Method, device, equipment and medium for perceiving network quality of home broadband users

Abstract

The present invention discloses a method, apparatus, device, and medium for network quality perception of home broadband users. The method comprises: constructing an LDA topic model based on historical complaints from complaining users to locate the complaints about their home broadband networks; analyzing user service data based on network issues and customer experience management theory to extract key performance indicator data for the home broadband network and construct a home broadband network quality indicator system; obtaining key performance indicator data for the home broadband network quality indicator system from user samples; inputting the key performance indicator data into an accessibility neural network, an integrity neural network, and a retention neural network to obtain accessibility scores, integrity scores, and retention scores; and performing a weighted summation of the aforementioned scores to obtain the home broadband user's network quality perception. The present invention can accurately locate user pain points and calculate home broadband user perception from a data perspective, thereby improving accuracy.

Description

Technical Field

[0001] The present invention relates to the field of communication networks, and in particular to a method, device, equipment and medium for perceiving network quality of home broadband users. Background Art

[0002] Home broadband services are a key competitive advantage for operators. Improving user satisfaction with these services has been a key focus for these operators in recent years. When customers can no longer tolerate poor network performance, they often file complaints. Customer Experience Management (CEM) addresses this need.

[0003] The basic theory of the CEM indicator system is to analyze user usage frequency, network, and service quality across various services, weighting and aggregating KPI performance indicators to construct a CEM indicator system that closely reflects user perception. This allows for the timely identification and location of issues that impact user experience. Currently, the Delphi and analytic hierarchy process methods are primarily used to determine the weights for network quality evaluation systems. However, the determination of indicator weights still relies heavily on the experience of individual technical personnel.

[0004] The Delphi method mainly involves the company forming a special forecasting organization, which includes several experts and corporate forecast organizers. According to the prescribed procedures, the experts' opinions or judgments on the future market are consulted back-to-back, and then a forecast is made. However, in the consultation process, the opinions of authoritative people will affect the opinions of others, and the process is relatively complicated and time-consuming. The hierarchical analysis method refers to a decision-making method that decomposes elements that are always related to decision-making into levels such as goals, criteria, and plans, and conducts qualitative and quantitative analysis on this basis. However, it cannot provide new plans for decision-making, and there are fewer quantitative data and more qualitative components, which is not easy to be convincing. Summary of the Invention

[0005] The present invention provides a method, device, equipment and medium for network quality perception of home broadband users. The method mines network pain points based on the LDA model, constructs a neural network to simulate and investigate users' perception of different network issues based on performance indicators, and uses an objective weighting method to assign weights. The method calculates home broadband user perception from a data perspective, improves accuracy, and utilizes accurate user perception to effectively improve network service quality and user satisfaction, and conducts more practical evaluation and optimization of network service content.

[0006] To achieve the above objectives, an embodiment of the present invention provides a method for perceiving network quality of a home broadband user, comprising:

[0007] An LDA topic model is constructed based on the historical complaints of complaining users to identify the issues they complained about regarding their home broadband network.

[0008] Classify the complaint issues as network issues of accessibility, integrity and retention;

[0009] Analyze user service data based on the aforementioned network issues and customer experience management theory, extract key performance indicator data of home broadband networks, and build a home broadband network quality indicator system;

[0010] Obtain key performance indicator data in the home broadband network quality indicator system for the surveyed user sample;

[0011] Inputting the key performance indicator data into the access neural network, integrity neural network, and retention neural network according to business attributes to obtain access score, integrity score, and retention score;

[0012] A weighted sum is taken of the accessibility score, the integrity score, and the retention score to obtain the network quality perception of the home broadband user.

[0013] The LDA topic model is constructed based on the historical complaint content of the complaining user to locate the complaint issues of the complaining user regarding the home broadband network, specifically including:

[0014] Obtain the historical complaint content of the complaining user;

[0015] Segment the complaint content d of each complaining user and filter out meaningless words to obtain the corpus set W = {w1, w2, ..., w n};

[0016] Count the words after the above segmentation and get p(w i |d), which represents the frequency of the i-th word appearing in the d-th complaint content;

[0017] For each w in the corpus set W i Randomly assign a subject as the initial subject;

[0018] Resample each w by Gibbs sampling method i The topic to which is assigned is updated in the corpus until Gibbs sampling converges.

