension score of each dimension of each broadband user according to the indicator score, and obtain the overall perceived score of each broadband user according to the dimension score, specifically including:

[0097] Obtain the indicator weight of each perceived indicator according to the difference degree of the indicator scores of the same perceived indicator of multiple broadband users, obtain the dimension score of each dimension of each broadband user according to each indicator score and each indicator weight, obtain the dimension weight of each dimension according to the difference degree of the dimension scores of the same dimension of multiple broadband users, and obtain the overall perceived score of each broadband user according to each dimension score and each dimension weight.

[0098] In this embodiment, the overall perceived evaluation of users is presented in 4 sub - dimensions (home side, network side, service side, and business side). The score of each sub - dimension is calculated based on the scores and weights of single indicators. The entropy weight method is used to determine the weights of each indicator, and the weighted scores of each category (home side, network side, application side, and service side) are calculated in turn. Finally, a total of three - layer scoring results are obtained, one score for each indicator, one score for each of the four dimensions, and one overall score. The entropy weight method evaluates the information content of each indicator based on the entropy concept in information theory and assigns weights accordingly. The basic idea is that if the degree of variation (i.e., information content) of an indicator is greater, the more information it provides, and the greater the weight of this indicator should be; conversely, if the degree of variation of an indicator is smaller, the less information it provides, and the smaller the weight of this indicator should be. The degree of variation is the difference of the same indicator among different users, which measures the dispersion of a certain broadband performance indicator (such as access bandwidth compliance rate, access delay, packet loss rate, etc.) among different user groups. When the degree of variation of a certain indicator is large, it means that the values of this indicator among different users vary greatly.

[0099] In one embodiment, the indicator weight of each perception indicator is obtained according to the difference degree of the indicator scores of the same perception indicator of multiple broadband users. The dimension score of each dimension of each broadband user is obtained according to each indicator score and each indicator weight. The dimension weight of each dimension is obtained according to the difference degree of the dimension scores of the same dimension of multiple broadband users. The overall perceived score of each broadband user is obtained according to each dimension score and each dimension weight, specifically including:

[0100] Calculate the information entropy of the perception indicator of each \(j_i\in[1,J_i]\) among all \(N\) broadband users where \(k = 1 / \ln(N)\) Calculate the indicator weight of the perception indicator of each \(j_i\in[1,J_i]\) where \(d\) ji \(= 1 - e\) ji Calculate the dimension score of each dimension of each broadband user

[0101] Calculate the information entropy of each dimension \(i\in[1,I]\) among all \(N\) broadband users where \(k = 1 / \ln(N)\) Calculate the dimension weight of each dimension \(i\in[1,I]\) where \(d\) i \(= 1 - e\) i Calculate the overall perceived score of each broadband user

[0102] In this embodiment, the steps of calculating the indicator weight by the entropy weight method are as follows:

[0103] Calculate the information entropy: where k > 0, and generally k = 1 / ln(N), is the index score of the jith index of the nth sample data (user), and converting it to a percentage value (i.e., a value between 0 and 1) for calculation can unify the symbols;

[0104] Calculate the entropy weight: d ji = 1 - e ji , and the entropy weight reflects the amount of information of the jith index;

[0105] Calculate the weights of each index: where Ji is the number of indexes in the ith dimension;

[0106] Calculate the scores of each category of users:

[0107] After calculating all the sub - dimension scores of each user according to the above steps, then use the entropy weight method to calculate the weights of the sub - dimension scores, and use weighted calculation to obtain the overall score of the user. In addition to calculating the comprehensive score using weights, it can also be like Figure 5 shown, directly use the average to calculate the comprehensive score. Finally, define the index scores, scores of each sub - category, and overall perception scores as excellent, medium, and poor through division. [0, 60) is defined as poor, [60, 90) is defined as medium, and [90, 100] is defined as excellent.

