A method, apparatus and device for predicting quality of service

Abstract

The application discloses a service quality prediction method, device and equipment, the method comprises the following steps: acquiring at least one index parameter related to service response success rate; inputting the at least one index parameter into a service quality evaluation model for processing to obtain a target service quality level corresponding to the service response success rate, the target service quality level belongs to one of a plurality of service quality levels, the plurality of service quality levels are obtained by self-adaptive service quality level division according to the service response success rate through a preset algorithm; acquiring a time sequence sample set composed of a plurality of target service quality levels contained in a sliding time window; inputting the time sequence sample set into a service quality prediction model for processing to obtain a service quality level at a next moment. Through the above manner, the application can perform self-adaptive division of service quality levels, and through the service quality evaluation model and the service quality prediction model, the accuracy of service quality prediction is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of network communication technology, in particular to a service quality prediction method, device and equipment. BACKGROUND

[0002] With the increasing competition in the mobile communication market, network quality becomes a key factor to enhance the competitiveness of enterprises. In order to comprehensively measure the network quality level and timely perceive the changes in network operation from the customer's perspective, it is necessary to improve mobile network quality management, mine and expand new network indicators, and carry out targeted maintenance and optimization work to maintain the leading advantage in network quality. The network indicator in-depth research based on this aims to establish a network quality indicator system based on customer perception.

[0003] The service scenario is complex, and the indicators differ under different services. The existing technology analyzes according to a single service or indicator, and the accuracy of the result is not high. When scoring and grading, a predetermined threshold needs to be set, and without data reference, it is impossible to judge its reference and evaluation value; the service quality cannot be effectively predicted. SUMMARY

[0004] In view of the above problems, the embodiments of the present application are proposed to provide a service quality prediction method, device and equipment which can overcome the above problems or at least partially solve the above problems.

[0005] According to an aspect of an embodiment of the present application, a service quality prediction method is provided, comprising:

[0006] obtaining at least one indicator parameter related to service response success rate;

[0007] inputting the at least one indicator parameter into a service quality evaluation model for processing to obtain a target service quality grade corresponding to the service response success rate, the target service quality grade belonging to one of a plurality of service quality grades, the plurality of service quality grades being obtained by self-adaptive service quality grade division according to the service response success rate through a preset algorithm;

[0008] obtaining a time sequence sample set composed of a plurality of target service quality grades contained in a sliding time window;

[0009] inputting the time sequence sample set into a service quality prediction model for processing to obtain a service quality grade at the next moment.

[0010] According to another aspect of an embodiment of the present application, a service quality prediction device is provided, comprising:

[0011] an obtaining module configured to obtain at least one indicator parameter related to service response success rate;

[0012] a processing module configured to input the at least one index parameter into a service quality evaluation model to obtain a target service quality level corresponding to the service response success rate, the target service quality level belonging to one of a plurality of service quality levels, the plurality of service quality levels being obtained by self-adapting service quality level division according to the service response success rate through a preset algorithm; obtain a time sequence sample set composed of a plurality of target service quality levels contained in a sliding time window; and input the time sequence sample set into a service quality prediction model to obtain a service quality level at a next time.

[0013] According to another aspect of the embodiments of the present application, a computing device is provided, comprising a processor, a memory, a communication interface and a communication bus, the processor, the memory and the communication interface being capable of communicating with each other through the communication bus;

[0014] The memory is configured to store at least one executable instruction, the executable instruction enabling the processor to perform operations corresponding to the above-mentioned service quality prediction method.

[0015] According to another aspect of the embodiments of the present application, a computer storage medium is provided, the storage medium storing at least one executable instruction, the executable instruction enabling a processor to perform operations corresponding to the above-mentioned service quality prediction method.

[0016] According to the above-mentioned embodiments of the present application, at least one index parameter related to the service response success rate is obtained; the at least one index parameter is input into a service quality evaluation model to obtain a target service quality level corresponding to the service response success rate, the target service quality level belonging to one of a plurality of service quality levels, the plurality of service quality levels being obtained by self-adapting service quality level division according to the service response success rate through a preset algorithm; a time sequence sample set composed of a plurality of target service quality levels contained in a sliding time window is obtained; thereby self-adapting service quality level division is realized, and the accuracy of service quality prediction is improved.

[0017] The above description is only a summary of the technical solutions of the embodiments of the present application, in order to more clearly understand the technical means of the embodiments of the present application, the embodiments of the present application can be implemented according to the content of the specification, and in order to make the above-mentioned and other purposes, features and advantages of the embodiments of the present application more obvious and easy to understand, the following specific embodiments of the embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0018] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0019] Figure 1 A flowchart of the service quality prediction method provided by an embodiment of the present invention is shown;

[0020] Figure 2 This diagram illustrates the accuracy of experimental data in the data preprocessing method of the business quality prediction method provided in this embodiment of the invention.

[0021] Figure 3 A schematic diagram of the structure of the service quality prediction device provided in an embodiment of the present invention is shown;

[0022] Figure 4 A schematic diagram of another service quality prediction device provided in an embodiment of the present invention is shown;

[0023] Figure 5 A schematic diagram of the structure of a computing device provided in an embodiment of the present invention is shown. Detailed Implementation

[0024] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0025] Figure 1 A flowchart of a service quality prediction method provided by an embodiment of the present invention is shown. Figure 1 As shown, the method includes the following steps:

[0026] Step 11: Obtain at least one metric parameter related to the business response success rate;

[0027] Step 12: Input the at least one indicator parameter into the business quality assessment model for processing to obtain the target business quality level corresponding to the business response success rate. The target business quality level is one of multiple business quality levels. The multiple business quality levels are obtained by adaptively classifying business quality levels based on the business response success rate using a preset algorithm.

