Rate prediction method and device based on ping service data

Abstract

An embodiment of the present invention discloses a rate prediction method and device based on ping service data. The method includes: inputting an indicator affecting the MCS value in the ping service data into an MCS prediction model to obtain an MCS prediction value for the ping service data, wherein the MCS prediction model is pre-trained based on corresponding indicator samples and MCS value labels in the actual rate data; obtaining a modulation ratio indicator for the ping service data based on the MCS prediction value; obtaining a prediction index item based on the indicator affecting the MCS value in the ping service data, the MCS prediction value, and the modulation ratio, and inputting the prediction index item into a rate prediction model to obtain the rate of the ping service data, wherein the rate prediction model is pre-trained based on corresponding prediction index item samples and rate conversion value labels in the actual rate data. The embodiment of the present invention can predict the actual rate of an operator through ping service data, and has the advantages of low rate testing cost and high testing efficiency for telecom operators.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a method, device, electronic device and storage medium for predicting a rate based on ping service data. Background Art

[0002] The existing method for obtaining the speed indicators of competing operators is to conduct manual traversal testing in a predetermined area, that is, manually use a test terminal to drive traversing the road to conduct FTP (File Transfer Protocol) download service tests to collect the speed of competing operators.

[0003] However, due to the high cost of competing operators' SIM cards (Subscriber Identity Modules), it's impossible to conduct a full-network, full-road speed test, making it impossible to effectively and timely monitor competing operators' speeds. Existing methods for acquiring competing operators' speed indicators generate a large amount of traffic during the test. With average speeds of hundreds of Mbps per grid area, test durations measured in hours, and the high cost of 5G traffic, a single test cycle based on 94 grids costs tens of thousands of yuan. This prohibitive testing cost makes it impossible to complete a full network traversal test, making it impossible to effectively and timely monitor competing operators' speed indicators. Furthermore, competing operators have already achieved 5G base station sharing, and existing methods for acquiring competing operators' speed indicators cannot test two operators simultaneously on the same vehicle. Otherwise, the speed would be halved when both operators occupy the same 5G base station. Therefore, the same area must be tested twice, resulting in low test efficiency. Summary of the Invention

[0004] In view of the problems existing in the prior art, embodiments of the present invention provide a method, device, electronic device, and storage medium for predicting a rate based on ping service data.

[0005] In a first aspect, an embodiment of the present invention provides a rate prediction method based on ping service data, comprising:

[0006] Inputting an indicator affecting the MCS value in the ping service data into an MCS prediction model to obtain an MCS prediction value of the ping service data, wherein the MCS prediction model is pre-trained based on corresponding indicator samples and MCS value labels in the real rate data;

[0007] Obtaining a modulation ratio indicator of the ping service data based on the MCS prediction value;

[0008] A prediction index item is obtained based on the indicator affecting the MCS value in the ping service data, the MCS prediction value and the modulation ratio, and the prediction index item is input into a rate prediction model to obtain the rate of the ping service data, wherein the rate prediction model is pre-trained based on the corresponding prediction index item samples and rate conversion value labels in the actual rate data.

[0009] Furthermore, before inputting the indicator affecting the MCS value in the ping service data into the MCS prediction model to obtain the MCS prediction value of the ping service data, the method includes:

[0010] Get real rate data;

[0011] Extracting an index affecting an MCS value and an MCS value from the true rate data;

[0012] The indicators affecting the MCS value and the MCS value extracted from the real rate data are used as data samples and labels respectively, and the MCS prediction model is trained based on the lasso linear regression algorithm.

[0013] Furthermore, the rate prediction method based on ping service data further includes:

[0014] Before using the index affecting the MCS value and the MCS value extracted from the true rate data as data samples and labels, respectively, a feature enhancement algorithm is used to perform feature enhancement on the index affecting the MCS value and the MCS value extracted from the true rate data;

[0015] Before inputting the indicators affecting the MCS value in the ping service data into the MCS prediction model to obtain the MCS prediction value of the ping service data, the feature enhancement algorithm is used to perform feature enhancement on the indicators affecting the MCS value in the ping service data.