[0108] In one embodiment, the method further includes:

[0109] Predict the occurrence of network faults and / or locate the location where network faults occur based on the index scores, dimension scores, and overall scores of multiple broadband users;

[0110] Obtain the index scores, dimension scores, and overall scores of multiple broadband users during the period of using a certain network service, and evaluate the quality of a certain network service according to the index scores, dimension scores, and overall scores corresponding to the period;

[0111] Market network services to a certain broadband user based on the index scores, dimension scores, and overall scores of the certain broadband user.

[0112] In this embodiment, it is also possible to achieve a closed - loop of quality - difference repair through policy matching, quickly diagnose user perception problems based on an algorithm model, match corresponding repair strategies, and support relevant departments to improve capabilities, that is, the policy adjustment function part as Figure 3 shown. Specifically, it can be like Figure 8As shown, based on the model algorithm, it is used to perceive, diagnose and locate the problems of broadband users, and visually present various possible problems; it associates the diagnosed problems of broadband users with the policy library, matches the disposal suggestions for the corresponding diagnosed problems and presents them visually; it collects the current capacity building requirements of each department of home customers, and then evaluates whether the diagnosis function of the broadband user perception portrait can provide corresponding capacity support. For those that meet the capacity requirements, the interfaces are connected to output the capacity to support capacity improvement. This function can support all customers (marketing), smart home (on-site diagnosis), customer service (complaint support), and also provide the middle platform later (empowering customers through the operator APP); after the strategy is executed, subsequent tracking and evaluation can be carried out. To evaluate the quality of the strategy, a complete evaluation system is required. This system needs to collect various data of the strategy for analysis and calculation. For example, the number of user touches, the number of handled cases, the number of successful cases, the number of consultations, the number of complaints, the subsequent trajectories of users, etc., calculate the evaluation index data, and comprehensively evaluate these indexes. Finally, score this strategy. For those with unsatisfactory evaluation of the specific problems and the corresponding repair strategy effects, continue to polish and optimize the algorithm, and link with the front-line departments in the cities at the strategy level to collect, adjust and optimize more problem strategies until the model is accurate and available.

[0113] Specific application scenarios may include:

[0114] Network fault warning and diagnosis (smart home on-site, customer service complaint support). For network fault warning, by analyzing the network quality data used by broadband users in real time, when abnormal network behavior or quality decline is detected, a warning can be issued in time so as to take measures in advance to avoid the occurrence of network faults;

[0115] Fault problem location: After a network fault occurs, through the user perception portrait, combined with data such as network topology and device status, the root cause of the fault can be quickly located. After calculation, it takes 0.33 hours for the smart home engineer to receive the fault order, analyze and dispose of the fault cause. After the root cause is located through perception diagnosis, the average duration of the work order is shortened by 0.2 hours, and labor costs can also be saved;

[0116] Personalized recommendation (empowering through the middle platform Unicom APP), user experience optimization. By collecting and analyzing the perception data of broadband users when using various services (such as video live broadcast, online games, Internet TV, etc.), the quality of the service can be evaluated, which helps to understand the user's satisfaction with the service, and then optimize the service quality. The optimization suggestions can analyze the source of user anomalies according to the scores of different levels and put forward targeted suggestions to improve the user experience;

[0117] Precision Marketing (All Customers): Based on the user perception portrait, information such as users' interest preferences and usage habits can be understood. According to this information, more precise marketing strategies can be formulated. For example, through the home-side scoring, it can be identified which user families have problems. The network-side scoring mainly evaluates the performance and quality of the operator's network infrastructure. The content-side reflects the performance and quality of the broadband service itself. The service-side scoring mainly evaluates the service quality provided by the operator. According to the user's service usage situation, video membership services can be recommended to users who like to watch videos, and game acceleration services can be recommended to users who like online games, etc.

