Digital television testing system

Abstract

The invention discloses a digital television test system, relates to the technical field of communication and electronic information, and is used for solving the problems of rough user classification and lack of dynamic adjustment capability based on user feedback in the existing test system. Personalization and dynamic optimization of digital television test configuration are achieved, firstly, digital television operation data and user operation behavior data are collected, a segmented weighting method is adopted to initialize test parameters, and users are divided into a high-frequency type, an intermediate-frequency type and a low-frequency type through fuzzy clustering. For intermediate frequency users, a convolutional neural network model is constructed to analyze behavior characteristics of the intermediate frequency users, whether a high-precision test scheme is needed or not is judged, and a feedback updating module dynamically adjusts user classification according to comparison between a satisfaction score and a deviation interval, so that the test matching degree and the resource utilization efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of communication and electronic information technology, and more particularly to a digital television testing system. Background Art

[0002] With the rapid development of digital television (DTV) technology, users' demands for program quality, responsiveness, and intelligent service experience are increasing. Traditional DTV test systems typically rely on static test cases, standardized indicators, and fixed parameter configurations to test digital signal transmission performance, audio and video quality, interface compatibility, and other aspects. However, in actual applications, different users' operational behaviors vary significantly, such as viewing frequency, channel switching behavior, and feature usage preferences, all of which have varying degrees of impact on system performance and testing requirements.

[0003] The existing technology has the following deficiencies: At present, the test parameter initialization is static and lacks adaptability to different user scenarios. In addition, the test plan design has a weak correlation with user behavior and cannot effectively identify users who require high-precision testing. At the same time, there is a lack of a dynamic classification mechanism based on user feedback, and the test plan cannot be adjusted in time, resulting in unreasonable allocation of test resources and low system test efficiency. It is difficult to meet the needs of the intelligent development of digital TV. Therefore, a digital TV test system is proposed.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a digital television test system, which solves the problems raised in the above-mentioned background technology by applying a segmented weighted initialization algorithm, a fuzzy clustering user classification method, a test scheme identification mechanism based on a convolutional neural network, and a feedback-driven dynamic update strategy.

[0006] To achieve the above object, the present invention provides the following technical solution: a digital television test system, comprising a data collection module, a user classification module, a program analysis module and a feedback update module, wherein the modules are signal-connected; The data collection module collects digital TV operation data and user operation behavior data; The user classification module initializes the test parameters using a segmented weighted method based on the digital TV operation data, and uses fuzzy clustering to analyze the user usage frequency based on the user operation behavior data, dividing the users into high-frequency users, medium-frequency users, and low-frequency users, and matching the classified users with the initialized test parameters; If the solution analysis module identifies the current user as a medium-frequency user, it collects the current user's operation behavior data and builds a convolutional neural network model based on the operation behavior data to analyze whether the current user needs a high-precision test solution; The feedback update module collects user feedback data, calculates the deviation interval based on the user feedback data, and updates the current user classification by comparing the current user feedback data with the deviation interval.

[0007] In a preferred embodiment, CPU occupancy rate, startup time and response time are collected as digital TV operation data; The number of channel switching, daily cumulative viewing time, and number of function button clicks are collected as user operation behavior data.

[0008] In a preferred embodiment, the digital television operation data is normalized and weight coefficients are set for different parameters to obtain initialization test parameters; The user operation behavior data is used as input to classify users using fuzzy clustering method to obtain the category to which the user belongs; Match the user's category with the corresponding initialization test parameters to obtain a personalized test configuration plan.

[0009] In a preferred embodiment, medium frequency users are screened out based on the categories to which they belong; The user operation behavior data of the medium frequency user within the preset monitoring time interval starting from the current moment is called, and is sorted in chronological order to obtain a behavior feature matrix.

[0010] In a preferred embodiment, the behavior feature matrix is normalized to obtain a standardized feature matrix; a convolutional neural network model is constructed using the standardized feature matrix as input to output the matching probability between the current user behavior and the high-precision test configuration.

[0011] In a preferred embodiment, a precision determination threshold is set and compared with the matching probability of the current user behavior and the high-precision test configuration; if the matching probability of the current user behavior and the high-precision test configuration is greater than or equal to the precision determination threshold, the high-precision test scheme is enabled; If the probability of matching the current user behavior with the high-precision test configuration is less than the accuracy judgment threshold, the medium-precision test solution configuration remains unchanged.

[0012] In a preferred embodiment, user feedback data is collected, including the user's rating value for the current test solution; the user feedback data is archived according to the user identifier to form a feedback data record set.

