Network performance index data distribution prediction method and device, equipment and storage medium

Abstract

The embodiment of the application provides a network performance index data distribution prediction method and device, electronic equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring network performance index data; performing forward noise adding processing on the network performance index data by using a preset diffusion model to obtain network performance index noise data; acquiring a de-noising function of the preset diffusion model according to the network performance index data and the network performance index noise data; performing reverse de-noising on the network performance index noise data according to the de-noising function to obtain network performance index sample data; and calculating the data distribution of the network performance index sample data by using the preset diffusion model to obtain a network performance index distribution diagram. The embodiment of the application can improve the accuracy of network performance index data distribution prediction.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method and apparatus for predicting the distribution of network performance index data, an electronic device, and a storage medium. Background Technology

[0002] Currently, most methods for predicting network performance indicators in 5G networks employ deterministic prediction models based on multi-layer neural networks. The prediction results obtained from training these models represent a deterministic one-to-one functional relationship. However, in real-world scenarios, the relationship between network characteristics and network performance indicators exhibits a degree of uncertainty. This uncertainty arises from the dynamic changes in real, complex wireless networks and the complex and ever-changing environmental factors. Deterministic models fail to capture this uncertainty, leading to inaccurate predictions of network performance indicator data distribution. Therefore, improving the accuracy of network performance indicator data distribution predictions has become an urgent technical challenge. Summary of the Invention

[0003] The main objective of this application is to provide a method, apparatus, electronic device, and storage medium for predicting the distribution of network performance index data, aiming to improve the accuracy of predicting the distribution of network performance index data.

[0004] To achieve the above objectives, a first aspect of this application proposes a method for predicting the distribution of network performance index data, the method comprising:

[0005] Obtain network performance metrics data;

[0006] The network performance index data is subjected to forward noise processing using a preset diffusion model to obtain network performance index noise data.

[0007] Based on the network performance index data and the network performance index noise data, obtain the denoising function of the preset diffusion model;

[0008] Based on the denoising function, the network performance index noise data is denoised in reverse to obtain network performance index sample data.

[0009] The data distribution of the network performance index sample data is calculated using the preset diffusion model to obtain the network performance index distribution map.

[0010] In some embodiments, obtaining the denoising function of the preset diffusion model based on the network performance index data and the network performance index noise data includes:

[0011] The time length for adding noise to the network performance index data to the network performance index noise data in the preset diffusion model is obtained.

[0012] The number of noise addition steps from the network performance index data to the network performance index noise data is calculated based on the preset time step and the noise addition time length, wherein the preset time step is the time length for performing one noise addition step on the network performance index data.

[0013] Perform a Gaussian Fourier transform on the number of noise-adding steps to obtain the noise-adding step feature data;

[0014] The noise step count feature data and the network performance index data are combined to obtain feature combination data;

[0015] The feature combination data and the network performance index noise data are convolved to obtain the target feature convolution data;

[0016] The target feature convolution data is deconvolutionally processed to obtain target feature deconvolution data;

[0017] The target features are deconvolved and linearly mapped to obtain a denoising function.

[0018] In some embodiments, the step of performing convolution processing on the feature combination data and the network performance indicator noise data to obtain target feature convolution data includes:

[0019] The feature combination data is added to the network performance index noise data to obtain the first feature merged data;

[0020] Perform a convolution process on the first feature merged data to obtain feature convolution data;

[0021] The feature convolutional data and the feature combination data are added together to obtain the second feature merged data;

[0022] The second feature merged data is subjected to a second convolution process to obtain the target feature convolution data.

[0023] In some embodiments, the step of performing deconvolution processing on the target feature convolutional data to obtain target feature deconvolutional data includes:

[0024] The target feature convolutional data and the feature combination data are added together to obtain the third feature merged data;

[0025] Perform a deconvolution process on the third feature merged data to obtain feature deconvolution data;

[0026] The deconvolutional data of the features is concatenated with the merged data of the second features to obtain the concatenated data of the first features.

[0027] The first feature concatenation data and the feature combination data are added together to obtain the fourth feature merged data;

[0028] The fourth feature merged data is subjected to a second deconvolution process to obtain the target feature deconvolution data.

