User satisfaction determination method and device, electronic equipment and readable storage medium

Abstract

The invention discloses a user satisfaction determination method and apparatus, an electronic device and a readable storage medium. The method comprises the steps of obtaining network use data of a user in a preset time period; dividing the network use data into a plurality of sub-data sets based on a preset active rule; wherein the plurality of sub-data sets are respectively corresponding to different user activity degrees; respectively processing the plurality of sub data sets to obtain satisfaction degree features respectively corresponding to the plurality of sub data sets; and determining the satisfaction degree of the user based on the satisfaction degree characteristics corresponding to the plurality of sub-data sets.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of artificial intelligence, and particularly relates to a user satisfaction determination method and device, electronic equipment and a readable storage medium. BACKGROUND

[0002] In related technologies, when the user satisfaction of a home broadband is managed, a threshold method and an artificial intelligence (AI) algorithm are mainly used for implementation. The data used by the threshold method and the AI algorithm is data related to complaints extracted from a large amount of original data. However, the data only represents part of the cases in which the home broadband service is below an acceptable level, so that there is a data bias when the user satisfaction is evaluated, and the influence of multiple possible cases of the home broadband on the user satisfaction cannot be covered, so that the user satisfaction cannot be accurately predicted based on the above method. SUMMARY

[0003] Embodiments of the present application provide a user satisfaction determination method, device, electronic equipment and readable storage medium, which can solve the problem that there is a data bias when the user satisfaction is evaluated, the influence of multiple possible cases of the home broadband on the user satisfaction cannot be covered, and the user satisfaction cannot be accurately predicted based on the above method.

[0004] In a first aspect, the embodiments of the present application provide a user satisfaction determination method, which includes: obtaining network usage data of a user in a preset time period; dividing the network usage data into multiple sub-data sets based on a preset activity rule; wherein the multiple sub-data sets respectively correspond to different user activity levels; processing the multiple sub-data sets respectively to obtain satisfaction features respectively corresponding to the multiple sub-data sets; and determining the satisfaction of the user based on the satisfaction features respectively corresponding to the multiple sub-data sets.

[0005] In a second aspect, the embodiments of the present application provide a user satisfaction determination device, which includes: an obtaining module configured to obtain network usage data of a user in a preset time period; a first determining module configured to divide the network usage data into multiple sub-data sets based on a preset activity rule; wherein the multiple sub-data sets respectively correspond to different user activity levels; a second determining module configured to process the multiple sub-data sets respectively to obtain satisfaction features respectively corresponding to the multiple sub-data sets; and a third determining module configured to determine the satisfaction of the user based on the satisfaction features respectively corresponding to the multiple sub-data sets.

[0006] In a third aspect, an electronic device is provided, which includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, and the program or instructions, when executed by the processor, implement the steps of the method according to the first aspect.

[0007] In a fourth aspect, a readable storage medium is provided, which stores a program or instructions, and the program or instructions, when executed by a processor, implement the steps of the method according to the first aspect.

[0008] In a fifth aspect, a chip is provided, which includes a processor and a communication interface, the communication interface is coupled to the processor, and the processor is configured to execute a program or instructions to implement the steps of the method according to the first aspect.

[0009] In a sixth aspect, a computer program product is provided, which includes a computer program stored in a non-transitory computer readable storage medium, and the computer program includes a program or instructions, and the program or instructions, when executed, implement the steps of the method according to the first aspect.

[0010] In the embodiments of the present application, by obtaining the network usage data of the user in the preset time period, the network usage data is divided into a plurality of sub-data sets based on a preset active rule, the plurality of sub-data sets correspond to different user activity levels respectively, and the plurality of sub-data sets are processed respectively to obtain satisfaction features corresponding to the plurality of sub-data sets respectively, and the satisfaction of the user is determined based on the satisfaction features corresponding to the plurality of sub-data sets respectively. Considering that different activity levels of the user have different experiences of network usage, the network usage data is classified according to the activity level of the user by the preset active rule, the activity level of the user is analyzed in detail, various situations of the user in the preset time period are covered, and the state of the user is accurately described. On the basis of accurately describing the state of the user, the plurality of sub-data sets after division are processed respectively, the satisfaction features corresponding to the plurality of sub-data sets obtained by comprehensive processing are determined, and the satisfaction of the user is determined, so that the user is analyzed in detail and comprehensively, and the satisfaction of the user is accurately determined. BRIEF DESCRIPTION OF DRAWINGS

[0011] Figure 1 is a flowchart of a user satisfaction determination method provided by the embodiments of the present application; Figure 2a is a division diagram of network usage data provided by the embodiments of the present application; Figure 2b is another division diagram of network usage data provided by the embodiments of the present application; Figure 3is a flowchart of another user satisfaction determination method provided by an embodiment of the present application; Figure 4 is a structural diagram of a Transformer network provided by an embodiment of the present application; Figure 5 is a flowchart of a user satisfaction determination and management method provided by an embodiment of the present application; Figure 6 is a flowchart of a training method of a user satisfaction reasoning network provided by an embodiment of the present application; Figure 7 is a time information embedding and position encoding diagram of data provided by an embodiment of the present application; Figure 8 is a structural diagram of a user satisfaction determination apparatus provided by an embodiment of the present application; Figure 9 is a structural diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0012] The technical solutions in the embodiments of the present application will be clearly described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art belong to the scope of protection of the present application.

