Scoring method for mobile application service, and storage medium and electronic device

Abstract

Provided in the embodiments of the present disclosure are a score method for a mobile application service, and a storage medium and an electronic device. The method comprises: on the basis of a preset time interval, segmenting traffic data generated by a mobile application service into a plurality of traffic data segments, and performing feature serialization to determine serialized numerical features of each traffic data segment; and inputting the serialized numerical features of the traffic data segments into a mobile application service scoring model, so as to output the score of the mobile application service.

Description

Scoring method, storage medium and electronic device for mobile application business

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] The present disclosure is based on Chinese Patent Publication No. 2024102878349 filed on March 13, 2024, entitled “Rating Method, Storage Medium and Electronic Device for Mobile Application Business”, and claims the priority of the patent disclosure, and all the disclosed contents thereof are incorporated into the present disclosure by reference. Technical Field

[0003] The embodiments of the present disclosure relate to the field of communications, and in particular, to a scoring method, storage medium, and electronic device for mobile application services. Background Art

[0004] With the development of communication technology, mobile applications have become extremely popular. To timely perceive network quality and provide customers with a better mobile application experience (especially for services that rely heavily on network quality, such as video, VR, and voice), operators need a method that can comprehensively and quantitatively reflect the user experience of specific applications from multiple dimensions, taking into account service characteristics. On the other hand, for security reasons, service providers often use encrypted traffic to transmit interaction data between terminal applications and servers, which also greatly complicates the application experience quality scoring process.

[0005] The application scoring methods used in related technologies rely on the assumption that messages are transmitted in plaintext, in order to obtain the decoded data within them. Furthermore, even in plaintext transmission scenarios, messages do not necessarily carry all the decoded data required for scoring. Furthermore, encrypted transmission is currently an established trend, and in this case, all decoded data cannot be extracted from the message payload, making this method not universally applicable. Summary of the Invention

[0006] Embodiments of the present disclosure provide a scoring method for a mobile application service, a storage medium, and an electronic device.

[0007] According to one embodiment of the present disclosure, a scoring method for mobile application services is provided, comprising: cutting traffic data generated by the mobile application service into a plurality of traffic data segments according to preset time intervals, and performing feature serialization to determine the serialized numerical features of each of the traffic data segments; inputting the serialized numerical features of the traffic data segments into a mobile application service scoring model, and outputting a score for the mobile application service.

[0008] According to another embodiment of the present disclosure, a computer-readable storage medium is provided, in which a computer program is stored. The computer program is configured to execute the steps in the above method embodiment when running.

[0009] According to another embodiment of the present disclosure, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in the above method embodiment. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] FIG1 is a hardware structure block diagram of a computer terminal according to a method for scoring a mobile application service according to an embodiment of the present disclosure;

[0011] FIG2 is a flow chart of a scoring method for a mobile application service according to an embodiment of the present disclosure;

[0012] FIG3 is a flow chart of a scoring method for a mobile application service according to yet another embodiment of the present disclosure;

[0013] FIG4 is a flow chart of a data collection method according to an embodiment of the present disclosure;

[0014] FIG5 is a schematic diagram of the composition of original sample data according to an embodiment of the present disclosure;

[0015] FIG6 is a flow chart of a scoring formula design method according to an embodiment of the present disclosure;

[0016] FIG7 is a flowchart of a tag calculation method according to an embodiment of the present disclosure;

[0017] FIG8 is a flowchart of a method for determining an original vector sequence according to an embodiment of the present disclosure;

[0018] FIG9 is a schematic diagram of feature vector splicing according to an embodiment of the present disclosure;

[0019] FIG10 is a schematic diagram of features of a serialized packet capture file according to an embodiment of the present disclosure;

[0020] FIG11 is a flowchart of a method for generating a mobile service scoring model according to an embodiment of the present disclosure;

[0021] FIG12 is a partial schematic diagram of a network structure according to an embodiment of the present disclosure;

[0022] FIG13 is a second partial schematic diagram of a network structure according to an embodiment of the present disclosure;

[0023] FIG14 is a structural block diagram of a scoring apparatus for a mobile application service according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings and in conjunction with embodiments.