[0013] In a preferred embodiment, statistical analysis is performed on the feedback data record sets of high-frequency users, medium-frequency users, and low-frequency users, respectively, and the mean and standard deviation of each type of feedback data record set are calculated.

[0014] In a preferred embodiment, the feedback deviation interval of each type of user is defined according to the mean and standard deviation of each type of feedback data record set.

[0015] In a preferred embodiment, the current user feedback score is compared with the feedback deviation interval of each category; if the current user feedback score falls within the feedback deviation interval of a certain category, the current user category is updated to that category; If the current user feedback rating value falls within the feedback deviation range of the current category, the original category will remain unchanged; if the current user feedback rating value does not fall within the feedback deviation range of any category, it will be recorded as abnormal feedback and the category will not be updated for the time being.

[0016] Technical effects and advantages of the present invention: 1. The present invention includes a data collection module, a user classification module, a solution analysis module and a feedback update module, aiming to realize the personalization and dynamic optimization of the test configuration. First, the data collection module collects the digital TV operation data and the user operation behavior data, and the user classification module uses the segmented weighted method to initialize the test parameters according to the digital TV operation data. Then, the user usage frequency is analyzed by fuzzy clustering, and the users are divided into three categories: high frequency, medium frequency and low frequency. For medium frequency users, the solution analysis module further constructs a convolutional neural network model, analyzes their behavioral characteristics, and determines whether a high-precision test solution is needed. Finally, the feedback update module collects user satisfaction scores, conducts comparative analysis based on the deviation interval with the feedback data of various users, and dynamically adjusts the user classification results to realize the refinement and intelligent adjustment of the test configuration, thereby improving the personalized matching degree and resource utilization efficiency of the test solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The present invention is a flow chart of an implementation of a digital television test system.

[0018] Figure 2 The figure is a schematic diagram of the steps of a digital television testing system of the present invention. DETAILED DESCRIPTION

[0019] 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. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] Example 1, please refer to Figures 1 to 2 A digital television test system includes a data collection module, a user classification module, a program analysis module, and a feedback update module. The modules are connected by signals and have the following functions: The data collection module collects digital TV operation data and user operation behavior data.

[0021] The user classification module initializes test parameters using a segmented weighted approach based on digital TV operating data. It then uses fuzzy clustering to analyze user usage frequency based on user behavior data, categorizing users into high-frequency, medium-frequency, and low-frequency users. The classified users are then matched with the initialized test parameters.

[0022] If the solution analysis module identifies the current user as a medium-frequency user, it collects the current user's operation behavior data and builds a convolutional neural network model based on the operation behavior data to analyze whether the current user needs a high-precision test solution.

[0023] The feedback update module collects user feedback data, calculates the deviation interval based on the user feedback data, and updates the current user classification by comparing the current user feedback data with the deviation interval.

[0024] The specific implementation is as follows: In the data collection module, digital TV operating data includes CPU usage, startup time, and response time. Through the system-level performance monitoring interface, the central processing unit (CPU usage) used by the digital TV during operation is collected in real time as a percentage. This parameter reflects system load and is collected in % units. Boot time is collected using a standard time measurement mechanism, recording in seconds the time it takes for the digital TV to receive the power-on command and for the user interface to fully load, used to assess startup efficiency. Response time is collected in milliseconds, recording the time it takes for the digital TV to receive a user command and for the system to respond, used to measure the immediacy of interaction.

[0025] User operation behavior data includes the number of channel switches, daily cumulative viewing time, and the number of function button clicks. The number of channel switches is calculated by analyzing remote control input logs, and the data unit is times. The daily cumulative viewing time is the total time users actually watch, recorded in minutes by the front-end playback log. The number of function button clicks is collected through the interface interaction event monitoring interface, and the data unit is times.

[0026] After the data collection module completes the collection of the above parameters, it passes the digital TV operation data and user operation behavior data as input parameters to the user classification module for subsequent test parameter initialization and user behavior analysis.

[0027] It should be noted that the system-level performance monitoring interface is a software interface integrated into the digital TV operating system, used to collect, monitor and record the usage of hardware resources in real time; the remote control input log is a time series log in the digital TV operating system used to record the operating instructions issued by the user through the remote control; the foreground playback log is a data log that records the playback content and duration of the digital TV in normal playback status, and is used to track user viewing behavior.

[0028] Through system-level performance monitoring interfaces and real-time collection of multi-dimensional user behavior data, comprehensive monitoring of digital TV operating status and user habits is achieved. This provides accurate and comprehensive input parameters for the user classification module, improving the accuracy and reliability of user classification. Based on detailed operating indicators and behavioral data, the test system can dynamically adjust test parameters and plans, enhancing the pertinence and adaptability of test configurations, effectively improving the efficiency and accuracy of digital TV testing, reducing resource waste, and enhancing the user experience and the level of intelligent system maintenance.