[0029] In some embodiments, the step of performing a linear mapping on the deconvolution of the target features to obtain a denoising function includes:

[0030] The target feature deconvolution data is concatenated with the first feature merged data to obtain the second feature concatenated data.

[0031] The second feature concatenation data and the feature combination data are added together to obtain the fifth feature merged data;

[0032] The fifth feature merged data is subjected to three convolution processes to obtain feature merged convolution data;

[0033] The feature-merged convolutional data is subjected to feature mapping to obtain the denoising function of the preset diffusion model.

[0034] In some embodiments, the preset diffusion model includes a noise adjustment function, and the step of using the preset diffusion model to perform forward noise processing on the network performance index data to obtain network performance index noise data includes:

[0035] Random Gaussian noise data is generated using the preset noise adjustment function, wherein the preset noise adjustment function is a function built into the preset diffusion model to simulate the noise contained in the communication network channel under real-world conditions;

[0036] The random Gaussian noise data and the network performance index data are merged to obtain the network performance index noise data.

[0037] In some embodiments, before obtaining the denoising function of the preset diffusion model based on the network performance index data and the network performance index noise data, the method further includes:

[0038] A two-dimensional real-axis distribution plot of the network performance index noise data is drawn to obtain the network performance index noise data distribution plot;

[0039] Find the peak value of the network performance indicator noise data from the network performance indicator noise data distribution map to obtain the network performance indicator noise peak data;

[0040] Based on the network performance indicator noise peak data, the network performance indicator noise data in the network performance indicator noise data distribution map is partitioned to obtain the data interval to the left of the peak and the data interval to the right of the peak.

[0041] Calculate the mean of the data interval to the left of the peak to obtain the mean of the data on the left.

[0042] Calculate the sum of the data interval to the left of the peak value to obtain the sum of the data on the left.

[0043] Calculate the mean of the data interval to the right of the peak to obtain the mean of the data on the right.

[0044] Calculate the sum of the data interval to the right of the peak value to obtain the sum of the data on the right.

[0045] If the mean of the data on the left side is different from the mean of the data on the right side, or if the sum of the data on the left side is different from the sum of the data on the right side, forward noise processing is performed on the network performance index noise data to obtain new network performance index noise data.

[0046] If the mean of the data on the left side is the same as the mean of the data on the right side, and the sum of the data on the left side is the same as the sum of the data on the right side, the denoising function of the preset diffusion model is obtained based on the network performance index data and the network performance index noise data.

[0047] To achieve the above objectives, a second aspect of this application provides a network performance index data distribution prediction device, the device comprising:

[0048] The data acquisition module is used to acquire network performance indicator data;

[0049] The data noise-adding module is used to perform forward noise-adding processing on the network performance index data using a preset diffusion model to obtain network performance index noise data.

[0050] The sample data acquisition module is used to obtain the denoising function of the preset diffusion model based on the network performance index data and the network performance index noise data, and to perform reverse denoising on the network performance index noise data based on the denoising function to obtain network performance index sample data.

[0051] The data distribution prediction module is used to calculate the data distribution of the network performance index sample data using the preset diffusion model, and obtain the network performance index distribution map.

[0052] To achieve the above objectives, a third aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect.

[0053] To achieve the above objectives, a fourth aspect of the present application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect.

[0054] The network performance index data distribution prediction method, apparatus, electronic device, and storage medium proposed in this application obtain network performance index sample data by searching for network performance index data, using a preset diffusion model to perform forward noise addition and reverse noise removal on the network performance index data, and then generating a network performance index distribution map based on the network performance index sample data, thereby filling the gaps in network performance index data and improving the accuracy of network performance index data distribution prediction. Attached Figure Description

[0055] Figure 1 This is a flowchart of the network performance index data distribution prediction method provided in the embodiments of this application;

[0056] Figure 2 yes Figure 1 The flowchart of step S102 in the document;

[0057] Figure 3 yes Figure 1 A flowchart of the steps preceding step S103;

[0058] Figure 4 yes Figure 1 The flowchart of step S103 in the process;

[0059] Figure 5 yes Figure 3 The flowchart of step S405 in the document;