[0013] The terms "first", "second", and the like in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than that illustrated or described herein, and the objects distinguished by "first", "second", etc. are usually a category, not limited to the number of objects, for example, the first object can be one or more. In addition, "and / or" in the specification and claims indicates at least one of the connected objects, and the character " / ", generally indicates that the front and rear associated objects are in an "or" relationship.

[0014] The user satisfaction determination method, apparatus, electronic device and readable storage medium provided by the embodiments of the present application will be described in detail below with reference to the drawings, through specific embodiments and their application scenarios.

[0015] Figure 1 A flowchart of a user satisfaction determination method provided by an embodiment of the present application is shown, which can be executed by an electronic device. Referring to Figure 1 , the method can include the following steps.

[0016] At step 102, network usage data of the user in a preset time period is acquired.

[0017] The network usage data includes not only the behavior habit related data of the user using the network, such as connection duration, access frequency, access time, etc., but also the response of the network when the user uses the network, such as the frequency and quantity of data flow, network quality, service response time, etc.

[0018] At step 104, the network usage data is divided into a plurality of sub-data sets based on preset active rules, wherein the plurality of sub-data sets correspond to different user activity levels respectively.

[0019] The user activity level can be measured from multiple aspects based on the diversity of the network usage data, so as to process the network usage data in a targeted manner.

[0020] At step 106, the plurality of sub-data sets are processed respectively to obtain satisfaction characteristics corresponding to the plurality of sub-data sets respectively.

[0021] The sub-data sets corresponding to different user activity levels are processed in a targeted manner, so as to avoid the characteristics of the data being ignored due to no differentiation, the state of the user being misunderstood, and the user satisfaction being evaluated incorrectly.

[0022] At step 108, the satisfaction of the user is determined based on the satisfaction characteristics corresponding to the plurality of sub-data sets respectively.

[0023] In the embodiments of the present application, the network usage data of the user in a preset time period is acquired, the network usage data is divided into a plurality of sub-data sets based on preset active rules, the plurality of sub-data sets correspond to different user activity levels respectively, the plurality of sub-data sets are processed respectively to obtain satisfaction characteristics corresponding to the plurality of sub-data sets respectively, and the satisfaction of the user is determined based on the satisfaction characteristics corresponding to the plurality of sub-data sets respectively. Considering that different activity levels of the user have different experiences of using the network, the network usage data is classified according to the activity level of the user by using the preset active rules, the activity level of the user is analyzed in detail, various situations of the user in the preset time period are covered, and the state of the user is accurately described. On the basis of accurately describing the state of the user, the plurality of sub-data sets after division are processed in a targeted manner, the satisfaction characteristics corresponding to the plurality of sub-data sets obtained by comprehensive processing are used to determine the satisfaction of the user, the user is analyzed in detail and comprehensively, and the user satisfaction is accurately determined.

[0024] In one implementation, the above network usage data may include: network data traffic, network connection duration, and network access frequency. The above step 104 divides the network usage data into multiple sub-datasets based on a preset activity rule, which may include the following steps.

[0025] Step 1041: According to the preset activity rule, respectively judge the network data traffic, the network connection duration, and the network access frequency, and determine the traffic characteristics, duration characteristics, and access characteristics corresponding to the network data traffic, the network connection duration, and the network access frequency respectively.

[0026] Among them, for network traffic data x, threshold values X1 and X2 are set in the preset activity rule, and X1 < X2. The traffic characteristics are as follows: (1) Active period: x > X2, indicating that when the user is using the network, the frequency and quantity of data traffic exceed a certain value; (2) Inactive period: X1 < x ≤ X2, indicating that the user uses the network less and the data traffic decreases to a lower level. (3) Silent period: x ≤ X1, indicating that when the user does not access the network at all, the data traffic is close to zero.

[0027] For network connection duration y, threshold values Y1 and Y2 are set in the preset activity rule, and Y1 < Y2. The duration characteristics are as follows: (1) Active period: y > Y2, and the user's connection duration to the network is longer during the active period. (2) Inactive period: Y1 < y ≤ Y2, indicating that although the user is connected to the network, the active time is short. (3) Silent period: y ≤ Y1, indicating that the user is not connected to the network at all and the connection duration is zero.