[0029] After receiving the digital TV operation data and user operation behavior data transmitted by the data collection module, the user classification module first initializes the test parameters. During the initialization process, based on the three parameters of CPU usage, startup time, and response time in the digital TV operation data, a segmented weighted method is used to standardize and weight the parameters. The specific operation is as follows: Normalize each original parameter value. For any original parameter X, let its normalized value be X', then the normalization formula is ,in, and They are the minimum and maximum values of the parameter in historical data respectively.

[0030] After normalization is completed, weight coefficients are set for different parameters. The weights are determined based on the degree of influence of the parameters on the performance. The specific calculation formula for the initialization test parameter T is: , where C' is the normalized value of CPU occupancy, S' is the normalized value of startup time, and R' is the normalized value of response time. is the CPU usage weight, is the startup duration weight, is the response time weight, and satisfy .

[0031] After the test parameters are initialized, user behavior data is used as input to classify users using a fuzzy clustering method. Clustering factors include the number of channel switches, daily cumulative viewing time, and the number of function button clicks.

[0032] Construct feature vectors for all users’ operation behavior data , clustering is performed using the fuzzy C-means clustering algorithm. is the number of channel switching times, The daily cumulative viewing time, The number of function button clicks.

[0033] Let the cluster center be (j=1,2,3, corresponding to high-frequency, medium-frequency and low-frequency users respectively), the fuzzy membership matrix is ,in Indicates that the i-th user belongs to the j-th class, satisfying .

[0034] The following objective function is used for minimization , where m>1 is the fuzzy coefficient, is the Euclidean distance.

[0035] After clustering is completed, the classification of each user is determined according to the maximum membership principle, that is, ; After completing the clustering of user usage frequency, the user categories are obtained, namely high-frequency users, medium-frequency users, and low-frequency users. The corresponding initialization test parameters T are matched to generate a personalized test configuration plan. The specific process is as follows: Based on historical data statistics and actual test requirements, define the classification threshold range of the test parameter T. The settings are shown in the following table:

[0036] in, and The threshold constant is obtained by professionals through experimental methods and meets the following requirements: .

[0037] For each user, based on their clustering labels (i.e., high-frequency users, medium-frequency users, and low-frequency users), and their corresponding test parameters T, the following matching rules are executed according to the above table: If the user is a high-frequency user and , then it is configured as high-precision test; If the user is a medium frequency user and , then it is configured as medium precision test; If the user is a low-frequency user and , it is configured as a basic test.

[0038] It should be noted that the segmented weighted method refers to an algorithm that divides the test parameters of different dimensions into several logical segments, assigns corresponding weights according to the importance of the parameters in each segment, and performs weighted synthesis on the original parameter values. It is used to achieve orderly unification of multi-dimensional test indicators in the initialization stage; the fuzzy clustering method is a clustering algorithm that allows samples to belong to multiple clusters at the same time. In this embodiment, the fuzzy C-means algorithm is used for frequency classification modeling of user usage behavior.

[0039] By using a segmented weighting method to scientifically normalize and weight digital TV operating parameters, the system accurately reflects the combined impact of each parameter on system performance, improving the rationality and precision of test parameter initialization. A fuzzy C-means clustering algorithm is combined with multi-dimensional analysis of user operating behavior to achieve dynamic and flexible classification of user usage frequency, effectively resolving the information loss problem associated with traditional hard classification. By precisely matching user classification results with initialization test parameters, the system generates personalized and hierarchical test configuration plans, ensuring the targeted and effective testing while optimizing test resource allocation, improving test efficiency and user satisfaction, and significantly enhancing the intelligence and adaptability of the digital TV test system.

[0040] In the solution analysis module, after receiving the classification result output by the user classification module, it is determined whether the current user is a medium-frequency user. If the current user is classified as a medium-frequency user, the user operation behavior data of the current user within the preset monitoring time interval starting at the current moment is called. The above preset monitoring time interval is seven days in this embodiment. The collected indicators include the number of channel switching times , Daily cumulative viewing time and the number of function button clicks .

[0041] Arrange the above three data into a behavioral feature matrix in chronological order , where the behavior vector on day t is , forming a matrix .

[0042] Normalize the above matrix. The normalization uses the maximum and minimum normalization formula, that is, ; After normalization, the standardized feature matrix is obtained , as the input of the convolutional neural network model.