[0060] Figure 6 yes Figure 3 The flowchart of step S406 in the document;

[0061] Figure 7 yes Figure 3 The flowchart of step S407 in the document;

[0062] Figure 8 This is a schematic diagram of the network performance index data distribution prediction device provided in the embodiments of this application;

[0063] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0065] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0067] First, let's analyze some of the terms used in this application:

[0068] Artificial intelligence (AI) is a new branch of computer science that studies, develops, and applies theories, methods, technologies, and systems to simulate, extend, and expand human intelligence. It aims to understand the essence of intelligence and produce intelligent machines that can react in a way similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. AI can simulate the information processes of human consciousness and thought. Furthermore, AI utilizes digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceiving the environment, acquiring knowledge, and using that knowledge to achieve optimal results.

[0069] Currently, most methods for predicting network performance indicators in 5G networks employ deterministic prediction models based on multi-layer neural networks. The prediction results obtained from training these models represent a deterministic one-to-one functional relationship. However, in real-world scenarios, the relationship between network characteristics and network performance indicators exhibits a degree of uncertainty. This uncertainty arises from the dynamic changes in real, complex wireless networks and the complex and ever-changing environmental factors. Deterministic models fail to capture this uncertainty, leading to inaccurate predictions of network performance indicator data distribution. Therefore, improving the accuracy of network performance indicator data distribution predictions has become an urgent technical challenge.

[0070] Based on this, embodiments of this application provide a method and apparatus for predicting the distribution of network performance index data, an electronic device, and a storage medium, aiming to improve the diversity of lighting in smart mirror cabinets.

[0071] The network performance index data distribution prediction method, apparatus, electronic device, and storage medium provided in this application are specifically described through the following embodiments. First, the network performance index data distribution prediction method in this application is described.

[0072] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0073] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0074] The network performance indicator data distribution prediction method provided in this application relates to the field of artificial intelligence technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the network performance indicator data distribution prediction method, but is not limited to the above forms.

[0075] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0076] Figure 1 This is an optional flowchart of the network performance index data distribution prediction method provided in the embodiments of this application. Figure 1 The method may include, but is not limited to, steps S101 to S105.

[0077] Step S101: Obtain network performance indicator data;

[0078] Step S102: Use a preset diffusion model to perform forward noise processing on the network performance index data to obtain network performance index noise data;

[0079] Step S103: Obtain the denoising function of the preset diffusion model based on the network performance index data and the network performance index noise data.

[0080] Step S104: Based on the denoising function, reverse denoising is performed on the network performance index noise data to obtain network performance index sample data.

[0081] Step S105: Calculate the data distribution of the network performance index sample data using a preset diffusion model to obtain the network performance index distribution map.

[0082] Based on steps S101 to S105 shown in the embodiments of this application, this application provides a method, apparatus, electronic device, and storage medium for predicting the distribution of network performance index data, aiming to improve the accuracy of network performance index data distribution prediction. This application obtains network performance index data by searching for relevant data from network devices. After obtaining the network performance index data, a preset diffusion model is used to forward-noise the network performance index data to obtain network performance index noise data, simulating the impact of various random factors and interference signals on network devices in network communication. Further, if the network performance index noise data has a normal distribution characteristic, a denoising function of the preset diffusion model is obtained based on the network performance index data and the network performance index noise data. After obtaining the denoising function, the denoising function is used to reverse-denoise the network performance index noise data to obtain network performance index sample data. Finally, the preset diffusion model is used to calculate the data distribution of the network performance index sample data to obtain a network performance index distribution map. This application improves the accuracy of network performance index data distribution prediction by increasing the number of network performance index data samples.

[0083] In one application scenario, taking network device spectral efficiency as an example, spectral efficiency data is obtained from a database storing network feature data in the communication network. Further, a pre-defined diffusion model is used to add forward noise to the spectral efficiency, resulting in spectral efficiency noise data containing noise. This simulates the impact of various random factors and interference signals in network communication on the network device's spectral efficiency. Further, it is detected whether the spectral efficiency noise data has a normal distribution characteristic. If the spectral efficiency noise data has a normal distribution characteristic, a denoising function of the pre-defined diffusion model is calculated based on the spectral efficiency data and the spectral efficiency noise data. Based on the denoising function, the spectral efficiency noise data is then reverse-denoised to obtain spectral efficiency sample data. Finally, the data distribution of the spectral efficiency sample data is calculated using the pre-defined diffusion model to obtain a predicted distribution map of the network spectral efficiency.