[0025] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0026] The method embodiments provided in the embodiments of the present disclosure can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a computer terminal as an example, FIG1 is a hardware structure block diagram of a computer terminal of a scoring method for a mobile application business in an embodiment of the present disclosure. As shown in FIG1 , the computer terminal may include one or more (only one is shown in FIG1 ) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 configured to store data, wherein the above-mentioned computer terminal may also include a transmission device 106 and an input and output device 108 configured to have a communication function. It can be understood by those skilled in the art that the structure shown in FIG1 is only for illustration, and it does not limit the structure of the above-mentioned computer terminal. For example, the computer terminal may also include more or fewer components than those shown in FIG1 , or have a configuration different from that shown in FIG1 .

[0027] The memory 104 can be configured to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the scoring method for the mobile application business in the embodiment of the present disclosure. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, implementing the above-mentioned method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely located relative to the processor 102, and these remote memories may be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0028] The transmission device 106 is configured to receive or transmit data via a network. A specific example of the aforementioned network may include a wireless network provided by a communications provider of a computer terminal. In one embodiment, the transmission device 106 includes a network interface controller (NIC) that can be connected to other network devices via a base station to enable communication with the Internet. In another embodiment, the transmission device 106 may be a radio frequency (RF) module that is configured to communicate with the Internet wirelessly.

[0029] The following methods are commonly used in related technologies to score application experience (taking video services as an example):

[0030] The first step is to obtain message decoding data: extract fields such as video bit rate and resolution that are highly relevant to user experience from the payload of the user-plane message.

[0031] The second step is calculating factors influencing user experience: The decoded data extracted in the first step is combined according to business logic to derive various factors influencing the score. For example, the number and duration of video freezes can be calculated using traffic and bitrate information. Certain decoded data, such as video resolution, can also be directly considered as influencing factors. These factors are typically mapped to a numerical range on the same scale to eliminate dimensionality.

[0032] The third step is to calculate the final score based on the impact factors: For each impact factor obtained in the second step, different weights are assigned according to experience or specific needs, and each impact factor is weighted and calculated to obtain the final score for the business.

[0033] FIG2 is a flow chart of a method for scoring a mobile application service according to an embodiment of the present disclosure. As shown in FIG2 , the process includes the following steps:

[0034] Step S202 : cutting the traffic data generated by the mobile application service into a plurality of traffic data segments according to a preset time interval, and performing feature serialization to determine the serialized numerical features of each traffic data segment.

[0035] In one embodiment, after dividing the traffic data generated by a mobile application service into multiple traffic data segments, the basic characteristics of the traffic within each traffic data segment are determined. The basic characteristics include uplink traffic, downlink traffic, and the number of uplink packets or downlink packets. Each basic feature in each traffic data segment is then represented in vector form. All the basic features in vector form in each traffic data segment are concatenated to determine the original vector of each traffic data segment, thereby determining the original vector sequence corresponding to the traffic data generated by the mobile application service.

[0036] Step S204: inputting the serialized numerical features of the traffic data segments into a mobile application service scoring model, and outputting a score of the mobile application service.

[0037] In one embodiment, the original vector sequence is further encoded through a self-attention model to generate an enhanced vector sequence, and then the enhanced vector sequence is subjected to regression analysis through LSTM to generate a score for the mobile application service.