[0019] Furthermore, performing a weighted summation of the accessibility score, the integrity score, and the retention score to obtain the network quality perception of the home broadband user specifically includes:

[0020] The weight of each network problem score is calculated using an objective weighting method, specifically:

[0021]

[0022] The comparative strength of network problem scores is calculated using the above formula, where: represents the mean value of the j-th network question score, Sj represents the standard deviation of the j-th network question score;

[0023]

[0024] The conflict of network problem scores is calculated using the above formula; where r ij represents the correlation between the score of the i-th network problem and the score of the j-th network problem;

[0025] C j =S j *R j

[0026] The above formula is used to calculate the information content of network problem scores; where C j represents the information content of the j-th network question score;

[0027]

[0028] The weight of the network problem score is calculated using the above formula; where w j represents the weight of the j-th network question score;

[0029] CEM=w i *cem i

[0030] The above formula is used to calculate the user's home broadband network quality perception, where cem i ∈(accessibility score, integrity score, retention score).

[0031] Accordingly, an embodiment of the present invention further provides a device for perceiving network quality of a home broadband user, comprising:

[0032] The complaint content topic model module is used to build an LDA topic model based on the historical complaint content of the complaining user and locate the complaint issues of the complaining user regarding the home broadband network;

[0033] A network quality indicator system acquisition module is used to classify the complaint issues into network issues related to accessibility, integrity, and retention, analyze user service data based on the network issues and customer experience management theory, extract key performance indicator data of home broadband networks, and build a home broadband network quality indicator system;

[0034] A data collection module is used to obtain key performance indicator data in the home broadband network quality indicator system from the surveyed user sample;

[0035] A network problem scoring module is used to input the key performance indicator data into the accessibility neural network, integrity neural network, and retention neural network according to business attributes to obtain accessibility scores, integrity scores, and retention scores;

[0036] The network quality calculation module is used to perform weighted summation on the accessibility score, the integrity score and the retention score to obtain the network quality perception of the home broadband user.

[0037] Furthermore, the complaint content topic model module specifically includes:

[0038] Obtain the historical complaint content of the complaining user;

[0039] Segment the complaint content d of each complaining user and filter out meaningless words to obtain the corpus set W = {w1, w2, ..., w n};

[0040] Count the words after the above segmentation and get p(w i |d), which represents the frequency of the i-th word appearing in the d-th complaint content;

[0041] For each w in the corpus set W i Randomly assign a subject as the initial subject;

[0042] Resample each w by Gibbs sampling method i The topic to which is assigned is updated in the corpus until Gibbs sampling converges.

[0043] Furthermore, the network quality calculation module specifically includes:

[0044] The weighted sum of the accessibility score, integrity score, and retention score is performed to obtain the network quality perception of the home broadband user, specifically including:

[0045] The weight of each network problem score is calculated using an objective weighting method, specifically:

[0046]

[0047] The comparative strength of network problem scores is calculated using the above formula, where: represents the mean value of the j-th network question score, S j represents the standard deviation of the j-th network question score;

[0048]

[0049] The conflict of network problem scores is calculated using the above formula; where r ij represents the correlation between the score of the i-th network problem and the score of the j-th network problem;

[0050] C j =S j *Rj

[0051] The above formula is used to calculate the information content of network problem scores; where C j represents the information content of the j-th network question score;

[0052]

[0053] The weight of the network problem score is calculated using the above formula; where w j represents the weight of the j-th network question score;

[0054] CEM=w i *cem i

[0055] The above formula is used to calculate the user's home broadband network quality perception, where cem i ∈(accessibility score, integrity score, retention score).

[0056] Correspondingly, an embodiment of the present invention also provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for perceiving network quality of home broadband users when executing the program.