[0084] In step S101 of some embodiments, network performance index data refers to index data that can reflect the performance of wireless network devices, such as signal receiving power and spectral efficiency. It should be noted that network performance index data can be obtained by searching relevant parameters of wireless network devices.

[0085] In step S102 of some embodiments, the preset diffusion model can be a DDPM (Diffusion Models) diffusion generation model, used to learn and generate a probability distribution from given network performance index data, thereby generating new network performance index sample data. Further, this invention typically uses a noise adjustment function in the preset diffusion model to add noise to the network performance index data. It should be noted that the random noise generated by the noise adjustment function can be Gaussian noise with a normal distribution, and the added random noise should become increasingly larger as the number of noise addition steps increases.

[0086] Please see Figure 2 In some embodiments, step S102 may include, but is not limited to, steps S201 to S202:

[0087] Step S201: Generate random Gaussian noise data using a preset noise adjustment function;

[0088] Step S202: Merge the random Gaussian noise data with the network performance index data to obtain network performance index noise data.

[0089] In step S201 of some embodiments, the preset noise adjustment function is a function built into the preset diffusion model to simulate the noise contained in the communication network channel under real environmental conditions. Random Gaussian noise is generated by using the preset noise adjustment function built into the preset diffusion model, wherein the random Gaussian noise is noise data that follows a Gaussian normal distribution.

[0090] In step S202 of some embodiments, network performance indicator noise data is obtained by splicing random Gaussian noise into the network performance indicator data.

[0091] Steps S201 to S202 as shown in the embodiments of this application generate different random Gaussian noises through a preset noise adjustment function in a preset diffusion model, and then concatenate the random Gaussian noises with network performance index data to obtain network performance index noise data. This simulates the impact of various random factors and interference signals on network devices in network communication, making the prediction results of network performance index data distribution more accurate.

[0092] In step S103 of some embodiments, the denoising function refers to a function that removes noise from the data, and is typically used in the reverse denoising process in a preset diffusion model.

[0093] Please see Figure 3 In some embodiments, steps S301 to S309 may be included before step S103:

[0094] Step S301: Draw a two-dimensional real axis distribution diagram of the network performance index noise data to obtain the network performance index noise data distribution diagram;

[0095] Step S302: Find the peak value of the network performance indicator noise data from the network performance indicator noise data distribution map to obtain the network performance indicator noise peak data.

[0096] Step S303: Based on the network performance indicator noise peak data, partition the network performance indicator noise data in the network performance indicator noise data distribution map to obtain the data interval to the left of the peak and the data interval to the right of the peak.

[0097] Step S304: Calculate the mean of the data interval to the left of the peak to obtain the mean of the data on the left.

[0098] Step S305: Calculate the sum of the data interval to the left of the peak to obtain the sum of the data to the left.

[0099] Step S306: Calculate the mean of the data interval to the right of the peak to obtain the mean of the data on the right.

[0100] Step S307: Calculate the sum of the data interval to the right of the peak to obtain the sum of the data on the right.

[0101] Step S308: If the mean of the data on the left side is different from the mean of the data on the right side, or the sum of the data on the left side is different from the sum of the data on the right side, perform forward noise processing on the network performance index noise data to obtain new network performance index noise data.

[0102] Step S309: If the mean of the data on the left side is the same as the mean of the data on the right side, and the sum of the data on the left side is the same as the sum of the data on the right side, obtain the denoising function of the preset diffusion model based on the network performance index data and the network performance index noise data.

[0103] In step S301 of some embodiments, the two-dimensional real-axis distribution plot refers to a two-dimensional spatial coordinate system with a horizontal axis and a vertical axis. Based on the network performance indicator noise data, a two-dimensional spatial coordinate system is drawn to obtain the network performance indicator noise data distribution plot.