[0118] The following application effects are achieved in this embodiment: A set of standard broadband perception evaluation index systems are established, including more than 40 indicators in four dimensions: home side, network side, application perception side, and service side; A set of scientific index calculation models are established, with the ability to evaluate the perception of all network users as excellent, medium, and poor, conduct end-to-end perception diagnosis based on the family unit, and output handling suggestions; Based on the perception evaluation index system, the group's broadband perception diagnosis ability is built to empower the market, customer service, smart home, and customers, assist user marketing strategies and customer perception repair and improvement; Realize the ability to assist in dispatching integration of communication and sensing for home-side index diagnosis.

[0119] Embodiment 2:

[0120] As Figure 2 shown, this application provides a broadband user perception evaluation device, and the device includes:

[0121] An acquisition module 1, configured to acquire multiple perception indicators of each dimension of multiple broadband users from multiple dimensions, and the perception indicators include non-labeled perception indicators;

[0122] An index scoring module 2, connected to the acquisition module 1, configured to acquire the index scores of each perception indicator of each broadband user, including acquiring the index scores of each non-labeled perception indicator of each broadband user according to multiple characteristic quantiles of the perception distribution of each non-labeled perception indicator by multiple broadband users;

[0123] A comprehensive scoring module 3, connected to the index scoring module 2, configured to obtain the dimension scores of each dimension of each broadband user according to the index scores, and obtain the overall score of the perception of each broadband user according to the dimension scores.

[0124] In an implementation manner, the acquisition module 1 specifically includes:

[0125] A user acquisition unit, configured to acquire all broadband users of the operator within a preset range;

[0126] A data acquisition unit, connected to the user acquisition unit, is configured to obtain multiple perception indicators of each broadband user in each dimension from multiple dimensions at a preset frequency of each perception indicator within a preset duration;

[0127] Among them, the multiple dimensions include the user side, the network side, the application side, and the service side. The perception indicators also include label-type perception indicators. The label representation of the label-type perception indicators is yes or no, and the corresponding indicator scores are 100 points or 0 points.

[0128] In an implementation manner, the data acquisition unit specifically includes:

[0129] A data collection unit, configured to obtain service B domain data, operation O domain data, complaint data, order arrangement data, and alarm data of an operator within a preset range. Among them, the O domain data includes SA data collected by setting a service assurance SA board on a broadband access server BRAS or a broadband gateway BNG in a metropolitan area network;

[0130] A user-side indicator acquisition unit, connected to the data collection unit, is configured to obtain multiple package adaptation parameters of each broadband user according to the B domain data, and determine whether the mobile hotspot WiFi performance, the first part performance of the optical modem, and the router performance on the user side of each broadband user meet the corresponding package adaptation parameters according to the O domain data, so as to obtain multiple label-type perception indicators on the user side; and obtain the measured values of the second part performance of the optical modem and the network television IPTV performance on the user side of each broadband user according to the O domain data, so as to obtain multiple non-label-type perception indicators on the user side;

[0131] A network-side indicator acquisition unit, connected to the data collection unit, is configured to obtain the measured values of the optical line terminal OLT device performance, switch performance, and broadband access server BAS performance on the network side of each broadband user according to the O domain data, so as to obtain multiple non-label-type perception indicators on the network side;

[0132] An application-side indicator acquisition unit, connected to the data collection unit, is configured to obtain the application indicators and quality-difference application indicators of each broadband user according to the SA data, so as to obtain multiple non-label-type perception indicators on the application side;

[0133] A service-side indicator acquisition unit, connected to the data collection unit, is configured to obtain the indicators representing the services obtained by each broadband user according to the complaint data, order arrangement data, and alarm data, so as to obtain multiple non-label-type perception indicators on the service side.