[0038] Through the embodiment of the present disclosure, since the traffic data generated by the mobile application business is cut into multiple traffic data segments, in order to be able to use the mobile application business scoring model to score the mobile application business after cutting, the traffic data segments are feature serialized, and numerical features of the serialized traffic data segments are generated. The serialized features are input into the mobile application business scoring model, and then the score is output. The scoring method in the embodiment of the present disclosure has no requirements on the transmission method of the traffic data, and both plain text transmission and encrypted transmission are possible. For encrypted transmitted data, the score of the mobile application business can be obtained without decoding. Therefore, it can solve the problem that the scoring method for mobile application business in the related art is not universal due to the requirement that the traffic data is transmitted in plain text, thereby achieving the effect of a wider range of use of the method for analyzing the sequence by cutting the traffic data generated by the mobile application business and further processing it to generate a sequence representing the mobile application business.

[0039] In one embodiment, when scoring a mobile application service, regardless of whether the traffic data generated by the mobile application service is encrypted, the traffic data can be segmented and feature-serialized, and then a mobile application scoring model can be used to output a score for the mobile application service. FIG3 is a flow chart of a mobile application service scoring method according to another embodiment of the present disclosure. As shown in FIG3 , the process includes the following steps:

[0040] Step S301: Loading a mobile application service rating model on a network device.

[0041] Step S302: Detect traffic data of the target mobile application service.

[0042] Specifically, the network device detects the traffic data of the target mobile application service, such as the traffic data of a specific video application, through means such as Deep Packet Inspection (DPI).

[0043] Step S303: Process the traffic data into serialized numerical features.

[0044] Specifically, the traffic data generated by the application is processed into serialized numerical features through the same processing method as the feature serialization in step S202 in the scoring method for mobile application services shown in FIG. 2 .

[0045] Step S304: input the serialized numerical features into the mobile application service rating model, and output the rating of the mobile application service.

[0046] Before step S304 of the embodiment of the present disclosure, it also includes: cutting the captured files in the original sample data according to preset time intervals, and performing feature serialization to determine the serialized numerical features of the original sample data, wherein the original sample data is collected and obtained from mobile application business data; scoring the metadata in the original sample data according to the scoring formula of the mobile application business to determine the label of the original sample data; training a neural network model for processing sequence data according to the serialized numerical features and the label of the original sample data to determine the mobile application business scoring model.

[0047] In one embodiment, the original sample data includes the captured packet files and metadata. The captured packet files are essentially the same as the traffic data in step S202 above, both of which are messages generated by the mobile application business. The difference is that the traffic data in step S202 is the messages generated in real time by the mobile application business when the mobile application business scoring model is actually used for scoring, while the captured packet files are stored as captured packet files during the stage of determining the mobile application business scoring model in order to provide samples required for model training. Figure 4 is a flow chart of the data collection method according to an embodiment of the present disclosure. As shown in Figure 4, the process includes the following steps:

[0048] Step S401: Open the packet capture software.

[0049] Specifically, open and run the packet capture software to capture the messages generated during application execution as part of the raw sample data. Figure 5 is a schematic diagram of the raw sample data structure according to an embodiment of the present disclosure. Along the timeline of application use, there is a metadata file that stores various metadata generated during the intermediate process, including resolution 720P, bit rate 2Mbit / s, and freeze time 20s; there is also a packet capture file that stores the message data generated during application operation.

[0050] Step S402: Open a specific mobile application.

[0051] Specifically, open the mobile application that needs to be analyzed and start the operation.

[0052] Step S403: applying different network conditions.

[0053] Specifically, during the use of the application, different network conditions can be imposed on it, such as good and poor signals, heavy or light device loads, etc. The purpose of this is to cover different scenarios as much as possible so that the trained model has better generalization capabilities.

[0054] Step S404: Record various metadata during the operation.

[0055] Specifically, various metadata generated during the use of the application are recorded, such as the bit rate, resolution, and freeze duration of video applications, as basic elements for making labels.

[0056] Step S405: store the captured packet file and the acquired metadata.

[0057] Specifically, after viewing is completed, the packet capture file obtained in step S401 and the annotation file obtained in step S404 are saved as a set of original sample data.