[0104] In step S302 of some embodiments, the network performance indicator noise data distribution map is analyzed to obtain the peak value of the network performance indicator noise data in the network performance indicator noise data distribution map, that is, the network performance indicator noise peak data.

[0105] In step S303 of some embodiments, the network performance index noise data in the network performance index noise data distribution map is partitioned using the network performance index noise peak data as the boundary, to obtain the data interval to the left of the peak and the data interval to the right of the peak.

[0106] In steps S304 to S307 of some embodiments, the mean and sum of network performance index noise data within the data interval to the left of the peak are calculated to obtain the mean and sum of left-side data. Then, the mean and sum of network performance index noise data within the data interval to the right of the peak are calculated to obtain the mean and sum of right-side data.

[0107] In step S308 of some embodiments, it is determined whether the mean of the left data is the same as the mean of the right data, and whether the sum of the left data is the same as the sum of the right data. If the mean of the left data is different from the mean of the right data, or the sum of the left data is different from the sum of the right data, forward noise processing is performed on the network performance indicator noise data to obtain new network performance indicator noise data.

[0108] In step S309 of some embodiments, if the mean value of the data on the left side is the same as the mean value of the data on the right side, and the sum value of the data on the left side is the same as the sum value of the data on the right side, the denoising function of the preset diffusion model is obtained based on the network performance index data and the network performance index noise data.

[0109] Steps S301 to S309 of the embodiments of this application determine whether the network performance indicator noise data has a normal distribution characteristic by calculating the average value, sum value and peak value of the network performance indicator noise data interval, thus clarifying the data distribution characteristics of the network performance indicator noise data.

[0110] Please see Figure 4 In some embodiments, step S103 may include, but is not limited to, steps S401 to S407:

[0111] Step S401: Obtain the time length for adding noise from the network performance index data to the network performance index noise data in the preset diffusion model;

[0112] Step S402: Calculate the number of noise addition steps from network performance index data to network performance index noise data based on the preset time step and noise addition time length, wherein the preset time step is the time length for one noise addition step on the network performance index data.

[0113] Step S403: Perform Gaussian Fourier transform on the number of noise-adding steps to obtain the feature data of the number of noise-adding steps;

[0114] Step S404: Combine the noisy step count feature data with the network performance index data to obtain the feature combination data;

[0115] Step S405: Perform convolution processing on the feature combination data and network performance index noise data to obtain target feature convolution data;

[0116] Step S406: Perform deconvolution processing on the target feature convolution data to obtain target feature deconvolution data;

[0117] Step S407: Perform linear mapping on the deconvolution of the target features to obtain the denoising function.

[0118] In step S401 of some embodiments, the noise addition time length refers to the time length used between the conversion from network performance indicator data to network performance indicator noise data.

[0119] In step S402 of some embodiments, the number of preset time steps consumed by the noise addition time length is calculated to obtain the number of noise addition steps for converting network performance index data into network performance index noise data.

[0120] In step S403 of some embodiments, a one-dimensional noise-adding step matrix is ​​created based on the noise-adding step count, and the one-dimensional noise-adding step matrix is ​​subjected to Gaussian filtering to obtain noise-adding step feature data.

[0121] In steps S404 to S407 of some embodiments, the feature combination data of the combination of noise step number feature data and network performance index data and the network performance index noise data are input into a preset U-shaped mesh model to obtain a denoising function.

[0122] Steps S401 to S407 of the embodiments of this application involve calculating the number of noise-adding steps from network performance index data to network performance index noise data, obtaining noise-adding step feature data, and then using a preset U-shaped mesh model to calculate the denoising function of a preset diffusion model based on the network performance index data and the network performance index noise data, thereby improving the convergence speed of the preset diffusion model. It should be noted that the preset U-shaped mesh model refers to a convolutional neural network model based on a multilayer perceptron.

[0123] Please see Figure 5 In some embodiments, step S405 may include, but is not limited to, steps S501 to S504:

[0124] Step S501: Add the features of the feature combination data and the network performance index noise data to obtain the first feature merging data;

[0125] Step S502: Perform a convolution process on the first feature merged data to obtain feature convolution data;

[0126] Step S503: Add the features of the feature convolution data and the feature combination data to obtain the second feature merged data;

[0127] Step S504: Perform a second convolution process on the merged second feature data to obtain the convolutional data of the target feature.