[0028] For network access frequency z, threshold values Z1 and Z2 are set in the preset activity rule, and Z1 < Z2. The access characteristics are as follows: (1) Active period: z > Z2, and the user's access frequency to the network is higher during the active period. (2) Inactive period: Z1 < z ≤ Z2, indicating that although the user accesses the network occasionally, the access frequency is low. (3) Silent period: z ≤ Z1, indicating that the user does not access the network at all and the access frequency is zero.

[0029] Step 1042: According to the traffic characteristics, the time information corresponding to the traffic characteristics, the duration characteristics, the time information corresponding to the duration characteristics, the access characteristics, and the time information corresponding to the access characteristics, determine the user activity level in multiple sub-periods within the preset period.

[0030] Among them, for the complex state of the user, the preset activity rule further includes: (1) High-frequency user judgment condition ①: Satisfy .

[0031] (2) Low-frequency user judgment condition ②: Satisfy , where X1>0, Y1>0.

[0032] (3) Long-term user judgment condition ③: meets , where Y2>0.

[0033] (4) Short-term user judgment condition ④: Satisfies , where Y1>0, Y2>0.

[0034] (5) High-flow user judgment condition ⑤: meets , where X2>0.

[0035] (6) Low-flow user judgment condition ⑥: meets , where X1>0, X2>0.

[0036] Based on this, regarding the user's network usage data:

[0037] The user's activity level can be determined in multiple sub-time periods: if the user's network usage data meets ①③⑤, the user's activity level is determined to be active; if the user's network usage data meets ④⑥, the user's activity level is determined to be inactive; if the user's network usage data meets ②, the user's activity level is determined to be silent.

[0038] Step 1043: Divide the network usage data into multiple subsets based on the user activity levels of the multiple sub-time periods.

[0039] Among them, see Figure 2a and Figure 2b Multiple sub-time periods can include active periods, inactive periods, and silent periods. User activity levels are represented by active, inactive, and silent periods. Multiple subset datasets can include active period data, inactive period data, and silent period data.

[0040] In this embodiment, user network usage data within a preset time period is judged based on preset activity rules to determine the user activity level of multiple sub-time periods within that preset time period. The network usage data is then divided according to the user activity level, and visualized as follows: Figure 2a and Figure 2bAs shown. Based on the types of network usage data, each is first judged separately. According to preset activity rules, the network data traffic, network connection duration, and network access frequency are judged to determine the traffic characteristics, duration characteristics, and access characteristics corresponding to the network data traffic, network connection duration, and network access frequency, respectively. Then, time information is comprehensively judged. Based on the traffic characteristics, the time information corresponding to the traffic characteristics, the duration characteristics, the time information corresponding to the duration characteristics, the access characteristics, and the time information corresponding to the access characteristics, the user activity level of multiple sub-time periods in the preset time period is determined. The user's status within the preset time period is described in detail, and the user's network usage data within the preset time period is divided into multiple subsets to facilitate subsequent targeted processing of each subset.

[0041] In one implementation, the aforementioned network usage data needs to consider its time information and network parameters. Time information has already been considered. In this embodiment, the impact of network parameters on determining satisfaction characteristics is added. Step 1042, which determines the user activity level of multiple sub-time periods within the preset time period based on the traffic characteristics, the time information corresponding to the traffic characteristics, the duration characteristics, the time information corresponding to the duration characteristics, the access characteristics, and the time information corresponding to the access characteristics, may include the following steps.

[0042] Step 10421: Obtain the network parameters of the user's network usage under the time information, and determine the weight corresponding to the time information based on the network parameters.

[0043] Among them, network parameters are static or dynamic state information configured and generated to maintain network operation, indicating how the network is built and how it operates, such as gateway addresses, routing protocols, firewall rules, or bandwidth limits.

[0044] Step 10422: Determine the user activity level of multiple sub-time periods based on the traffic characteristics, the time information corresponding to the traffic characteristics, the duration characteristics, the time information corresponding to the duration characteristics, the access characteristics, the time information corresponding to the access characteristics, and the weights corresponding to each of the time information.

[0045] In this embodiment, considering the strong correlation between changes in user satisfaction and different network parameters, the network parameters used by users at each time point are mapped as weights to participate in the calculation of user activity levels. For example, users with better network parameters are more likely to use the network more actively, and the weight corresponding to those network parameters is greater. Therefore, when determining user activity levels across multiple sub-time periods, both the influence of user behavior and network parameters are considered, thereby improving the accuracy of determining user activity levels and enabling precise segmentation of network usage data.