[0043] The solution analysis module loads a pre-trained convolutional neural network model and uses user operation behavior data as input to analyze its behavioral feature patterns in the time series dimension to determine whether a high-precision testing solution is needed.

[0044] The convolutional neural network model is composed of the following structures in sequence, and the operation functions of each structure are as follows: The input layer is used to receive the normalized user behavior feature matrix , the matrix represents three types of behavioral data: the number of channel switches, viewing time, and function button clicks of the current user in 7 consecutive days. The matrix rows represent the time series dimension, and the columns represent the behavioral dimension.

[0045] The first convolutional layer uses 16 one-dimensional convolution kernels with a kernel size of 3, a stride of 1, and a padding of "valid". The convolution operation formula is as follows: ; in, is the weight of the j-th convolution kernel in the k-th dimension, is the bias term and ReLU is the activation function.

[0046] The maximum pooling layer performs a maximum pooling operation on the convolution output with a pooling window size of 2 and a stride of 2, taking the maximum value of every two consecutive time steps to reduce the feature dimension and retain significant features.

[0047] The second convolutional layer receives the pooled feature map and applies 32 one-dimensional convolution kernels with a kernel size of 3 and a stride of 1, also activated by ReLU.

[0048] Second maximum pooling layer: Consistent with the structure of the first pooling layer, continue to perform downsampling in the time dimension.

[0049] The flattening layer flattens the two-dimensional time series feature map into a one-dimensional vector as the input of the fully connected layer.

[0050] The fully connected layer receives the flattened feature vector, performs linear transformation and nonlinear activation, and outputs an intermediate representation vector.

[0051] The output layer contains one node, uses the Sigmoid activation function, and outputs a value y∈(0,1), which represents the matching probability between the current user behavior and the high-precision test configuration. The output formula is: ; Among them, h is the output vector of the fully connected layer, w is the output layer weight vector, and b is the bias term.

[0052] Set the accuracy judgment threshold δ∈(0,1). This accuracy judgment threshold is obtained by professionals through experimental optimization and will not be described in detail here. The specific judgment logic is as follows: If the model output value y ≥ δ, it is determined that the current user behavior complexity and interaction intensity are high, and a high-precision testing solution needs to be enabled; If the model output value y<δ, it is considered that the current behavior characteristics do not meet the high-precision test trigger standard, and the medium-precision test solution configuration remains unchanged.

[0053] It should be noted that the convolutional neural network model is a type of feedforward neural network with a local receptive field, weight sharing and downsampling mechanism, which is used in this embodiment to mine the time series features in user operation behavior data; a one-dimensional convolution kernel refers to a weight window that slides on one-dimensional time series input data to perform a convolution operation; the filling method "valid" is a boundary processing strategy in the convolution operation that is defined as not filling the input edges with zero; the activation function is a functional unit that introduces nonlinear factors into the neural network, which is used to perform nonlinear mapping on the weighted input of neurons; maximum pooling is a downsampling operation used to extract the dominant response of the feature map in the convolutional neural network and compress the spatial dimension; downsampling refers to the operation of reducing the length or dimension of the feature map by methods such as pooling to reduce the computational complexity and the risk of overfitting; nonlinear activation refers to the activation function applying a nonlinear transformation to the input data; the Sigmoid activation function is an S-shaped nonlinear function, which will not be described in detail here.

[0054] By introducing a deep learning model based on convolutional neural networks, it is possible to effectively mine the time series features and underlying patterns in user operational behavior data, enabling accurate identification of the behavioral complexity and interaction intensity of medium-frequency users. This solution transcends the limitations of traditional static rule-based judgments, enhances the intelligence of test solution selection, and ensures that high-precision test resources are accurately allocated to users who truly need them, avoiding resource waste. Furthermore, the use of multi-layer convolution and pooling structures effectively extracts key features and reduces computational complexity, improving system response speed and judgment accuracy. This significantly enhances the adaptability and dynamic adjustment capabilities of the digital TV test system, optimizing the efficiency and effectiveness of the overall test process.

[0055] The feedback update module collects user feedback data through the front-end interactive interface. The collected data includes the user's rating of the current test plan, denoted as s∈[0,5], in points. The rating value is directly selected by the user through a five-level satisfaction scale. Each user needs to complete a rating operation after the test plan usage cycle ends.

[0056] User feedback data is archived by user identifier to form a feedback data record set: ; in, Indicates the total number of users in the current category. represents the i-th user belonging to the category c∈{high frequency, medium frequency, low frequency}, For its corresponding rating value, all rating values are stored independently by category.