[0016] Furthermore, obtaining a modulation ratio index of ping service data based on the MCS prediction value includes:

[0017] A preset modulation mode mapping table is queried to obtain a modulation ratio index corresponding to the MCS prediction value, wherein the modulation mode mapping table includes a mapping relationship between the MCS prediction value and the modulation ratio index.

[0018] Furthermore, before obtaining a prediction index item based on the indicator affecting the MCS value in the ping service data, the MCS prediction value, and the modulation ratio, and inputting the prediction index item into a rate prediction model to obtain the rate of the ping service data, the method further includes:

[0019] Extracting corresponding prediction index items and rate impact items from the actual rate data;

[0020] Converting the rate impact item according to a preset conversion algorithm to obtain the rate conversion value;

[0021] The corresponding prediction index items extracted from the real rate data are used as data samples, and the rate conversion values ​​are used as labels, and the rate prediction model is trained based on the lasso linear regression algorithm.

[0022] Furthermore, the rate impact item includes the frequency domain RB scheduling number and the time domain SLOT scheduling number, and the rate conversion value is obtained by the following formula:

[0023] Rate conversion value = actual rate * (1300 / number of frequency domain RB scheduling) * (250 / number of time domain SLOT scheduling).

[0024] Furthermore, the rate prediction method based on ping service data further includes:

[0025] Before training the rate prediction model based on the lasso linear regression algorithm using the corresponding prediction indicator items extracted from the true rate data as data samples and the rate conversion values ​​as labels, feature enhancement is performed on the corresponding prediction indicator items extracted from the true rate data and the rate conversion values;

[0026] Before inputting the prediction index item into a rate prediction model to obtain the rate of the ping service data, feature enhancement is performed on the prediction index item.

[0027] In a second aspect, an embodiment of the present invention provides a rate prediction device based on ping service data, comprising:

[0028] An MCS value acquisition module is configured to input an indicator affecting the MCS value in the ping service data into an MCS prediction model to obtain an MCS prediction value for the ping service data, wherein the MCS prediction model is pre-trained based on corresponding indicator samples and MCS value labels in the real rate data;

[0029] A modulation ratio acquisition module, configured to obtain a modulation ratio indicator of ping service data based on the MCS prediction value;

[0030] A rate prediction module is used to obtain a prediction index item based on the indicator affecting the MCS value in the ping service data, the MCS prediction value and the modulation ratio, and input the prediction index item into a rate prediction model to obtain the rate of the ping service data, wherein the rate prediction model is pre-trained based on the corresponding prediction index item samples and rate conversion value labels in the actual rate data.

[0031] In a third aspect, an embodiment of the present invention further provides 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 rate prediction method based on ping service data as described in the first aspect is implemented.

[0032] In a fourth aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the rate prediction method based on ping service data as described in the first aspect.

[0033] It can be seen from the above technical solution that the embodiment of the present invention predicts the MCS value based on the indicator that affects the MCS value in the ping service data, and then obtains the modulation ratio indicator of the ping service data based on the MCS prediction value, so that the indicator in the ping service data that is not equivalent to the download service is converted and then input into the rate prediction model, thereby obtaining the rate of the ping service data. The actual rate of the operator can be predicted through the ping service data, which has the advantages of low testing cost and high testing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0035] Figure 1 This is a flow chart of a method for predicting a rate based on ping service data provided by one embodiment of the present invention;

[0036] Figure 2 is a flow chart of a rate prediction method based on ping service data provided by another embodiment of the present invention;

[0037] Figure 3 This is a mapping diagram of MCS prediction values ​​and modulation modes provided by an embodiment of the present invention;

[0038] Figure 4 This is an MCS prediction data set diagram provided by an embodiment of the present invention;

[0039] Figure 5 This is a diagram of MCS value prediction results provided by an embodiment of the present invention;

[0040] Figure 6 This is a ping service rate prediction data set diagram provided by an embodiment of the present invention;

[0041] Figure 7 This is a diagram showing the prediction results of the ping service rate value provided by an embodiment of the present invention;

[0042] Figure 8 This is a ping service rate value prediction result analysis diagram provided by an embodiment of the present invention;

[0043] Figure 9 This is a real test rate diagram provided by an embodiment of the present invention;

[0044] Figure 10 This is a ping service rate prediction graph provided by an embodiment of the present invention;

[0045] Figure 11 This is a structural block diagram of a rate prediction device based on ping service data provided by an embodiment of the present invention;

[0046] Figure 12 It is a structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following embodiments of the present invention are further described in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0048] The following describes a method, device, electronic device, and storage medium for predicting a rate based on ping service data according to embodiments of the present invention with reference to the accompanying drawings.