[0057] Accordingly, an embodiment of the present invention further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for perceiving network quality of home broadband users.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] The present invention uses the LDA topic model to mine the pain points of home broadband networks of complaining users, conducts text mining on the complaint content of complaining users, accurately locates the pain points of home broadband users, and thus determines network problems. Based on network problems and combined with customer experience management theory, a home broadband network quality index system is constructed; then a BP neural network is constructed for the three major network problems, and the user's perception of different network problems is simulated from the performance indicators of home broadband satisfaction survey; finally, the weights of the three major network problem scores are calculated using an objective weighting method, and then the network quality perception of home broadband users is calculated. The network problem perception of home broadband users is simulated through a deep learning self-learning method, and weighted using an objective weighting method. The home broadband user perception is calculated from a data perspective to improve accuracy, so as to effectively improve network service quality and user satisfaction by using accurate network quality perception, increase user stickiness, and further guide precise marketing and customer service, thereby facilitating better and more practical evaluation of poor quality of home broadband users and more accurate early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a flow chart of a method for perceiving network quality of a home broadband user provided by an embodiment of the present invention;

[0061] Figure 2 This is a diagram of the architecture of the home broadband network quality indicator body constructed in an embodiment of the present invention;

[0062] Figure 3 It is a structural diagram of a network quality perception device for home broadband users provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0064] See also Figure 1 , is a flow chart of a method for perceiving network quality of a home broadband user provided by an embodiment of the present invention.

[0065] The method for perceiving network quality of a home broadband user provided by an embodiment of the present invention includes steps S1 to S6, which are specifically as follows:

[0066] S1: Build an LDA topic model based on the historical complaint content of the complaining user to locate the complaint issues of the complaining user regarding the home broadband network;

[0067] S2, classifying the complaint problem as a network problem of accessibility, integrity, and retention;

[0068] S3, analyzing user service data based on the network issues and customer experience management theory, extracting key performance indicator data of home broadband networks, and building a home broadband network quality indicator system;

[0069] S4, obtaining key performance indicator data in the home broadband network quality indicator system of the surveyed user sample;

[0070] S5, inputting the key performance indicator data into the accessibility neural network, the integrity neural network, and the retention neural network according to the business attributes to obtain the accessibility score, the integrity score, and the retention score;

[0071] S6. Perform weighted summation on the accessibility score, the integrity score, and the retention score to obtain the network quality perception of the home broadband user.

[0072] In step S1, the LDA topic model is constructed based on the historical complaint content of the complaining user to locate the complaint issues of the complaining user regarding the home broadband network, specifically including:

[0073] Obtain the historical complaint content of the complaining user;

[0074] Segment the complaint content d of each complaining user and filter out meaningless words to obtain the corpus set W = {w1, w2, ..., w n};

[0075] Count the words after the above segmentation and get p(w i |d), which represents the frequency of the i-th word appearing in the d-th complaint content;

[0076] For each w in the corpus set W i Randomly assign a subject as the initial subject;

[0077] Resample each w by Gibbs sampling method i The topic to which is assigned is updated in the corpus until Gibbs sampling converges.

[0078] In a specific embodiment, based on the results of the topic model, it was found that users' dissatisfaction with home broadband networks mainly focused on problems such as "broadband disconnection", "webpage opening slow", "webpage cannot open", "TV freeze" and "unable to dial". Based on customer experience management theory, the complaint details can be classified into three categories of network problems: "accessibility", "integrity" and "retention". Specifically, "unable to dial" is classified as an accessibility problem, "webpage opening slow", "webpage cannot open" and "TV freeze" are classified as integrity problems, and "broadband disconnection" is classified as a retention problem.

[0079] It should be noted that customer experience management, or CEM, refers to the collection and analysis of customer feedback data by enterprises through all channels and all touchpoints, from which problems and opportunities in products or services are discovered, and targeted improvements and optimizations are made to enhance customer experience, improve customer satisfaction and loyalty, and achieve continuous user growth.

[0080] In specific embodiments, see Figure 2 , is a diagram of the home broadband network quality indicator structure constructed in an embodiment of the present invention.

[0081] Based on the CEM system, by analyzing the frequency of users' use of various services, network and service quality, and weighted aggregation of underlying data, combined with the entire process of home broadband services, key performance indicator data of home broadband networks is extracted to build a home broadband network quality indicator system.