[0046] In one implementation, step 106 above processes multiple subsets of data to obtain satisfaction features corresponding to each subset of data, and may include the following steps.

[0047] Step 1061: Process each of the subsets through the Transformer network corresponding to each subset to obtain the first satisfaction feature.

[0048] Among them, such as Figure 3 As shown, the network usage data, after the above processing, is divided into active period data, inactive period data, and silent period data. These three subsets are processed through their respective Transformer networks to determine the first satisfaction feature for each subset, which serves as the user's satisfaction feature for one aspect of the sub-time period.

[0049] Among them, such as Figure 4 As shown, the Transformer network is a neural network based on an attention mechanism, suitable for processing dependencies in long sequences of data. This Transformer network includes an embedding layer, at least one self-attention encoder, at least one self-attention decoder, a Linear layer, and a Softmax layer. Through self-attention and positional encoding, the Transformer network can capture the relationships between different positions in long sequences of data, where these positions correspond to the network parameters in this embodiment.

[0050] Step 1062: Process each of the subsets through the Temporal Convolutional Network (TCN) corresponding to each subset to obtain the second satisfaction feature.

[0051] Among them, the three subsets of data—active period data, inactive period data, and silent period data—are further processed through their respective TCNs to determine their respective second satisfaction features, which serve as another aspect of user satisfaction features in the sub-time period.

[0052] Temporal Convolutional Networks (TCNs) are a type of neural network based on convolutional structures that can handle long-term dependency problems. TCNs utilize convolutional structures to capture local dependencies in long-sequence data and effectively handle dependencies in long-sequence data through residual connections and dilated convolutions.

[0053] Step 1063: Perform feature fusion processing on the first satisfaction feature and the second satisfaction feature to determine the user's satisfaction feature for each sub-time period.

[0054] In this embodiment, the Transformer network and TCN are combined to process the subsets of user activity levels in a targeted manner. Each subset is processed by the Transformer network and TCN corresponding to each subset to capture the correlation between network usage data and user satisfaction. The first satisfaction feature and the second satisfaction feature are then fused together to compensate for the correlations that have not been noticed, making the obtained satisfaction features more accurate.

[0055] In one implementation, after determining the user's satisfaction based on the satisfaction features corresponding to the multiple subsets of data in step 108 above, the method further includes the following steps.

[0056] Step 110: If the user's satisfaction level is lower than a preset satisfaction threshold, update the network parameters of the Transformer network and / or the TCN based on the user's dissatisfaction and the network status; wherein the user's dissatisfaction is determined based on the user's satisfaction level.

[0057] The network state can be categorized into whether the network is in a latency adaptation period and whether it is experiencing frequent fluctuations. When the network is in a latency adaptation period, there are two possible outcomes: one is to wait until the latency adaptation period is over and continue using the current Transformer network and / or TCN; the other is to continue assessing whether the network is experiencing frequent fluctuations, which further includes two outcomes: one is that the network is not experiencing frequent fluctuations, the network state is in the latency adaptation period, but there are no other anomalies, and the current Transformer network and / or TCN can continue to be used; the other is that the network is experiencing frequent fluctuations, and the current Transformer network and / or TCN will no longer be suitable for the current network state, thus requiring updates to the network parameters of the Transformer network and / or TCN to adapt to the current network state, thereby accurately determining user satisfaction.

[0058] Step 112: Determine the user's satisfaction using the Transformer network and / or TCN after the network parameters are updated.

[0059] In this embodiment, if the user's satisfaction level is lower than a preset satisfaction threshold, it indicates that the current user satisfaction is low, and it is necessary to analyze the reasons for the low user satisfaction. Since the user's satisfaction is low, multiple judgments are made based on the user's dissatisfaction and the network state to confirm the network state. The user satisfaction is then determined using a Transformer network and / or TCN adapted to the current network state, ensuring the accuracy of the user satisfaction assessment.

[0060] In one implementation, step 110 above, which updates the network parameters of the Transformer network and / or the TCN based on the user's dissatisfaction and the network status, may include the following steps.

[0061] Step 1101: When the user's dissatisfaction exceeds a preset dissatisfaction threshold, the network is in a delayed adaptation period, and the network fluctuates frequently, determine the correction vector of the network parameters of the Transformer network and / or the TCN based on the network usage data and the preset network status.

[0062] The preset network status is determined based on the network parameters of the user's network when the user's satisfaction is greater than the preset satisfaction threshold, and is used as a standard.

[0063] Step 1102: Update the network parameters of the Transformer network and / or the TCN based on the correction vector of the network parameters of the Transformer network and / or the TCN and the loss value of the Transformer network and / or the TCN.