[0128] In steps S501 to S504 of some embodiments, the feature combination data and the network performance index noise data are added together to obtain the first feature merged data. Then, the first feature merged data is convolved once to obtain the feature convolution data. Further, the feature convolution data and the feature combination data are added together to obtain the second feature merged data. Then, the second feature merged data is convolved a second time to obtain the target feature convolution data.

[0129] Steps S501 to S504 as shown in the embodiments of this application involve adding noise step feature data and network performance index data to the network performance index noise data, so that the preset U-shaped mesh model can learn the conditional distribution of the network performance index noise data when performing convolution processing on the network performance index noise data.

[0130] Please see Figure 6 In some embodiments, step S406 may also include, but is not limited to, steps S601 to S605:

[0131] Step S601: Add the target feature convolutional data and the feature combination data together to obtain the third feature merged data;

[0132] Step S602: Perform a deconvolution process on the merged data of the third feature to obtain the deconvolution data of the features;

[0133] Step S603: Concatenate the deconvolution data of the features with the merged data of the second features to obtain the concatenated data of the first features;

[0134] Step S604: Add the first feature concatenation data and the feature combination data together to obtain the fourth feature merged data;

[0135] Step S605: Perform a second deconvolution process on the merged data of the fourth feature to obtain the deconvolution data of the target feature.

[0136] In steps S601 to S605 of some embodiments, the target feature convolutional data is transformed into target feature deconvolutional data with a dimension twice larger than that of the network performance indicator noise data by performing two deconvolutions of different multiples on the target feature convolutional data.

[0137] Steps S601 to S605 as shown in the embodiments of this application, by performing two deconvolutions on the target feature convolution data at different multiples, can restore the low-dimensional data that appears after the convolution operation to high-dimensional data. Furthermore, the step of splicing the feature deconvolution data with the second feature merged data can recover some of the spatial dimensionality information loss caused by the convolution operation.

[0138] Please see Figure 7 In some embodiments, step S407 includes, but is not limited to, steps S701 to S704:

[0139] Step S701: Concatenate the deconvolution data of the target feature with the merged data of the first feature to obtain the concatenated data of the second feature;

[0140] Step S702: Add the second feature concatenation data and the feature combination data together to obtain the fifth feature merged data;

[0141] Step S703: Perform three convolution processes on the fifth feature merged data to obtain feature merged convolution data;

[0142] Step S704: Perform feature mapping on the feature-merged convolutional data to obtain the denoising function of the preset diffusion model.

[0143] In step S701 of some embodiments, the dimension of the first feature merging data is transformed into the dimension of the target feature deconvolution data, and then the target feature deconvolution data is concatenated with the first feature merging data to obtain the second feature concatenation data.

[0144] In step S702 of some embodiments, the dimension of the feature combination data is transformed into the dimension of the second feature concatenation data, and then the second feature concatenation data and the feature combination data are added together to obtain the fifth feature merged data.

[0145] In step S703 of some embodiments, the fifth feature merged data is subjected to three convolution processes to restore the dimension of the fifth feature merged data to the dimension when the network performance index noise data was input, thus obtaining feature merged convolution data.

[0146] In step S704 of some embodiments, the activation function in the U-shaped grid model is used to perform feature mapping on the feature-merged convolutional data to obtain a denoising function.

[0147] Steps S701 to S704 of the embodiments of this application involve performing operations such as feature concatenation, feature addition, and feature convolution on the deconvolutional data of the target features to restore the dimension of the deconvolutional data of the target features to the dimension of the network performance index noise data input. Then, the activation function in the U-shaped grid model is used to perform feature mapping on the feature merged convolutional data to obtain the denoising function of the preset diffusion model, thereby improving the convergence speed of the preset diffusion model.

[0148] In step S104 of some embodiments, after obtaining the denoising function, the network performance index noise data is reversed using a preset diffusion model based on the denoising function to obtain network performance index sample data, which fills the gap in the network performance index data and improves the fault tolerance rate of the network performance index data distribution prediction.