[0082] In step S4, specifically, from multiple data sources such as DPI, IPTV, IHGU, Radius, QCA, etc., the fields of multiple data sources are matched in the shared layer, home broadband end-to-end and other systems according to the user numbers of the surveyed user samples, so as to obtain the key performance indicator data in the home broadband network quality indicator system.

[0083] Based on the constructed home broadband network quality indicator system, after collecting the corresponding key performance indicator data, a BP neural network was constructed for the three major network problems to simulate the user's network quality perception in different sub-processes to characterize the user's perception.

[0084] The BP neural network used in the present invention is a multi-layer feedforward network trained by error back propagation. Its basic idea is to use the gradient descent method to update the weight w and bias b. Its calculation process is divided into a forward calculation process and a reverse calculation process.

[0085] During the forward propagation process, the input pattern is processed layer by layer from the input layer to the hidden unit layer and then transferred to the output layer. The state of each layer of neurons only affects the state of the neurons in the next layer.

[0086] If the expected output cannot be obtained in the output layer, back propagation is performed to return the error signal along the original connection path, and the error signal is minimized by modifying the weights of each neuron.

[0087] Business experts divide the key performance indicator data into accessibility, integrity, and retention according to business attributes; and then construct accessibility neural network, integrity neural network, and retention neural network respectively.

[0088] First, the three neural networks are trained, where the weight parameters are adjusted through learning so that they can predict the label information of the input sample data, thereby obtaining the key performance index data x i to y i The mapping relationship;

[0089] y i =Wx i +b

[0090] Among them, x i They represent accessibility key performance indicator data, integrity key performance indicator data, and retention key performance indicator data, respectively; w represents the weight matrix of the input layer, hidden layer, or output layer; b represents the bias matrix; y i They represent accessibility score, integrity score and retention score respectively.

[0091] Furthermore, a weighted sum is performed on the accessibility score, the integrity score, and the retention score to obtain the network quality perception of the home broadband user.

[0092] The weight of each network problem score is calculated using the objective weighting method, specifically:

[0093]

[0094] The comparative strength of network problem scores is calculated using the above formula, where: represents the mean value of the j-th network question score, S j represents the standard deviation of the j-th network question score;

[0095]

[0096] The conflict of network problem scores is calculated using the above formula; where r ij represents the correlation between the score of the i-th network problem and the score of the j-th network problem;

[0097] C j =S j *R j

[0098] The above formula is used to calculate the information content of network problem scores; where C j represents the information content of the j-th network question score;

[0099]

[0100] The weight of the network problem score is calculated using the above formula; where w j represents the weight of the j-th network question score;

[0101] It should be noted that the weighting method described above is an objective weighting method based on data volatility. Its concept relies on two indicators: contrast intensity (volatility) and conflict (correlation). Contrast intensity is represented by standard deviation; a larger standard deviation indicates greater volatility and thus a higher weight. Conflict is represented by the correlation coefficient; a larger correlation coefficient indicates less conflict and thus a lower weight. When calculating the weight, contrast intensity and conflict are multiplied and normalized to obtain the final weight.

[0102] CEM=w i *cem i

[0103] Finally, the above formula is used to calculate the user's home broadband network quality perception, where cem i ∈(accessibility score, integrity score, retention score).

[0104] In summary, the network quality perception method for home broadband users provided by the embodiment of the present invention uses the LDA topic model to parse the complaint content, combines customer experience management theory with the entire process of home broadband services to construct a network quality index system; uses the BP neural network to simulate and investigate users' network quality perception on different network issues, and uses the CRITIC weighting method to calculate the weight of the network problem score, thereby avoiding manual intervention and improving the accuracy of home broadband users' network perception. Among them, by performing text mining on the complaint content of the complaining users, the pain points of home broadband users are accurately located; the network problem perception of home broadband users is simulated through the self-learning method of deep learning, and the objective weighting method is used to assign weights, and the perception of home broadband users is calculated from a data perspective to improve accuracy; accurate network quality perception is used to effectively improve network service quality and user satisfaction, increase user stickiness, further guide precise marketing and customer service, and conduct more practical evaluation and optimization of network service business content.