[0058] In one embodiment, there may be multiple types of scoring formulas for mobile application services, including assigning different weights to metadata related to user experience in mobile application services, and designing and generating weighted quality scoring formulas for mobile application services. Taking the weighted quality scoring formula as an example, FIG6 is a flow chart of a scoring formula design method according to an embodiment of the present disclosure. As shown in FIG6, the process includes the following steps:

[0059] Step S601: Determine evaluation dimensions for a specific mobile application service.

[0060] Specifically, taking the video service as an example, "clarity" and "smoothness" are selected as evaluation dimensions.

[0061] Step S602: convert the evaluation dimension into metadata.

[0062] Specifically, the evaluation dimension selected in step S501 is converted into metadata that can actually be obtained during sample collection. For example, "clarity" can be mapped to "resolution" and "bitrate," two metadata related to user experience, while "smoothness" can be mapped to "freeze duration," another metadata related to user experience.

[0063] Step S603: Map each metadata to a scoring interval.

[0064] Specifically, each metadata can be considered an independent factor that affects the score. In order to quantify the score, each metadata needs to be mapped to a numerical range in a different way. Taking the percentage system as an example, for the metadata "resolution", the following mapping method can be planned:

[0065] Greater than 1080p, corresponding to 100 points;

[0066] Greater than 720p and less than or equal to 1080p, proportionally corresponding to the 90-100 score range;

[0067] Greater than 480p and less than or equal to 720p, the score is proportionally corresponding to the range of 70-90;

[0068] Greater than 360p and less than or equal to 480p, proportionally correspond to the 50-70 point range;

[0069] Less than 360p, proportionally corresponds to the range of 0-50 points.

[0070] The above is just an example and can be adjusted based on experience and actual needs.

[0071] As for other metadata, they can also be mapped to a numerical range of 0-100 according to their own characteristics.

[0072] Step S604: assign a weight to each metadata.

[0073] Specifically, the score is determined by the comprehensive evaluation of each metadata, but the importance of different metadata varies. Therefore, it is necessary to assign appropriate weights to each metadata according to actual needs. For example, the weight of resolution can be set to 0.3, the weight of bit rate can be set to 0.3, and the weight of the number of freezes can be set to 0.4.

[0074] Step S605: Determine a scoring formula for the mobile application service.

[0075] Specifically, the final formula for calculating the score should be the weighted value of each metadata. The specific process is: first extract the metadata, then obtain the score corresponding to each metadata according to the mapping method in 603, and finally sum these scores with the weights assigned in 604.

[0076] It should be noted that the method shown in FIG6 is applied to the video service, and similar methods are also used in designing scoring formulas for other services.

[0077] In one embodiment, after determining the scoring formula for the mobile application service, the metadata in the original sample data is scored according to the determined scoring formula to determine the label of the original sample data. Taking the scoring formula as a weighted quality scoring formula as an example, FIG7 is a flow chart of the label calculation method according to an embodiment of the present disclosure. As shown in FIG7, the process includes the following steps:

[0078] Step S701: extract all metadata from a copy of original sample data.

[0079] Specifically, all metadata in a piece of original sample data, such as resolution, bit rate, freeze duration, etc., are extracted. These metadata are the basic units for generating the score value.

[0080] Step S702: Map each metadata into a score.

[0081] Specifically, each raw data is independently mapped to a specific numerical value, and each numerical value is an independent consideration of the final score.

[0082] Step S703: Calculate the label value by weighted summation.

[0083] Specifically, these values ​​are weighted and summed to obtain the final score value of the original sample data, which is also the label value of the original sample data.

[0084] It should be noted that the method in the embodiment of the present disclosure needs to be applied to every piece of original data.

[0085] In an exemplary embodiment, the packet capture file in the original sample data is cut according to a preset time interval and feature serialization is performed, including: cutting the packet capture file in the original sample data into multiple traffic data segments according to the preset time interval, and determining the basic features of the traffic in each traffic data segment; representing each basic feature in each traffic data segment in vector form, splicing all the basic features in vector form in each traffic data segment, determining the original vector of each traffic data segment, and then determining the original vector sequence corresponding to the packet capture file.