[0057] Perform statistical analysis on the feedback data record sets of high-frequency users, medium-frequency users, and low-frequency users, and calculate the mean of each type of feedback data record set and standard deviation The calculation formula is as follows: ; Based on the above mean and standard deviation, the feedback deviation interval of each type of user is defined as , where λ is the interval adjustment factor.

[0058] The current user feedback rating value is recorded as , compare the current user feedback score value with the feedback deviation interval of each category.

[0059] like , then update the current user classification to c′; like , the original classification remains unchanged; like , it will be recorded as abnormal feedback and the classification will not be updated temporarily.

[0060] It should be noted that the front-end interactive interface refers to the graphical user interface system presented on the user's device, which is used to realize data input and feedback interaction between the user and the test system.

[0061] The feedback update module systematically collects and statistically analyzes user feedback scores, constructs feedback deviation intervals using mean and standard deviation, and dynamically corrects and optimizes user classification. This mechanism effectively improves the accuracy and adaptability of user classification, ensuring that test plans can promptly reflect changes in user needs and satisfaction, and enhancing the system's self-learning and intelligent adjustment capabilities. Furthermore, identifying abnormal feedback avoids the negative impact of misclassification, improves system stability and reliability, and promotes the continuous optimization of the digital TV test system and enhances user experience.

[0062] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0063] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0064] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0065] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0066] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0067] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0068] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0069] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0070] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0071] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A digital television test system, characterized in that: It includes data collection module, user classification module, solution analysis module and feedback update module, and the signal connections between each module; The data collection module collects digital TV operation data and user operation behavior data; The user classification module initializes the test parameters using a segmented weighted method based on the digital TV operation data, and uses fuzzy clustering to analyze the user usage frequency based on the user operation behavior data, dividing the users into high-frequency users, medium-frequency users, and low-frequency users, and matching the classified users with the initialized test parameters; If the solution analysis module identifies the current user as a medium-frequency user, it collects the current user's operation behavior data and builds a convolutional neural network model based on the operation behavior data to analyze whether the current user needs a high-precision test solution; The feedback update module collects user feedback data, calculates the deviation interval based on the user feedback data, and updates the current user classification by comparing the current user feedback data with the deviation interval.

2. A digital television test system according to claim 1, characterized in that: Collect CPU usage, startup time and response time as digital TV operation data; The number of channel switching, daily cumulative viewing time, and number of function button clicks are collected as user operation behavior data.

3. A digital television test system according to claim 2, characterized in that: Normalizing the digital TV operation data and setting weight coefficients for different parameters to obtain initialization test parameters; The user operation behavior data is used as input to classify users using fuzzy clustering method to obtain the category to which the user belongs; Match the user's category with the corresponding initialization test parameters to obtain a personalized test configuration plan.

4. A digital television test system according to claim 3, characterized in that: Filter out medium-frequency users based on their categories; The user operation behavior data of the medium frequency user within the preset monitoring time interval starting from the current moment is called, and is sorted in chronological order to obtain a behavior feature matrix.

5. A digital television test system according to claim 4, characterized in that: The behavior feature matrix is normalized to obtain a standardized feature matrix; a convolutional neural network model is constructed using the standardized feature matrix as input to output the matching probability between the current user behavior and the high-precision test configuration.

6. A digital television test system according to claim 5, characterized in that: Set an accuracy determination threshold and compare it with the matching probability of the current user behavior and the high-precision test configuration; if the matching probability of the current user behavior and the high-precision test configuration is greater than or equal to the accuracy determination threshold, the high-precision test solution is enabled; If the probability of matching the current user behavior with the high-precision test configuration is less than the accuracy judgment threshold, the medium-precision test solution configuration remains unchanged.

7. A digital television test system according to claim 6, characterized in that: Collect user feedback data, including the user's rating of the current test plan; archive the user feedback data according to the user identifier to form a feedback data record set.

8. A digital television test system according to claim 7, characterized in that: Statistical analysis is performed on the feedback data record sets of high-frequency users, medium-frequency users, and low-frequency users, and the mean and standard deviation of each type of feedback data record set are calculated.

9. A digital television test system according to claim 8, characterized in that: The feedback deviation interval of each type of user is defined based on the mean and standard deviation of each type of feedback data record set.

10. A digital television test system according to claim 9, characterized in that: Compare the current user feedback score with the feedback deviation interval of each category; if the current user feedback score falls within the feedback deviation interval of a certain category, update the current user category to that category; If the current user feedback rating value falls within the feedback deviation range of the current category, the original category will remain unchanged; if the current user feedback rating value does not fall within the feedback deviation range of any category, it will be recorded as abnormal feedback and the category will not be updated for the time being.

Images