[0134] In an implementation manner, the indicator scoring module 2 specifically includes:

[0135] A percentage acquisition unit for acquiring the perceived evaluations of N' broadband users on their own broadband usage, where the perceived evaluations are divided into M≥3 levels, and acquiring the percentage p of broadband users in each perceived evaluation level m∈[1,M] m ;

[0136] A sorting unit for sorting the non-labeled perceived indicators with the same ji among all N broadband users according to the numerical values as where n∈[1,N] represents each broadband user, i∈[1,I] represents each dimension, and ji∈[1,ji] represents each perceived indicator of each dimension;

[0137] A quantile unit, connected to the percentage acquisition unit and the sorting unit, for acquiring the percentage p of broadband users in the perceived evaluation level m∈[1,M - 1] m The demarcation position among N broadband users, and acquiring the numerical values of M - 1 non-labeled perceived indicators located at the corresponding demarcation positions after sorting as the characteristic quantiles

[0138] An index scoring unit, connected to the quantile unit, for Assigning scores between 0 points and 100 points to each characteristic quantile, and using the interpolation method between the characteristic quantiles to obtain each Index score between 0 points and 100 points

[0139] In one embodiment, where:

[0140] The perceived evaluations are divided into 4 levels: excellent, good, medium, and poor;

[0141] p m The demarcation position is n = ceil(p m *N), where ceil represents rounding up by 1 digit;

[0142] Less than or equal to / greater than or equal to of The index score is 0 points, equal to of The index score is 60 points, greater than or equal to / less than or equal to of The index score is 100 points.

[0143] In one embodiment, the comprehensive scoring module 3 is specifically used for:

[0144] Obtain the metric weight of each perception metric based on the degree of difference in the metric scores of multiple broadband users for the same perception metric. Obtain the dimension score of each dimension of each broadband user according to each metric score and each metric weight. Obtain the dimension weight of each dimension based on the degree of difference in the dimension scores of the same dimension of multiple broadband users. Obtain the overall score of the perception of each broadband user according to each dimension score and each dimension weight.

[0145] In one embodiment, the comprehensive scoring module 3 specifically includes:

[0146] The dimension scoring unit is used to calculate the information entropy of the perception metrics of each ji∈[1, Ji] among the total N broadband users where k = 1 / ln(N), Calculate the metric weight of the perception metric of each ji∈[1, Ji] where d ji = 1 - e ji , calculate the dimension score of each dimension of each broadband user

[0147] The overall scoring unit is connected to the dimension scoring unit and is used to calculate the information entropy of each dimension of each i∈[1, I] among the total N broadband users where k = 1 / ln(N), Calculate the dimension weight of each dimension of each i∈[1, I] where d i = 1 - e i , calculate the overall score of the perception of each broadband user

[0148] In one embodiment, the device further includes a scoring application module, which is connected to the comprehensive scoring module 3 and specifically includes:

[0149] The fault location unit is used to predict the occurrence of network faults and / or locate the location where network faults occur based on the metric scores, dimension scores, and overall scores of multiple broadband users;

[0150] The service evaluation unit is used to obtain the metric scores, dimension scores, and overall scores of multiple broadband users during the period of using a certain network service, and evaluate the quality of a certain network service according to the metric scores, dimension scores, and overall scores of the corresponding period;

[0151] The marketing unit is used to market network services to a certain broadband user according to the metric scores, dimension scores, and overall scores of a certain broadband user.

[0152] Example 3:

[0153] Example 3 of this application provides a computer-readable storage medium, in which a computer program is stored. When the computer program runs on a processor, it implements the broadband user perception evaluation method described in Example 1, or implements the broadband user perception evaluation device described in Example 2.

[0154] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information such as computer-readable instructions, data structures, computer program units, or other data. The computer-readable storage medium includes, but is not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), flash memory, or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), digital versatile disc (DVD), or other optical disc storage, magnetic cassette, tape, magnetic disk storage, or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer.

[0155] In addition, this application can also provide a computer device, including a memory and a processor. A computer program is stored in the memory. When the processor runs the computer program stored in the memory, the processor executes the broadband user perception evaluation method described in Example 1. This computer device can be the broadband user perception evaluation device described in Example 2.