[0149] In step S105 of some embodiments, after obtaining the network performance index sample data, the data distribution of the network performance index sample data is calculated using a preset diffusion model to obtain a network performance index distribution map. It should be noted that the confidence interval of the network performance index distribution can be calculated based on the network performance index distribution map, and then the network performance index distribution can be predicted based on the confidence interval.

[0150] This application embodiment obtains network performance index data by searching relevant data of wireless network devices, uses a preset diffusion model to add noise to the network performance index data in the forward direction and denoise it in the reverse direction to obtain sample data of network performance indexes, and then generates a network performance index distribution map based on the sample data of network performance indexes, thereby filling the gaps in network performance index data and improving the accuracy of network performance index data distribution prediction.

[0151] Please see Figure 8 This application also provides a network performance indicator data distribution prediction device, which can implement the above-mentioned network performance indicator data distribution prediction method. The device includes:

[0152] Data acquisition module 801 is used to acquire network performance index data;

[0153] The data noise-adding module 802 is used to perform forward noise-adding processing on the network performance index data using a preset diffusion model to obtain network performance index noise data.

[0154] The sample data acquisition module 803 is used to obtain the denoising function of the preset diffusion model based on the network performance index data and the network performance index noise data, and to perform reverse denoising on the network performance index noise data based on the denoising function to obtain the network performance index sample data.

[0155] The data distribution prediction module 804 is used to calculate the data distribution of sample data of network performance indicators using a preset diffusion model, and obtain the network performance indicator distribution map.

[0156] The specific implementation of this network performance index data distribution prediction device is basically the same as the specific implementation of the network performance index data distribution prediction method described above, and will not be repeated here.

[0157] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned method for predicting the distribution of network performance indicators. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0158] Please see Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:

[0159] The processor 901 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0160] The memory 902 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 902 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 902 and is called and executed by the processor 901 using the network performance index data distribution prediction method of the embodiments of this application.

[0161] The input / output interface 903 is used to implement information input and output;

[0162] The communication interface 904 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0163] Bus 905 transmits information between various components of the device (e.g., processor 901, memory 902, input / output interface 903, and communication interface 904);

[0164] The processor 901, memory 902, input / output interface 903, and communication interface 904 are connected to each other within the device via bus 905.

[0165] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for predicting the distribution of network performance index data.

[0166] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0167] The network performance index data distribution prediction method, device, electronic device, and storage medium provided in this application obtain network performance index data by searching relevant data of wireless network devices, use a preset diffusion model to perform forward noise addition and reverse noise reduction on the network performance index data to obtain network performance index sample data, and then generate a network performance index distribution map based on the network performance index sample data, thereby filling the gaps in network performance index data and improving the accuracy of network performance index data distribution prediction.

[0168] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0169] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0170] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0171] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0172] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0173] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0174] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. The coupling or direct coupling or communication connection between the shown or discussed units may be through some interfaces, or indirect coupling or communication connection between the apparatus or units, and may be electrical, mechanical, or other forms.

[0175] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0176] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0177] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0178] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A method for predicting the distribution of network performance index data, characterized in that, The method includes: Obtain network performance metrics data; The network performance index data is subjected to forward noise processing using a preset diffusion model to obtain network performance index noise data. The time length for adding noise to the network performance index data to the network performance index noise data in the preset diffusion model is obtained. The number of noise addition steps from the network performance index data to the network performance index noise data is calculated based on the preset time step and the noise addition time length, wherein the preset time step is the time length for performing one noise addition step on the network performance index data. Perform a Gaussian Fourier transform on the number of noise-adding steps to obtain the noise-adding step feature data; The noise step count feature data and the network performance index data are combined to obtain feature combination data; The feature combination data and the network performance index noise data are convolved to obtain the target feature convolution data; The target feature convolution data is deconvolutionally processed to obtain target feature deconvolution data; The target features are deconvolved and linearly mapped to obtain a denoising function; Based on the denoising function, the network performance index noise data is denoised in reverse to obtain network performance index sample data. The data distribution of the network performance index sample data is calculated using the preset diffusion model to obtain the network performance index distribution map.