[0064] In the embodiments of this application, Figure 5 This is a flowchart illustrating a user satisfaction determination and management method provided in an embodiment of this application, such as... Figure 5As shown, user network usage data over a preset time period is input into a user satisfaction inference network for processing to obtain user satisfaction. This user satisfaction inference network includes the Transformer network and / or the TCN, a fully connected layer, and a Softmax layer. Multiple judgments are made based on user satisfaction. If the user satisfaction is lower than a preset satisfaction threshold, then user dissatisfaction is assessed. If the user dissatisfaction exceeds a preset dissatisfaction threshold, the network is in a delayed adaptation period, and the network experiences frequent fluctuations, it is determined that the current user satisfaction inference network is not suitable for the current network state. Therefore, based on the network usage data and the preset network state, a correction vector for the network parameters of the Transformer network and / or the TCN is determined. Then, combined with the loss value of the Transformer network and / or the TCN, the network parameters of the Transformer network and / or the TCN are updated so that the updated Transformer network and / or TCN are suitable for the current network state. Figure 5 The system uses a user satisfaction recovery inference network to accurately determine the user's satisfaction level.

[0065] In one implementation, the above Figure 1 The method shown can be implemented using a pre-trained user satisfaction inference network. Figure 6 The diagram illustrates a flowchart of a training method for a user satisfaction inference network provided in an embodiment of this application. This method can be executed by an electronic device. See also... Figure 6 The method may include the following steps.

[0066] Step 601: Obtain the training dataset for the user satisfaction inference network.

[0067] The training dataset includes users' network usage data over historical time periods. This data may include user behavior patterns such as network connection duration and network access frequency, as well as network response data such as network traffic, network quality, and service response time. The user in question is considered a dissatisfied user, meaning their satisfaction level is below a preset satisfaction threshold; satisfied users are excluded to improve training efficiency.

[0068] Step 602: Based on preset activity rules, classify network usage data according to user activity levels to determine multiple sub-training datasets. These sub-training datasets may include active period data, inactive period data, and silent period data.

[0069] Among them, the network usage data includes network data traffic x, network connection duration y, and network access frequency z. For the above network usage data, the preset activity rules may include: two thresholds X1 and X2 for network data traffic, where X1 < X2; two thresholds Y1 and Y2 for network connection duration, where Y1 < Y2; and two thresholds Z1 and Z2 for network access frequency, where Z1 < Z2.

[0070] Classifying the network usage data according to the user activity level based on the preset activity rules may include: (1) High-frequency user judgment condition ①: Satisfy .

[0071] (2) Low-frequency user judgment condition ②: Satisfy , where X1 > 0 and Y1 > 0.

[0072] (3) Long-duration user judgment condition ③: Satisfy , where Y2 > 0.

[0073] (4) Short-duration user judgment condition ④: Satisfy , where Y1 > 0 and Y2 > 0.

[0074] (5) High-traffic user judgment condition ⑤: Satisfy , where X2 > 0.

[0075] (6) Low-traffic user judgment condition ⑥: Satisfy , where X1 > 0 and X2 > 0.

[0076] It is possible to determine the user activity level in multiple sub-time periods: If the user's network usage data satisfies ①③⑤, determine that the user activity level is active; if the user's network usage data satisfies ④⑥, determine that the user activity level is inactive; if the user's network usage data satisfies ②, determine that the user activity level is silent.

[0077] Step 603: Perform time information embedding and positional encoding on multiple sub-training data sets.

[0078] Among them, for the input vectors, that is, active period data, inactive period data, and silent period data, performing time information embedding and positional encoding is as Figure 7 shown.

[0079] The time information embedding therein includes: For the input vector , there is , where is an eigenvalue in the input vector. The considered time information includes: month , date , week , holiday Is the network in an abnormal or faulty state? The time state information embedding expression is constructed as follows: The time-state information embedded in the input vector. , where k is the scaling factor and c is the cost constant, and k and c are usually taken as empirical values.

[0080] The positional encoding includes the following: Considering the significant non-linear relationship between changes in user satisfaction and changes in network parameters, and the large dynamic changes in the relative position of the input vector when network parameters deteriorate, the same method as the Google attention model is reused when calculating positional encoding in the Transformer network. and .

[0081] Step 604: Train the Transformer network and / or TCN corresponding to each of the multiple sub-training datasets independently, and then fuse and train the fully connected layer until convergence to obtain the trained user satisfaction inference network.