[0156] Among them, the memory is connected to the processor. The memory can use flash memory, read-only memory, or other memories. The processor can use a central processing unit or a single-chip microcomputer.

[0157] Examples 1-3 of this application provide a broadband user perception evaluation method, device, and medium. They obtain multiple perception indicators from multiple dimensions, obtain characteristic quantiles according to the perception distribution of multiple broadband users, use the characteristic quantiles to obtain the scores of each non-labeled perception indicator, obtain multi-dimensional scores by integrating multiple indicators, and obtain the overall score of the user's overall perception by integrating multiple dimensions. Three-layer scoring results of the overall score, dimension score, and indicator score can be obtained, which more comprehensively evaluates the perception provided by the broadband to the user. Based on the indicator scores obtained from the user perception, the broadband user perception evaluation results can reflect the true experience of the user using the broadband.

[0158] It is understood that the above embodiments are merely exemplary embodiments adopted to illustrate the principle of the present application. However, the present application is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present application, and these modifications and improvements are also regarded as the protection scope of the present application.

Claims

1. A broadband user perception evaluation method, characterized in that, The method includes: Obtaining multiple perception metrics for each dimension of multiple broadband users from multiple dimensions, where the perception metrics include non-label perception metrics; Obtaining the metric scores of each perception metric for each broadband user, including obtaining the metric scores of each non-label perception metric for each broadband user according to multiple characteristic quantiles of the perception distribution of each non-label perception metric by multiple broadband users; Obtaining the dimension scores of each dimension of each broadband user according to the metric scores, and obtaining the overall perception score of each broadband user according to the dimension scores.

2. The method according to claim 1, wherein Obtaining multiple perception metrics for each dimension of multiple broadband users from multiple dimensions specifically includes: Obtaining all broadband users of the operator within a preset range; Obtaining multiple perception metrics for each dimension of each broadband user from multiple dimensions at the preset frequency of each perception metric within a preset duration; Among them, the multiple dimensions include the user side, the network side, the application side, and the service side, the perception metrics also include label perception metrics, the label representation of the label perception metrics is yes or no, and the corresponding metric scores are 100 points or 0 points.

3. The method according to claim 2, wherein Obtaining multiple perception metrics for each dimension of each broadband user from multiple dimensions specifically includes: Obtaining the business B domain data, operation O domain data, complaint data, order dispatch data, and alarm data of the operator within a preset range, where the O domain data includes SA data collected by setting a service assurance SA board on the broadband access server BRAS or broadband gateway BNG in the metropolitan area network; Obtaining multiple package adaptation parameters for each broadband user according to the B domain data, and judging whether the mobile hotspot WiFi performance, the first part performance of the optical modem, and the router performance on the user side of each broadband user meet the corresponding package adaptation parameters to obtain multiple label perception metrics on the user side; Obtaining the measured values of the second part performance of the optical modem and the IPTV performance of the network television on the user side of each broadband user, and the measured values of the optical line terminal OLT device performance, switch performance, and broadband access server BAS performance on the network side according to the O domain data to obtain multiple non-label perception metrics on the user side and the network side; Obtaining the application metrics and poor quality application metrics of each broadband user according to the SA data to obtain multiple non-label perception metrics on the application side, and obtaining the metrics representing the services obtained by each broadband user according to the complaint data, order dispatch data, and alarm data to obtain multiple non-label perception metrics on the service side.