[0086] In an exemplary embodiment, determining the basic characteristics of the traffic in each traffic data segment includes calculating the uplink traffic, downlink traffic, the number of uplink packets, or the number of downlink packets of the traffic in each traffic data segment.

[0087] In one embodiment, the process of determining the original vector sequence corresponding to the captured packet file is shown in FIG8 , and the process includes the following steps:

[0088] Step S801: extract a packet capture file from the original sample data.

[0089] Step S802: Determine the sequence time interval.

[0090] Specifically, since the packet capture file is treated as a sequence, it must first be segmented to map the data to different positions in the sequence. Here, the packets are segmented by time, for example, using a 1-second interval (which can be adjusted based on the final result). The packet capture file can then be viewed as a sequence arranged in chronological order, with 1-second intervals as the basic unit.

[0091] Step S803: quantize each feature in each interval.

[0092] Specifically, each interval contains several basic features, such as the number of uplink packets, the number of downlink packets, uplink traffic, and downlink traffic. To facilitate algorithmic processing, each feature type needs to be represented as a vector. Taking the number of uplink packets as an example, we first need to group the packet lengths based on the overall distribution of the messages. The packet length refers to the length of the data packet, that is, the number of bytes contained in the data message. For example, the packet length can be divided into three intervals: [0, 500], [500, 1200], and [1200, 1500]. Assuming that in a specific interval, the corresponding number of packets in these three intervals is 9, 10, and 15, respectively, then the uplink packet length feature for this interval can be represented by the vector (9, 10, 15). The remaining features can also be converted into specific vectors using this method.

[0093] Step S804: All features in each interval are spliced ​​together.

[0094] Specifically, since for the timing algorithm, each basic unit often corresponds to only a single vector, and what is obtained in step S803 is the vector representation of each feature in a specific interval, in order to losslessly represent this information with a single vector, it is necessary to splice the vectors of all features to form the final representation of the interval feature. For example, if only the two features of the number of uplink packets and the number of downlink packets of an interval are considered, the vector representations of these two features are: (9, 10, 15) and (20, 110, 8, 50) respectively, then the spliced ​​vector is: (9, 10, 15, 20, 110, 80, 50). Figure 9 is a schematic diagram of feature vector splicing according to an embodiment of the present disclosure. As shown in Figure 9, the number of uplink packets, the number of downlink packets, the uplink traffic, and the downlink traffic are spliced ​​to generate the feature vector t at each moment. Figure 10 is a schematic diagram of the serialized packet capture file features according to an embodiment of the present disclosure. After processing each interval, the entire packet capture file is converted into a vector sequence, and the basic unit of the sequence is a spliced ​​vector.

[0095] In an exemplary embodiment, a neural network model for processing sequence data is trained based on the serialized numerical features and the labels of the original sample data to determine a mobile application business scoring model, including: determining an enhanced vector sequence by encoding the original vector sequence, inputting the enhanced vector sequence and the labels of the original sample data into the neural network model, training the neural network model, and determining the mobile application business scoring model.

[0096] In an exemplary embodiment, encoding the original vector sequence to determine the enhanced vector sequence includes: encoding the original vector sequence through a self-attention model to determine the enhanced vector sequence.

[0097] In an exemplary embodiment, the enhanced vector sequence and the label of the original sample data are input into the neural network model, and the neural network model is trained, including: inputting the enhanced vector sequence and the label of the original sample data into a long short-term memory network (LSTM), and the LSTM trains the neural network model by performing label regression analysis on the original sample data.

[0098] In one embodiment, as shown in FIG11 , the process of determining an enhanced vector sequence and training a neural network model to generate a mobile service scoring model includes the following steps:

[0099] Step S1101: Generate an enhanced vector sequence.