[0082] In some embodiments, the training dataset can be further optimized by first obtaining samples from a wider range of data sources, including user data with high, medium, and low satisfaction levels, as well as user feedback data under various different circumstances. This increases the diversity of the samples, enabling the user satisfaction inference network to better cover the satisfaction space. Simultaneously, the quality of the training data is controlled, and necessary data cleaning, annotation verification, and anomaly detection are performed to ensure the integrity and accuracy of the training data and avoid the impact of erroneous data on training. Specific methods can be found in [reference needed]. Figure 2a To smooth the training data, the size of the sliding window T is increased, for example, by expanding T to 10 vectors. Introducing more feature dimensions, such as user habits, network quality metrics, and service response time, enriches the network's input features, thereby improving the network's ability to predict satisfaction levels. Joint training with a dataset of satisfied users and a dataset of low-rated users further enhances the network's ability to perceive low-rated users, but this also increases the training difficulty. Therefore, this embodiment uses a dataset of dissatisfied users for training to improve training efficiency.

[0083] In other embodiments, after determining user satisfaction using the user satisfaction inference network described above, the method may further include evaluating the quality of the user satisfaction inference network.

[0084] Based on the mapping table between user dissatisfaction probability and satisfaction rating, as shown in Table 1, the satisfaction rating score ranges in Table 1 are fixed. The thresholds X1, X2, Y1, Y2, Z1, Z2 for network usage data are reused. Depending on the duration of the current complainant's dissatisfaction (e.g., 24 or 48 hours), the corresponding X1, X2, Y1, Y2, Z1, Z2 parameters are adjusted to ensure that the user dissatisfaction probability value in Table 1 corresponds to the rating range. This correspondence essentially learns the user's rating habits and the degree of subjectivity in mapping their network usage experience to score ranges. Then, a local case evaluation dataset is constructed. Data sets before and after the user complaint are categorized by region, and sets of data confirmed as typical after manual processing are combined into an evaluation dataset containing real input samples and corresponding labels or manually annotated results. This ensures that the evaluation dataset covers various situations and boundary conditions that the network may encounter.

[0085] Table 1.

[0086] In other embodiments, the user satisfaction inference network is adjusted and optimized based on the evaluation results of its quality. This includes adjusting the network's hyperparameters, increasing training data, and fine-tuning the network.

[0087] The methods for quality assessment, adjustment, and optimization of the user satisfaction inference network can be adjusted according to the application scenario and task requirements. For example, possible methods include combining user feedback and A / B testing.

[0088] In other embodiments, see Figure 5 After determining user satisfaction through the user satisfaction inference network, multiple judgments are made based on this satisfaction level. If the user satisfaction is below a preset satisfaction threshold, then a judgment is made regarding user dissatisfaction. If the user dissatisfaction exceeds a preset dissatisfaction threshold, the network is in a delayed adaptation period, or the network experiences frequent fluctuations, it is determined that the current user satisfaction inference network is not suitable for the current network state. Therefore, based on the network usage data and the preset network state, a correction vector for the user satisfaction inference network is determined. .

[0089] in,‖ Determined based on cosine distance: .

[0090] Then, by combining the loss value of the user satisfaction inference network, the user satisfaction inference network is updated to obtain the restored user satisfaction inference network. This restored user satisfaction inference network can be applied to the current network state and make accurate predictions.

[0091] In other embodiments, when the network status exceeds the adjustable parameter range, a "manual intervention required" status report is submitted to the backend. This report provides reference results for automatic network anomaly detection and fault location, and assists the backend in performing manual intervention. This includes manual and manual maintenance of the network anomaly or fault based on factors such as priority and ease of maintenance. Upon receiving a complaint from a home broadband user, the system also includes inserting a "complaint user satisfaction recovery" knowledge graph between the user satisfaction recovery inference network and the user satisfaction inference network, based on a proactive "manual intervention" strategy. This assists relevant personnel handling the complaint in troubleshooting and selecting solutions to restore user satisfaction as quickly as possible.

[0092] For example, if network latency frequently fluctuates from within 100 milliseconds to 150 milliseconds, optimizing network parameters and increasing bandwidth can reduce these frequent latency fluctuations and prevent further decline in user satisfaction. Conversely, when latency exceeds 200 milliseconds, further increasing bandwidth and optimizing routing can be implemented to restore user satisfaction as quickly as possible, thus achieving effective management of both network status and user satisfaction.

[0093] It should be noted that the user satisfaction determination method provided in this application embodiment can be executed by a user satisfaction determination device or a control module within that device for executing the user satisfaction determination method. This application embodiment uses the execution of the method by a user satisfaction determination device as an example to illustrate the user satisfaction determination device provided in this application embodiment.

[0094] Figure 8 This paper shows a schematic diagram of a user satisfaction determination device according to an embodiment of this application. See also: Figure 8 The device 800 may include: an acquisition module 81, a first determination module 82, a second determination module 83, and a third determination module 84.