[0049] Figure 1 FIG. 1 is a flow chart showing a method for predicting the rate of ping service data according to an embodiment of the present invention. Figure 1 As shown, combined with Figure 2 An embodiment of the present invention provides a method for predicting the rate based on ping service data, including the following contents:

[0050] S101: Input the indicators that affect the MCS value in the ping service data into the MCS prediction model to obtain the MCS prediction value of the ping service data, wherein the MCS prediction model is pre-trained based on corresponding indicator samples and MCS value labels in the real rate data.

[0051] Specifically, before executing the above steps, data preparation and data processing are required. First, select an indicator set, such as selecting 1,000 road test indicators that may be related to the rate of the test data. Secondly, select the data volume, that is, select the actual rate data of competing operators for post-modeling, for example, select a total of 300,000 rows of data. Thirdly, perform data processing. For example, if it is discovered that the data contains some erroneous values, and some numerical data contain special characters such as "#N / A", "#VALUE!", "NIL", " / 0", etc., these sampling points need to be deleted, and sampling points with a frequency domain value less than 88 (i.e., it does not have the characteristics of FTP large package download services) and a time domain value less than 300 (i.e., it does not have the characteristics of FTP large package download services) are also deleted.

[0052] After performing data preparation and data processing, the following feature screening steps are required. Step 1: Based on the service principle, find the road test indicators related to the downlink rate. Downlink rate = frequency domain RB scheduling number * time domain SLOT scheduling number * MCS * modulation mode * RANK * (1-BLER rate), and determine the download rate service-related indicators: frequency domain RB scheduling number, modulation mode, RANK, time domain SLOT scheduling number, and BLER indicator. At the same time, based on the service principle, it is found that MCS has a strong correlation with indicators that characterize the wireless environment, such as SINR, BLER, and CQI. Step 2: Based on the results of the first step, compare whether the characteristics of the download service and the ping service are the same, that is, whether the download rate-related indicators and the ping service indicators are the same. If they are not the same, a solution needs to be found. The final analysis results are shown in Table 1.

[0053] As can be seen from Table 1, download rate-related indicators and Ping service indicators are either the same or different, that is, equivalent or inequality. The download rate-related indicator DL_AvgMCS (i.e., MCS value) in the table is not equivalent to the Ping service indicator. Therefore, it is necessary to predict the MCS value based on the indicators that affect the MCS value in the Ping service data. The MCS value prediction can be achieved by training an MCS prediction model.

[0054]

[0055] Table 1

[0056] In one embodiment of the present application, the final indicators of the MCS prediction model are shown in Table 2, wherein the MCS modeling-related indicator items are indicators that affect the MCS value in the ping service data. The relevant indicators shown in Table 2 are SINR, RANK, UL_InitBLER and AvgCQI. The above indicators that affect the MCS value are input into the MCS prediction model to obtain the MCS prediction value of the ping service data. The MCS prediction model is pre-trained based on the corresponding indicator samples and MCS value labels in the real rate data. Because the MCS value of the ping service is very low and cannot be equivalent to the MCS value at the download rate, this step requires completing the first layer of modeling, that is, MCS prediction model modeling, and inputting the indicators that affect the MCS value in the ping service data into the MCS prediction model to obtain the MCS prediction value of the ping service data.

[0057]

[0058] Table 2

[0059] In one embodiment of the present invention, in step 101, before inputting the indicators affecting the MCS value in the ping service data into the MCS prediction model to obtain the MCS prediction value of the ping service data, the following steps are included: S1011: obtaining real rate data; S1012: extracting the indicators affecting the MCS value and the MCS value from the real rate data; S1013: using the indicators affecting the MCS value and the MCS value extracted from the real rate data as data samples and labels, respectively, and training the MCS prediction model based on the lasso linear regression algorithm.