[0105] See also Figure 3 , is a structural diagram of a network quality perception device for home broadband users provided by an embodiment of the present invention.

[0106] The home broadband user network quality perception device includes:

[0107] Complaint content topic model module 1 is used to build an LDA topic model based on the historical complaint content of the complaining user and locate the complaint issues of the complaining user regarding the home broadband network;

[0108] A network quality indicator system acquisition module 2 is configured to classify the complaint issues into network issues related to accessibility, integrity, and retention, analyze user service data based on the network issues and customer experience management theory, extract key performance indicator data of home broadband networks, and construct a home broadband network quality indicator system;

[0109] Data collection module 3, used to obtain key performance indicator data in the home broadband network quality indicator system of the surveyed user sample;

[0110] The network problem scoring module 4 is used to input the key performance indicator data into the accessibility neural network, the integrity neural network, and the retention neural network according to the service attributes to obtain the accessibility score, the integrity score, and the retention score;

[0111] The network quality calculation module 5 is configured to perform a weighted summation of the accessibility score, the integrity score, and the retention score to obtain the network quality perception of the home broadband user.

[0112] Furthermore, the complaint content topic model module specifically includes:

[0113] Obtain the historical complaint content of the complaining user;

[0114] Segment the complaint content d of each complaining user and filter out meaningless words to obtain the corpus set W = {w1, w2, ..., w n};

[0115] Count the words after the above segmentation and get p(w i |d), which represents the frequency of the i-th word appearing in the d-th complaint content;

[0116] For each w in the corpus set W j Randomly assign a subject as the initial subject;

[0117] Resample each w by Gibbs sampling method i The topic to which is assigned is updated in the corpus until Gibbs sampling converges.

[0118] Furthermore, the network quality calculation module specifically includes:

[0119] The weighted sum of the accessibility score, integrity score, and retention score is performed to obtain the network quality perception of the home broadband user, specifically including:

[0120] The weight of each network problem score is calculated using an objective weighting method, specifically:

[0121]

[0122] The comparative strength of network problem scores is calculated using the above formula, where: represents the mean value of the j-th network question score, S j represents the standard deviation of the j-th network question score;

[0123]

[0124] The conflict of network problem scores is calculated using the above formula; where r ij represents the correlation between the score of the i-th network problem and the score of the j-th network problem;

[0125] C j =S j *R j

[0126] The above formula is used to calculate the information content of network problem scores; where C j represents the information content of the j-th network question score;

[0127]

[0128] The weight of the network problem score is calculated using the above formula; where w j represents the weight of the j-th network question score;

[0129] CEM=w i *cem i

[0130] The above formula is used to calculate the user's home broadband network quality perception, where cem i ∈(accessibility score, integrity score, retention score).

[0131] It should be noted that the functions and technical effects achieved by each module and unit in the device are respectively the same as the functions and technical effects achieved by the method for network quality perception of home broadband users provided in the above embodiment, and will not be repeated here.

[0132] An embodiment of the present invention also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for network quality perception of home broadband users described in any of the above embodiments is implemented.

[0133] An embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program; wherein, when the computer program is running, it controls the device where the computer-readable storage medium is located to execute the method for network quality perception of home broadband users described in any of the above embodiments.

[0134] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus the necessary hardware platform, and of course, it can also be implemented entirely by hardware. Based on this understanding, all or part of the contribution of the technical solution of the present invention to the background art can be embodied in the form of a software product. This computer software product can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments of the present invention or certain parts of the embodiments.