[0100] It should be noted that since the goal of this embodiment is to solve the problem of mobile application business rating by borrowing the concept of sequence analysis, it is necessary to introduce a network structure that excels at handling sequence problems. Compared with the traditional use of recurrent neural network (RNN) family algorithms such as LSTM to build a network, the network structure described in this embodiment also introduces a self-attention model encoder (Transformer Encoder) as part of the network structure.

[0101] Specifically, as shown in Figure 12, each enhanced vector is represented by all other vectors in the sequence (taking r2 as an example in the figure). The Transformer Encoder uses the self-attention mechanism of the self-attention model to further encode the original vector feature t at any time to obtain the corresponding enhanced vector r2, thereby being able to obtain more correlations between elements within the sequence (global attention).

[0102] Step S1102: Send the enhanced vector sequence and the corresponding label to LSTM for training.

[0103] Specifically, after obtaining the enhanced vector sequence, as shown in Figure 13, the enhanced vector sequence r and the corresponding label h are fed into the LSTM network used for sequence processing. This network analyzes the label value corresponding to the feature by superimposing the information of each enhanced vector, that is, the business score. After the LSTM training is completed, the mobile application business scoring model is immediately determined.

[0104] Step S1103: Achieve the expected effect through parameter debugging.

[0105] Specifically, the corresponding parameters are adjusted to improve the model's performance. These parameters can be algorithm hyperparameters, such as the number of Transformer Encoder layers or the number of neurons in an LSTM, or feature-related parameters, such as the size of the interval or the dimension of the vector feature.

[0106] In an exemplary embodiment, after training the neural network model for processing sequence data based on the serialized numerical features and the labels of the original sample data, it also includes at least one of the following: adjusting the number of layers of the encoder of the self-attention model; adjusting the number of neurons of the LSTM.

[0107] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it 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 the present disclosure is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), including a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present disclosure.

[0108] This embodiment also provides a scoring device for mobile application services. The device is configured to implement the above-mentioned embodiments and preferred implementations, and details already described will not be repeated. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0109] Figure 14 is a structural block diagram of a scoring device for mobile application services according to an embodiment of the present disclosure. As shown in Figure 14, the device includes: a scoring design module, a data acquisition module, a label calculation module, a feature calculation module, a model training module, and an online operation module.

[0110] Scoring Design Module: This module develops scoring criteria. For example, from the perspective of user experience, it is necessary to consider which factors (such as smoothness and clarity) are included in the scoring of an application. These factors need to be mapped to the metadata collected during the collection process (such as resolution, bitrate, and duration of freezes). The scoring criteria are then quantified in the form of formulas to guide the subsequent labeling process.

[0111] Data Collection Module: This module collects raw data generated during application execution to support subsequent training. By running specific applications in different network environments, it generates raw sample data representing a variety of user experiences. Each set of raw data includes a packet capture file and metadata such as resolution, bitrate, and duration of freezes.

[0112] Label Calculation Module: This module calculates a label value for each sample, representing the score corresponding to the application usage. By substituting the metadata acquired by the Data Collection Module into the scoring formula provided by the Scoring Design Module, we can obtain a score for each sample, also known as the label value.

[0113] Feature calculation module: The function of this module is to process the packet capture files in the original samples obtained in the data acquisition module and make them into serialized numerical data as the feature input of the algorithm.

[0114] Model training module: This module uses a Transformer + LSTM network for training. The algorithm inputs are labels generated by the label calculation module and features generated by the feature calculation module. The algorithm is trained using this data, enabling the generated model to score the user experience quality based on the messages generated by the application.

[0115] Online operation module: This module first loads the trained model file on the network device. Then, when the user uses a specific application, it calculates the corresponding serialization features based on the messages generated by the application. Finally, it combines the features with the trained model file to score the application's user experience quality.