[0060] Among them, the Lasso linear regression algorithm is an algorithm for supervised learning, that is, a linear function is learned on a given training set, the correlation coefficient is solved under the constraint of the loss function, and finally the regression effect of the model is tested on the test set. The form of the linear model is as follows:

[0061] H = XW;

[0062] Where X is the feature and W is the weight. The goal is to find all W values, and then when a new X value appears, the output H of the function can be estimated. The error between the estimated value and the true value is expressed as the loss function of Lasso regression as follows:

[0063]

[0064] RSS (Residual Sum of Squares) is called the residual sum of squares, where N is the number of samples and λ is a constant coefficient that needs to be tuned. Where X is the regularization term, H is the feature, and W is the coefficient. We need to find the value of W that minimizes the RSS. Since the regularization term is not differentiable at zero, we use the non-gradient descent method, coordinate descent, to solve it. The coordinate descent algorithm selects a dimension for parameter update each time, and the dimension selection can be random or sequential. When the maximum update step size is less than a preset threshold after a round of updates, the iteration terminates.

[0065] Based on the above embodiments, in one embodiment of the present invention, the modeling process of the MCS prediction model is as follows: first, obtain the actual rate data; second, extract the indicators and MCS values ​​that affect the MCS value from the actual rate data, use the indicator items that affect the MCS as features X, such as indicators SINR, BLER, CQI, etc., and use the MCS value as the label H; finally, train the MCS prediction model based on the above-mentioned lasso linear regression algorithm.

[0066] In one embodiment of the present invention, the above-mentioned step 1013 also includes: before the indicators and MCS values ​​affecting the MCS values ​​extracted from the actual rate data are used as data samples and labels respectively, a feature enhancement algorithm is used to perform feature enhancement on the indicators and MCS values ​​affecting the MCS values ​​extracted from the actual rate data; before the indicators affecting the MCS values ​​in the ping service data are input into the MCS prediction model to obtain the MCS prediction value of the ping service data, a feature enhancement algorithm is used to perform feature enhancement on the indicators affecting the MCS values ​​in the ping service data.

[0067] Specifically, the feature enhancement algorithm constructs new features X from the original features x. These new features X and labels H are then fed into the Lasso regression model for training and modeling. Because simple linear regression models have low prediction accuracy, to improve this accuracy, this application employs a polynomial feature generation method using multiple linear functions. After multiple training and verification comparisons, the cubic function has the best generalization ability, so the model ultimately uses the cubic function to generate new features X.

[0068] S102: Based on the MCS prediction value, obtain a modulation ratio index of the ping service data.

[0069] Specifically, since the modulation method of the ping service is very low, it cannot be equivalent to the modulation method at the download rate. The relevant data columns include QPSK proportion, 16QAM proportion, 64QAM proportion, and 256QAM proportion. It is necessary to map the QAM modulation method through the theoretical mapping table based on the MCS value predicted in step 101.

[0070] In one embodiment of the present invention, obtaining the modulation ratio index of the ping service data based on the MCS prediction value in step 102 includes: querying a preset modulation mode mapping table to obtain the modulation ratio index corresponding to the MCS prediction value, wherein the modulation mode mapping table includes a mapping relationship between the MCS prediction value and the modulation ratio index, and the mapping table between the MCS and the modulation mode is as follows: Figure 3 shown.

[0071] S103: Obtain prediction index items based on the indicators affecting the MCS value, the MCS prediction value, and the modulation ratio in the ping service data, and input the prediction index items into the rate prediction model to obtain the rate of the ping service data, wherein the rate prediction model is pre-trained based on the corresponding prediction index item samples and rate conversion value labels in the actual rate data.

[0072] Specifically, after calculating the inequivalence between ping and download services through the first two steps, the characteristic items for rate prediction are ready. This step requires completing the second layer of modeling: the rate prediction model. After the rate prediction model is completed, the indicators influencing the MCS value, the predicted MCS value, and the modulation ratio obtained in steps 101 and 102 are combined into prediction indicators. These prediction indicators are input into the rate prediction model to determine the rate of the ping service data.