[0135] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for perceiving network quality of home broadband users, characterized in that: include: Build an LDA topic model based on the historical complaints of complaining users to identify the issues they complained about their home broadband networks. Classify the complaint issues as network issues of accessibility, integrity and retention; Analyze user service data based on the aforementioned network issues and customer experience management theory, extract key performance indicator data of home broadband networks, and build a home broadband network quality indicator system; Obtain key performance indicator data in the home broadband network quality indicator system for the surveyed user sample; Inputting the key performance indicator data into the access neural network, integrity neural network, and retention neural network according to business attributes to obtain access score, integrity score, and retention score; performing a weighted summation of the accessibility score, the integrity score, and the retention score to obtain a network quality perception of the home broadband user; The weighted summation of the accessibility score, the integrity score, and the retention score to obtain the network quality perception of the home broadband user specifically includes: The weight of each network problem score is calculated using an objective weighting method, specifically: The comparative strength of network problem scores is calculated using the above formula, where: represents the mean value of the j-th network question score, S j represents the standard deviation of the j-th network question score; The conflict of network problem scores is calculated using the above formula; where r kj represents the correlation between the kth network problem score and the jth network problem score; The above formula is used to calculate the information content of network problem scores; where C j represents the information content of the j-th network question score; The weight of the network problem score is calculated using the above formula; where w j represents the weight of the j-th network question score; The above formula is used to calculate the user's home broadband network quality perception, where cem j Indicates the j-th network problem score, cem j ∈(accessibility score, integrity score, retention score).

2. The method for perceiving network quality of home broadband users according to claim 1, wherein: The LDA topic model is constructed based on the historical complaint content of the complaining user to locate the complaint issues of the complaining user regarding the home broadband network, specifically including: Obtain the historical complaint content of the complaining user; Segment the complaint content d of each complaining user and filter out meaningless words to obtain the corpus set W = {w1,w2,...,w n }; Count the words after the above segmentation and get p(w i |d), which represents the frequency of the i-th word appearing in the d-th complaint content; For each w in the corpus set W i Randomly assign a subject as the initial subject; Resample each w by Gibbs sampling method i The topic to which is assigned is updated in the corpus until Gibbs sampling converges.

3. A network quality perception device for home broadband users, characterized in that: The device comprises: The complaint content topic model module is used to build an LDA topic model based on the historical complaint content of the complaining user and locate the complaint issues of the complaining user regarding the home broadband network; A network quality indicator system acquisition module is used to classify the complaint issues into network issues related to accessibility, integrity, and retention, analyze user service data based on the network issues and customer experience management theory, extract key performance indicator data of home broadband networks, and build a home broadband network quality indicator system; A data collection module is used to obtain key performance indicator data in the home broadband network quality indicator system from the surveyed user sample; A network problem scoring module is used to input the key performance indicator data into the accessibility neural network, integrity neural network, and retention neural network according to business attributes to obtain accessibility scores, integrity scores, and retention scores; a network quality calculation module, configured to perform a weighted summation of the accessibility score, the integrity score, and the retention score to obtain a network quality perception of a home broadband user; The network quality calculation module specifically includes: The weight of each network problem score is calculated using an objective weighting method, specifically: The comparative strength of network problem scores is calculated using the above formula, where: represents the mean value of the j-th network question score, S j represents the standard deviation of the j-th network question score; The conflict of network problem scores is calculated using the above formula; where r kj represents the correlation between the kth network problem score and the jth network problem score; C j =S j *R j The above formula is used to calculate the information content of network problem scores; where C j represents the information content of the j-th network question score; The weight of the network problem score is calculated using the above formula; where w j represents the weight of the j-th network question score; The above formula is used to calculate the user's home broadband network quality perception, where cem j Indicates the j-th network problem score, cem j ∈(accessibility score, integrity score, retention score).

4. A network quality perception device for home broadband users according to claim 3, characterized in that: The complaint content topic model module specifically includes: Obtain the historical complaint content of the complaining user; Segment the complaint content d of each complaining user and filter out meaningless words to obtain the corpus set W = {w1,w2,...,w n }; Count the words after the above segmentation and get p(w i |d), which represents the frequency of the i-th word appearing in the d-th complaint content; For each w in the corpus set W i Randomly assign a subject as the initial subject; Resample each w by Gibbs sampling method i The topic to which is assigned is updated in the corpus until Gibbs sampling converges.

5. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for network quality perception of a home broadband user as claimed in any one of claims 1 to 2 is implemented.

6. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a method for perceiving network quality of a home broadband user as claimed in any one of claims 1 to 2.

Images