[0095] The system includes: an acquisition module 81 for acquiring network usage data of a user within a preset time period; a first determination module 82 for dividing the network usage data into multiple subsets based on preset activity rules, wherein the multiple subsets correspond to different levels of user activity; a second determination module 83 for processing the multiple subsets to obtain satisfaction features corresponding to each subset; and a third determination module 84 for determining the user's satisfaction based on the satisfaction features corresponding to each subset.

[0096] In one implementation, the first determining module 82 described above can be used to determine the network data traffic, the network connection duration, and the network access frequency according to the preset activity rules, and to determine the traffic characteristics, duration characteristics, and access characteristics corresponding to the network data traffic, the network connection duration, and the network access frequency, respectively; to determine the user activity level of multiple sub-time periods in the preset time period according to the traffic characteristics, the time information corresponding to the traffic characteristics, the duration characteristics, the time information corresponding to the duration characteristics, the access characteristics, and the time information corresponding to the access characteristics; and to divide the network usage data into multiple subsets according to the user activity levels of the multiple sub-time periods.

[0097] In one implementation, the first determining module 82 described above can also be used to obtain network parameters of the user's network usage under the time information, determine the weight corresponding to the time information based on the network parameters, and determine the user activity level of multiple sub-time periods based on the traffic characteristics, the time information corresponding to the traffic characteristics, the duration characteristics, the time information corresponding to the duration characteristics, the access characteristics, the time information corresponding to the access characteristics, and the weight corresponding to each of the time information.

[0098] In one implementation, the second determining module 83 described above can be used to process each of the subsets through a Transformer network corresponding to each subset to obtain a first satisfaction feature; process each of the subsets through a Temporal Convolutional Network (TCN) corresponding to each subset to obtain a second satisfaction feature; and perform feature fusion processing on the first satisfaction feature and the second satisfaction feature to determine the satisfaction feature corresponding to each subset.

[0099] In one implementation, the aforementioned apparatus 800 may further include an update module, configured to update the network parameters of the Transformer network and / or the TCN based on the user's dissatisfaction and the network status when the user's satisfaction is lower than a preset satisfaction threshold; wherein the user's dissatisfaction is determined based on the user's satisfaction; the aforementioned third determination module 84 may be used to determine the user's satisfaction through the updated Transformer network and / or TCN.

[0100] In one implementation, the update module described above can be used to determine the correction vector of the network parameters of the Transformer network and / or the TCN based on the network usage data and the state of the preset network when the user's dissatisfaction exceeds a preset dissatisfaction threshold, the network is in a latency adaptation period, and the network fluctuates frequently; and update the network parameters of the Transformer network and / or the TCN based on the correction vector of the network parameters of the Transformer network and / or the TCN and the loss value of the Transformer network and / or the TCN.

[0101] The user satisfaction determination device in this application embodiment can be an electronic device or a component within an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices besides a terminal. For example, the electronic device can be a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM, or self-service machine, etc. This application embodiment does not specifically limit the device.

[0102] The user satisfaction determination device in this application embodiment can be a device with an operating system. The operating system can be Android, iOS, or other possible operating systems; this application embodiment does not specifically limit the specific operating system.

[0103] The user satisfaction determination device provided in this application embodiment can achieve... Figures 1 to 7 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0104] Based on the same technical concept, embodiments of this application also provide an electronic device for performing the above-described user satisfaction determination method. Figure 9This is a schematic diagram of the structure of an electronic device to implement the various embodiments of this application. The electronic device can vary significantly due to differences in configuration or performance, and may include a processor 901, a communications interface 902, a memory 903, and a communication bus 904. The processor 901, communications interface 902, and memory 903 communicate with each other via the communication bus 904. The processor 901 can call a computer program stored in the memory 903 and executable on the processor 901 to perform the various steps of the user satisfaction determination method embodiments described above, achieving the same technical effects. To avoid repetition, further details are omitted here.

[0105] It should be noted that the electronic devices in the embodiments of this application include servers, terminals, or other devices besides terminals. For example, automobiles, robots, and handheld devices.

[0106] The above electronic device structure does not constitute a limitation on the electronic device. An electronic device may include more or fewer components than illustrated, or combine certain components, or arrange them differently. For example, an input unit may include a Graphics Processing Unit (GPU) and a microphone, and a display unit may use a liquid crystal display (LCD), organic light-emitting diode (OLED), or other similar display panels. User input units include at least one of a touch panel and other input devices. A touch panel is also called a touchscreen. Other input devices may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, power buttons, etc.), trackballs, mice, and joysticks, which will not be elaborated further here.

[0107] Memory can be used to store software programs and various data. Memory can primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area can store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, memory can include volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (Synchlink DRAM, SLDRAM), and direct memory bus RAM (DRRAM).