[0073] In one embodiment of the present invention, in step 103, before obtaining prediction index items based on the indicators affecting the MCS value, the MCS prediction value and the modulation ratio in the ping service data, and inputting the prediction index items into the rate prediction model to obtain the rate of the ping service data, it also includes: S1031: extracting corresponding prediction index items and rate impact items from the actual rate data; S1032: converting the rate impact items according to a preset conversion algorithm to obtain a rate conversion value; S1033: using the corresponding prediction index items extracted from the actual rate data as data samples, and the rate conversion value as a label, and training the rate prediction model based on the lasso linear regression algorithm.

[0074] Specifically, corresponding prediction indicators are extracted from the actual rate data, as shown in Table 3. The rate modeling indicators correspond to the corresponding prediction indicators. Rate-influencing factors, such as the number of frequency-domain RB scheduling and the number of time-domain slot scheduling, are extracted from the actual rate data. These rate-influencing factors are converted using a preset conversion algorithm to obtain a rate-converted value. The corresponding prediction indicator is then used as the data sample X, and the rate-converted value is used as the label H. A rate prediction model is trained using the lasso linear regression algorithm.

[0075]

[0076] Table 3

[0077] In one embodiment of the present invention, the rate impact item includes the frequency domain RB scheduling number and the time domain SLOT scheduling number, and the rate conversion value is obtained by the following formula:

[0078] Rate conversion value = actual rate * (1300 / number of frequency domain RB scheduling) * (250 / number of time domain SLOT scheduling).

[0079] Specifically, since 5G scheduling includes time domain scheduling and frequency domain scheduling, it is related to influencing factors such as transmission (packet loss, jitter, delay) and server performance, but these influencing factors cannot be obtained. Through analysis of existing network data, it is found that 1300 times in the time domain and 250 RB levels in the frequency domain are the average levels under normal circumstances. At the same time, the downlink rate also needs to be converted proportionally. Therefore, the downlink rate conversion value = actual rate * (1300 / actual time domain scheduling) * (250 / actual frequency domain scheduling).

[0080] In one embodiment of the present invention, step 103 also includes: before using the corresponding prediction indicator items extracted from the actual rate data as data samples and the rate conversion values ​​as labels to train the rate prediction model based on the lasso linear regression algorithm, feature enhancement is performed on the corresponding prediction indicator items extracted from the actual rate data and the rate conversion values; before inputting the prediction indicator items into the rate prediction model to obtain the rate of the ping service data, feature enhancement is performed on the prediction indicator items.

[0081] Specifically, the feature enhancement algorithm constructs new features X from the original features x. These new features X and labels H are then fed into a Lasso regression model for training and modeling. In this embodiment, a cubic function is used to generate the new high-dimensional features X. The new features X and labels H are then fed into the Lasso regression model for training. The relationship between the features and the rate is discovered, and the model and model parameters W are saved. When the rate needs to be predicted, the original features are reconstructed and multiplied by the corresponding coefficient W to calculate the corresponding downlink rate.

[0082] The embodiment of the present invention first uses the first-layer MCS prediction model to predict the MCS prediction value of the ping service, then uses the predicted MCS prediction value to obtain the corresponding modulation ratio through the modulation mode mapping table, constructs the ping service rate prediction index item, and then brings it into the second-layer rate prediction model for prediction, thereby obtaining the downlink rate of the ping service data. The actual rate of the operator can be predicted through the ping service data, which has the advantages of low testing cost and high testing efficiency.

[0083] Based on the above embodiments, in one embodiment of the present invention, the rate prediction process based on ping service data includes the following steps:

[0084] A) Ping service data preparation: that is, data preparation for predicting MCS values. The ping service data in the pilot area is cleaned according to data processing rules, and useless feature columns are removed to form the MCS prediction data set (such as Figure 4 shown);

[0085] B) Enhancement of feature items for predicting MCS: The original features that affect MCS are used as input to generate new feature items X;

[0086] C) Predicting the MCS value: Multiply the new feature item X by the MCS model parameter W to calculate the MCS value, that is, the MCS value H = X*W. The prediction result is as follows: Figure 5 As shown;

[0087] D) Ping service data preparation: that is, feature data preparation for rate prediction. According to the predicted MCS, the mapping table is searched to obtain the corresponding modulation mode ratio, and the ping service rate prediction data set is constructed. Figure 6 As shown;