[0116] It should be noted that the above modules can be implemented through software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0117] The disclosed embodiments address the problem that conventional network application scoring methods can only partially address scenarios where traffic is unencrypted, resulting in limited applicability and a lack of universal applicability. The disclosed embodiments cover a wide range of scenarios and eliminate the need for complex feature engineering, resulting in high efficiency and versatility. They can be used to quantify the quality of online services, enabling operators to understand the real user experience when using applications, thereby ensuring network quality.

[0118] An embodiment of the present disclosure further provides a computer-readable storage medium, in which a computer program is stored. The computer program is configured to execute the steps of any one of the above method embodiments when run.

[0119] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0120] An embodiment of the present disclosure further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0121] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0122] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail here.

[0123] Obviously, those skilled in the art should understand that the modules or steps of the present disclosure described above can be implemented using a general-purpose computing device, they can be concentrated on a single computing device, or distributed across a network composed of multiple computing devices, they can be implemented using program code executable by the computing device, and thus, they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be performed in a different order than herein, or they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. Thus, the present disclosure is not limited to any particular combination of hardware and software.

[0124] The foregoing description is merely a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. Those skilled in the art will readily appreciate that various modifications and variations of the present disclosure are possible. Any modifications, equivalent substitutions, or improvements made within the principles of the present disclosure shall be included within the scope of protection of the present disclosure.

Claims

1. A scoring method for a mobile application business, comprising: Cutting the traffic data generated by the mobile application service into a plurality of traffic data segments according to a preset time interval, and performing feature serialization to determine the serialized numerical features of each of the traffic data segments; The serialized numerical features of the traffic data segments are input into a mobile application service scoring model, and a score of the mobile application service is output.

2. The method according to claim 1, wherein Before inputting the serialized numerical features of the traffic data segments into the mobile application business scoring model, the method further includes: Slicing the captured packet files in the original sample data according to the preset time intervals and performing feature serialization to determine the serialized numerical features of the original sample data, wherein the original sample data is collected and obtained from mobile application business data; Scoring the metadata in the original sample data according to a scoring formula for the mobile application business to determine a label for the original sample data; A neural network model for processing sequence data is trained according to the serialized numerical features and the labels of the original sample data to determine a mobile application business scoring model.

3. The method according to claim 2, wherein: The packet capture files in the original sample data are cut according to the preset time intervals and feature serialized, including: Cutting the packet capture file in the original sample data into multiple traffic data segments according to the preset time interval, and determining the basic characteristics of the traffic in each traffic data segment; Each basic feature in each traffic data segment is represented in vector form, all basic features in vector form in each traffic data segment are spliced ​​together, the original vector of each traffic data segment is determined, and then the original vector sequence corresponding to the captured packet file is determined.

4. The method according to claim 3, wherein: The step of training a neural network model for processing sequence data according to the serialized numerical features and the labels of the original sample data to determine a mobile application business scoring model includes: An enhanced vector sequence is determined by encoding the original vector sequence, the enhanced vector sequence and the label of the original sample data are input into the neural network model, the neural network model is trained, and a mobile application business scoring model is determined.

5. The method according to claim 4, wherein The step of determining the enhanced vector sequence by encoding the original vector sequence includes: The enhanced vector sequence is determined by encoding the original vector sequence through a self-attention model.

6. The method according to claim 4, wherein: Inputting the enhanced vector sequence and the label of the original sample data into the neural network model to train the neural network model includes: The enhanced vector sequence and the label of the original sample data are input into a long short-term memory network LSTM, and the LSTM trains the neural network model by performing label regression analysis on the original sample data.

7. The method according to claim 2, wherein: After training the neural network model for processing sequence data according to the serialized numerical features and the labels of the original sample data, the method further includes at least one of the following: Adjust the number of layers of the encoder of the self-attention model; Adjust the number of neurons in LSTM.

8. The method according to claim 3, wherein: Determining the basic characteristics of the traffic in each traffic data segment includes: Calculate the uplink traffic, downlink traffic, number of uplink packets or number of downlink packets of the traffic in each traffic data segment.

9. A computer-readable storage medium having a computer program stored therein, wherein: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images