4. The method according to any one of claims 1 to 3, characterized in that, Obtaining the metric scores of each non-label perception metric for each broadband user according to multiple characteristic quantiles of the perception distribution of each non-label perception metric by multiple broadband users specifically includes: Obtain the perceived evaluations of N' broadband users on their own broadband usage. The perceived evaluations are divided into M≥3 levels, and obtain the percentage p of broadband users in each perceived evaluation level m∈[1,M] m ; Sort the same non-label perception metrics of all N broadband users according to the numerical values as where n ∈ [1, N] represents each broadband user, i ∈ [1, I] represents each dimension, and ji ∈ [1, ji] represents each perception metric of each dimension; Obtain the percentage p of broadband users with a perceived evaluation level m ∈ [1, M - 1] m At the demarcation position among N broadband users, obtain the values of M - 1 non-labeled perceived indicators located at the corresponding demarcation position after sorting Is the characteristic quantile For each feature quantile Assign scores between 0 and 100, and use interpolation between feature quantiles to obtain each Index score between 0 and 100 5. The method according to claim 4, wherein Wherein: The perception evaluation is divided into 4 levels: excellent, good, medium, and poor; p m The demarcation position is n = ceil(p m *N), where ceil means rounding up by 1 bit; Less than or equal to / greater than or equal to of The index score of is 0 points, equal to of The index score of is 60 points, greater than or equal to / less than or equal to of The index score of is 100 points.

6. The method according to claim 4, wherein Obtaining the dimension scores of each dimension of each broadband user according to the metric scores, and obtaining the overall perception score of each broadband user according to the dimension scores specifically includes: Obtain the index weight of each perception index according to the difference degree of the index scores of multiple broadband users for the same perception index. Obtain the dimension score of each dimension of each broadband user according to each index score and each index weight. Obtain the dimension weight of each dimension according to the difference degree of the dimension scores of the same dimension of multiple broadband users. Obtain the overall score of the perception of each broadband user according to each dimension score and each dimension weight.

7. The method according to claim 6, characterized in that, Obtain the index weight of each perception index according to the difference degree of the index scores of multiple broadband users for the same perception index. Obtain the dimension score of each dimension of each broadband user according to each index score and each index weight. Obtain the dimension weight of each dimension according to the difference degree of the dimension scores of the same dimension of multiple broadband users. Obtain the overall score of the perception of each broadband user according to each dimension score and each dimension weight, specifically including: Calculate the information entropy of the perception index of each \(j_i\in[1,J_i]\) among all \(N\) broadband users where \(k = 1 / \ln(N)\) Calculate the index weight of the perception index of each \(j_i\in[1,J_i]\) where \(d\) ji \(= 1 - e\) ji and calculate the dimension score of each dimension of each broadband user Calculate the information entropy of each dimension for \(i\in[1, I]\) among all \(N\) broadband users where \(k = 1 / \ln(N)\) Calculate the dimension weight of each dimension for \(i\in[1, I]\) where \(d\) i \(= 1 - e\) i Calculate the overall perceived score of each broadband user 8. The method according to claim 6, wherein The method further includes: Predict the occurrence of network faults and / or locate the location where network faults occur according to the index scores, dimension scores, and overall scores of multiple broadband users. Obtain the index scores, dimension scores, and overall scores of multiple broadband users during the period of using a certain network service, and evaluate the quality of a certain network service according to the index scores, dimension scores, and overall scores of the corresponding period. Market network services to a certain broadband user according to the index score, dimension score, and overall score of the certain broadband user.

9. A broadband user perception evaluation device, characterized in that The device includes: An acquisition module, configured to acquire multiple perception indexes of each dimension of multiple broadband users from multiple dimensions, and the perception indexes include non-label perception indexes; An index scoring module, connected to the acquisition module, configured to acquire the index score of each perception index of each broadband user, including obtaining the index score of each non-label perception index of each broadband user according to multiple characteristic quantiles of the perception distribution of multiple broadband users for each non-label perception index; A comprehensive scoring module, connected to the index scoring module, configured to obtain the dimension score of each dimension of each broadband user according to the index score, and obtain the overall score of the perception of each broadband user according to the dimension score.

10. A computer-readable storage medium, characterized in that, A computer program is stored in the computer-readable storage medium, and when the computer program is run by a processor, the broadband user perception evaluation method according to any one of claims 1-8 is implemented.

Images