[0108] The processor may include one or more processing units; optionally, the processor integrates an application processor and a modem processor, wherein the application processor mainly handles operations related to the operating system, user interface, and applications, while the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into the processor.

[0109] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described user satisfaction determination method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0110] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0111] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described user satisfaction determination method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0112] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0113] This application also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes a program or instructions. When the program or instructions are executed, they implement the various processes of the above-described user satisfaction determination method embodiments and can achieve the same technical effect. To avoid repetition, they will not be described again here.

[0114] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0115] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0116] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for determining user satisfaction, characterized in that, include: Obtain user network usage data within a preset time period; Based on preset activity rules, the network usage data is divided into multiple subsets; wherein, each subset corresponds to a different level of user activity. The multiple subsets of data are processed separately to obtain the satisfaction features corresponding to each subset of data. The user's satisfaction is determined based on the satisfaction features corresponding to the multiple subsets of data.

2. The method according to claim 1, characterized in that, The network usage data includes: network data traffic, network connection duration, and network access frequency; the network usage data is divided into multiple subsets based on preset activity rules, including: Based on the preset activity rules, the network data traffic, the network connection duration, and the network access frequency are judged respectively to determine the traffic characteristics, duration characteristics, and access characteristics corresponding to the network data traffic, network connection duration, and network access frequency, respectively; Based on the traffic characteristics, the time information corresponding to the traffic characteristics, the duration characteristics, the time information corresponding to the duration characteristics, the access characteristics, and the time information corresponding to the access characteristics, the user activity level of multiple sub-time periods in the preset time period is determined; The network usage data is divided into multiple subsets based on the user activity levels across multiple sub-time periods.

3. The method according to claim 2, characterized in that, The step of determining the user activity level of multiple sub-time periods within the preset time period based on the traffic characteristics, the time information corresponding to the traffic characteristics, the duration characteristics, the time information corresponding to the duration characteristics, the access characteristics, and the time information corresponding to the access characteristics includes: Obtain the network parameters of the user's network usage under the given time information, and determine the weight corresponding to the time information based on the network parameters; Based on the traffic characteristics, the time information corresponding to the traffic characteristics, the duration characteristics, the time information corresponding to the duration characteristics, the access characteristics, the time information corresponding to the access characteristics, and the weights corresponding to each of the time information, the user activity levels of multiple sub-time periods are determined.

4. The method according to claim 1, characterized in that, The step of processing multiple subsets of data to obtain satisfaction features corresponding to each subset includes: Each of the aforementioned subsets is processed through the Transformer network corresponding to each subset to obtain the first satisfaction feature; Each of the aforementioned subsets is processed through the Temporal Convolutional Network (TCN) corresponding to each subset to obtain the second satisfaction feature; The first satisfaction feature and the second satisfaction feature are fused together to determine the satisfaction feature corresponding to each of the subsets.

5. The method according to claim 4, characterized in that, After determining the user's satisfaction based on the satisfaction features corresponding to the multiple subsets of data, the method further includes: If the user's satisfaction level is lower than a preset satisfaction threshold, the network parameters of the Transformer network and / or the TCN are updated based on the user's dissatisfaction and the network status; wherein, the user's dissatisfaction is determined based on the user's satisfaction level. The user's satisfaction is determined using the Transformer network and / or TCN after the network parameters are updated.

6. The method according to claim 5, characterized in that, The network status includes whether the network is in a latency adaptation period and whether the network is experiencing frequent fluctuations; updating the network parameters of the Transformer network and / or the TCN based on the user's dissatisfaction and the network status includes: When the user's dissatisfaction exceeds a preset dissatisfaction threshold, the network is in a delay adaptation period, and the network fluctuates frequently, the correction vector of the network parameters of the Transformer network and / or the TCN is determined based on the network usage data and the preset network status. The network parameters of the Transformer network and / or the TCN are updated based on the correction vector of the network parameters of the Transformer network and / or the TCN and the loss value of the Transformer network and / or the TCN.

7. A device for determining user satisfaction, characterized in that, include: The acquisition module is used to acquire user network usage data within a preset time period; The first determining module is used to divide the network usage data into multiple subsets based on preset activity rules; wherein the multiple subsets correspond to different user activity levels. The second determining module is used to process the multiple subsets of data respectively to obtain the satisfaction features corresponding to the multiple subsets of data respectively; The third determining module is used to determine the user's satisfaction based on the satisfaction features corresponding to the multiple subsets of data.

8. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the user satisfaction determination method as described in any one of claims 1 to 6.

9. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the user satisfaction determination method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program including programs or instructions that, when executed, implement the steps of the user satisfaction determination method as described in any one of claims 1 to 6.

Images