[0088] E) Enhancement of feature items for predicting the ping service rate in ping service data: using a polynomial feature enhancement algorithm to take the original features that affect the ping service download rate as input to generate a new feature item X;

[0089] F) Final prediction of ping service rate value: The new feature item X is multiplied by the rate prediction model parameter W to calculate the downlink rate, that is, the downlink rate H = X * W. The prediction result is as follows: Figure 7 As shown;

[0090] G) Analysis of prediction results: Comparing the predicted ping service rate values ​​and the actual test values ​​of key roads, it can be seen that the predicted rate is basically consistent with the actual rate, with only a deviation of 2.4%. The prediction results are as follows: Figures 8-10 shown.

[0091] The embodiment of the present invention predicts the MCS value based on the indicator that affects the MCS value in the ping service data, and then obtains the modulation ratio indicator of the ping service data based on the MCS prediction value, so as to convert the indicator that is not equivalent to the download service in the ping service data and then input it into the rate prediction model, thereby obtaining the rate of the ping service data. The actual rate of the operator can be predicted through the ping service data, which has the advantages of low testing cost and high testing efficiency.

[0092] Figure 11 FIG. 1 shows a schematic diagram of a rate prediction device based on ping service data according to an embodiment of the present invention. Figure 11 As shown, an embodiment of the present invention provides a rate prediction device based on ping service data, comprising:

[0093] The MCS value acquisition module 1110 is used to input the indicators that affect the MCS value in the ping service data into the MCS prediction model to obtain the MCS prediction value of the ping service data, wherein the MCS prediction model is pre-trained based on the corresponding indicator samples and MCS value labels in the real rate data.

[0094] The modulation ratio acquisition module 1120 is used to obtain a modulation ratio indicator of the ping service data based on the MCS prediction value.

[0095] The rate prediction module 1130 is used to obtain a prediction index item based on the indicator affecting the MCS value in the ping service data, the MCS prediction value and the modulation ratio, and input the prediction index item into a rate prediction model to obtain the rate of the ping service data, wherein the rate prediction model is pre-trained based on the corresponding prediction index item samples and rate conversion value labels in the actual rate data.

[0096] The rate prediction device based on ping service data provided by the embodiment of the present invention predicts the MCS value according to the indicator that affects the MCS value in the ping service data, and then obtains the modulation ratio indicator of the ping service data according to the MCS prediction value, so as to convert the indicator that is not equivalent to the download service in the ping service data and then input it into the rate prediction model, thereby obtaining the rate of the ping service data. The actual rate of the operator can be predicted through the ping service data, which has the advantages of low testing cost and high testing efficiency.

[0097] It should be noted that the specific implementation of the rate prediction device based on ping service data in an embodiment of the present invention is similar to the specific implementation of the rate prediction method based on ping service data in an embodiment of the present invention. Please refer to the description of the method part for details. In order to reduce redundancy, the details will not be repeated here.

[0098] Based on the same inventive concept, another embodiment of the present invention provides an electronic device, see Figure 12 , the electronic device specifically includes the following contents: a processor 1201, a memory 1202, a communication interface 1203 and a communication bus 1204;

[0099] The processor 1201, the memory 1202, and the communication interface 1203 communicate with each other via the communication bus 1204; the communication interface 1203 is used to implement information transmission between devices;

[0100] The processor 1201 is configured to call the computer program in the memory 1202 , and when the processor executes the computer program, all steps of the above-mentioned rate prediction method based on ping service data are implemented.

[0101] Based on the same inventive concept, another embodiment of the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, all steps of the above-mentioned rate prediction method based on ping service data are implemented.

[0102] In addition, the logical instructions in the above-mentioned memory can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0103] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the embodiments of the present invention. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0104] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus the necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the indicator monitoring method described in each embodiment or certain parts of the embodiment.

[0105] Furthermore, in the present invention, terms such as "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of such features. In the description of the present invention, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0106] In addition, in the present invention, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further limitations, an element defined by the statement "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0107] In addition, in the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0108] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A rate prediction method based on ping service data, characterized in that: include: Inputting an indicator affecting the MCS value in the ping service data into an MCS prediction model to obtain an MCS prediction value of the ping service data, wherein the MCS prediction model is pre-trained based on corresponding indicator samples and MCS value labels in the real rate data; Obtaining a modulation ratio indicator of the ping service data based on the MCS prediction value; A prediction index item is obtained based on the indicator affecting the MCS value in the ping service data, the MCS prediction value and the modulation ratio, and the prediction index item is input into a rate prediction model to obtain the rate of the ping service data, wherein the rate prediction model is pre-trained based on the corresponding prediction index item samples and rate conversion value labels in the actual rate data.