2. The method according to claim 1, characterized in that, The step of performing convolution processing on the feature combination data and the network performance index noise data to obtain target feature convolution data includes: The feature combination data is added to the network performance index noise data to obtain the first feature merged data; Perform a convolution process on the first feature merged data to obtain feature convolution data; The feature convolutional data and the feature combination data are added together to obtain the second feature merged data; The second feature merged data is subjected to a second convolution process to obtain the target feature convolution data.

3. The method according to claim 1, characterized in that, The step of performing deconvolution processing on the target feature convolutional data to obtain target feature deconvolutional data includes: The target feature convolutional data and the feature combination data are added together to obtain the third feature merged data; Perform a deconvolution process on the third feature merged data to obtain feature deconvolution data; The deconvolutional data of the features is concatenated with the merged data of the second features to obtain the concatenated data of the first features. The first feature concatenation data and the feature combination data are added together to obtain the fourth feature merged data; The fourth feature merged data is subjected to a second deconvolution process to obtain the target feature deconvolution data.

4. The method according to any one of claims 1 to 3, characterized in that, The step of performing a linear mapping on the deconvolution of the target features to obtain the denoising function includes: The target feature deconvolution data is concatenated with the first feature merged data to obtain the second feature concatenated data. The second feature concatenation data and the feature combination data are added together to obtain the fifth feature merged data; The fifth feature merged data is subjected to three convolution processes to obtain feature merged convolution data; The feature-merged convolutional data is subjected to feature mapping to obtain the denoising function of the preset diffusion model.

5. The method according to any one of claims 1 to 3, characterized in that, The preset diffusion model includes a noise adjustment function. The step of using the preset diffusion model to perform forward noise processing on the network performance index data to obtain network performance index noise data includes: Random Gaussian noise data is generated using the noise adjustment function. The random Gaussian noise data and the network performance index data are merged to obtain the network performance index noise data.

6. The method according to any one of claims 1 to 3, characterized in that, Before obtaining the denoising function of the preset diffusion model based on the network performance index data and the network performance index noise data, the method further includes: A two-dimensional real-axis distribution plot of the network performance index noise data is drawn to obtain the network performance index noise data distribution plot; Find the peak value of the network performance indicator noise data from the network performance indicator noise data distribution map to obtain the network performance indicator noise peak data; Based on the network performance indicator noise peak data, the network performance indicator noise data in the network performance indicator noise data distribution map is partitioned to obtain the data interval to the left of the peak and the data interval to the right of the peak. Calculate the mean of the data interval to the left of the peak to obtain the mean of the data on the left. Calculate the sum of the data interval to the left of the peak value to obtain the sum of the data on the left. Calculate the mean of the data interval to the right of the peak to obtain the mean of the data on the right. Calculate the sum of the data interval to the right of the peak value to obtain the sum of the data on the right. If the mean of the data on the left side is different from the mean of the data on the right side, or if the sum of the data on the left side is different from the sum of the data on the right side, forward noise processing is performed on the network performance index noise data to obtain new network performance index noise data. If the mean of the data on the left side is the same as the mean of the data on the right side, and the sum of the data on the left side is the same as the sum of the data on the right side, the denoising function of the preset diffusion model is obtained based on the network performance index data and the network performance index noise data.

7. A network performance index data distribution prediction device, applied to the network performance index data distribution prediction method according to any one of claims 1 to 6, characterized in that, The device includes: The data acquisition module is used to acquire network performance indicator data; The data noise-adding module is used to perform forward noise-adding processing on the network performance index data using a preset diffusion model to obtain network performance index noise data. The sample data acquisition module is used to obtain the denoising function of the preset diffusion model based on the network performance index data and the network performance index noise data, and to perform reverse denoising on the network performance index noise data based on the denoising function to obtain network performance index sample data. The data distribution prediction module is used to calculate the data distribution of the network performance index sample data using the preset diffusion model, and obtain the network performance index distribution map.

8. An electronic device, characterized in that, The electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the network performance index data distribution prediction method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the network performance index data distribution prediction method according to any one of claims 1 to 6.

Images