2. The rate prediction method based on ping service data according to claim 1, characterized in that: Before inputting the indicator affecting the MCS value in the ping service data into the MCS prediction model to obtain the MCS prediction value of the ping service data, the method includes: Get real rate data; Extracting an index affecting an MCS value and an MCS value from the true rate data; The indicators affecting the MCS value and the MCS value extracted from the real rate data are used as data samples and labels respectively, and the MCS prediction model is trained based on the lasso linear regression algorithm.

3. The rate prediction method based on ping service data according to claim 2, characterized in that: Also includes: Before using the index affecting the MCS value and the MCS value extracted from the true rate data as data samples and labels, respectively, a feature enhancement algorithm is used to perform feature enhancement on the index affecting the MCS value and the MCS value extracted from the true rate data; Before inputting the indicators affecting the MCS value in the ping service data into the MCS prediction model to obtain the MCS prediction value of the ping service data, the feature enhancement algorithm is used to perform feature enhancement on the indicators affecting the MCS value in the ping service data.

4. The method for predicting the rate based on ping service data according to any one of claims 1 to 3, characterized in that: The obtaining, based on the MCS prediction value, a modulation ratio indicator of the ping service data, includes: A preset modulation mode mapping table is queried to obtain a modulation ratio index corresponding to the MCS prediction value, wherein the modulation mode mapping table includes a mapping relationship between the MCS prediction value and the modulation ratio index.

5. The rate prediction method based on ping service data according to claim 1, characterized in that: Before obtaining a prediction index item based on an indicator affecting the MCS value in the ping service data, the MCS prediction value, and the modulation ratio, and inputting the prediction index item into a rate prediction model to obtain the rate of the ping service data, the method further includes: Extracting corresponding prediction index items and rate impact items from the actual rate data; Converting the rate impact item according to a preset conversion algorithm to obtain the rate conversion value; The corresponding prediction index items extracted from the real rate data are used as data samples, and the rate conversion values ​​are used as labels, and the rate prediction model is trained based on the lasso linear regression algorithm.

6. The rate prediction method based on ping service data according to claim 5, characterized in that: The rate impact item includes the frequency domain RB scheduling number and the time domain SLOT scheduling number. The rate conversion value is obtained by the following formula: Rate conversion value = actual rate * (1300 / number of frequency domain RB scheduling) * (250 / number of time domain SLOT scheduling).

7. The rate prediction method based on ping service data according to claim 5 or 6, characterized in that: Also includes: Before training the rate prediction model based on the lasso linear regression algorithm using the corresponding prediction indicator items extracted from the true rate data as data samples and the rate conversion values ​​as labels, feature enhancement is performed on the corresponding prediction indicator items extracted from the true rate data and the rate conversion values; Before inputting the prediction index item into a rate prediction model to obtain the rate of the ping service data, feature enhancement is performed on the prediction index item.

8. A rate prediction device based on ping service data, characterized in that: include: An MCS value acquisition module is configured to input an indicator affecting the MCS value in the ping service data into an MCS prediction model to obtain an MCS prediction value for the ping service data, wherein the MCS prediction model is pre-trained based on corresponding indicator samples and MCS value labels in the real rate data; A modulation ratio acquisition module, configured to obtain a modulation ratio indicator of the ping service data based on the MCS prediction value; A rate prediction module is used to obtain a prediction index item based on the indicator affecting the MCS value in the ping service data, the MCS prediction value and the modulation ratio, and input the prediction index item into a rate prediction model to obtain the rate of the ping service data, wherein the rate prediction model is pre-trained based on the corresponding prediction index item samples and rate conversion value labels in the actual rate data.

9. 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 rate prediction method based on ping service data is implemented according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the rate prediction method based on ping service data according to any one of claims 1 to 7 is implemented.

Images