[0028] Step 13: Obtain a time series sample set consisting of multiple target service quality levels contained in a sliding time window;

[0029] Step 14: Input the time series sample set into the service quality prediction model for processing to obtain the service quality level at the next time step.

[0030] In this embodiment, at least one indicator parameter related to the business response success rate is obtained; the at least one indicator parameter is input into a business quality assessment model for processing to obtain the target business quality level corresponding to the business response success rate; a time series sample set consisting of multiple target business quality levels contained in a sliding time window is obtained; the time series sample set is input into a business quality prediction model for processing to obtain the business quality level at the next moment; thereby achieving adaptive classification of business quality levels, and improving the accuracy of business quality prediction through the business quality assessment model and the business quality prediction model.

[0031] In an optional embodiment of the present invention, at least one metric parameter related to the success rate of the service response in step 11 includes: the first HTTP response time (FHRT), the last packet time, and the transmission control protocol response (TCP ACK) time.

[0032] In this embodiment, Business Response Success Rate (BISR) is used as the basis for classifying business quality levels. Specifically, Pearson correlation coefficient analysis can be used to analyze the correlation between the average value of the indicators related to BISR and BISR. The larger the absolute value of the Pearson correlation coefficient, the stronger the correlation; the smaller the absolute value of the Pearson correlation coefficient, the weaker the correlation. Based on the correlation between the indicator parameters and BISR, the indicator parameters with the highest correlation to BISR are selected as the input of the business quality assessment model (XGB_QE). The indicator parameters with the highest correlation to BISR include: FHRT (First HTTP Response Time), LCPT (Last Content Packet Time), and TAT (TCP ACK Time).

[0033] In the above embodiments of the present invention, before constructing the business quality assessment model, it is necessary to adaptively classify the business quality levels. Specifically, the fuzzy C-means algorithm (FCM) can be used for adaptive classification of business quality levels. Before adaptively classifying the business quality levels, the business data can also be processed accordingly. The specific process is as follows:

[0034] Step 110, Obtain source data;

[0035] Step 111: Preprocess the source data to obtain the preprocessed result;

[0036] Step 112: Based on the preprocessing results, the service quality level is adaptively divided using the FCM algorithm to obtain multiple service quality levels;

[0037] In step 12 above, the at least one indicator parameter is input into the business quality assessment model for processing to obtain a definite target business quality level. Finally, the sequence of all the definite target business quality levels is used as the business quality level sequence output by the business quality assessment model.

[0038] Here, the source data can be formatted as follows:

[0039]

[0040]

[0041] In this embodiment, the specific implementation process of step 111 includes:

[0042] Data cleaning: In the preprocessing stage, sample data of the source data is obtained. It is determined whether there is missing, non-numerical, or other invalid data in the sample data, including the categories and data required by each indicator algorithm. Effective preprocessing can make the model's results more accurate and is a crucial process when using machine learning methods. After obtaining the basic data and removing invalid data, it is input into the corresponding algorithm for processing.

[0043] Experiments were conducted to compare several preprocessing schemes and identify the optimal combination of preprocessing techniques. The preprocessing included at least one of the following: normalization techniques, feature selection, and synthetic minority oversampling (SMOTE) techniques.

[0044] The specific preprocessing data is as follows:

[0045] Serial number Preprocessing technique 1 No preprocessing 2 Regularization + feature selection 3 Regularization + RUS 4 Feature selection + RUS 5 Standardization + ROS 6 Feature selection + ROS 7 Standardization + SMOTE 8 Feature selection + SMOTE 9 Regularization + feature selection + RUS 10 Regularization + feature selection + ROS 11 Regularization + feature selection + SMOTE

[0046] Eleven preprocessing scenarios were compared to identify the optimal combination of preprocessing techniques. Experiments began with no preprocessing and continued with combinations of several preprocessing techniques. The experimental results for the preprocessing techniques are as follows: Figure 2 As shown.

[0047] This embodiment of the invention describes a method for adaptively classifying service quality levels based on an improved fuzzy C-means algorithm (FCM). Specifically, the FCM algorithm is a flexible fuzzy partitioning method. It treats each cluster generated by clustering as a fuzzy set. The membership degree of each sample is calculated to determine which cluster it belongs to. Clustering algorithms are more accurate and flexible than hard partitioning, but several issues need to be addressed:

[0048] 1) The choice of p-value plays an important role in the smoothness of noisy data and membership, and affects the clustering results.

[0049] 2) FCM is very sensitive to the selection of initial cluster centers because its objective function is a non-linear convex function. If the cluster centers are not selected properly, FCM can easily get trapped in local optima, thus affecting the algorithm's global judgment.

[0050] 3) In practice, the number of clusters is often unknown. The optimal number of clusters varies for different problems, affecting clustering performance. Therefore, it is necessary to select an appropriate number of clusters.

[0051] To further improve the performance of the FCM algorithm, a fast and independent Xie-beni exponential clustering algorithm is used to find the initial cluster centers of FCM, so as to avoid getting trapped in local optima and improve the quality and efficiency of FCM.

[0052] Specifically, a fuzzy measure of p=2 can be chosen. Considering the fuzzy membership and geometric structure of the dataset, the number of clusters is adaptively determined using the Xie_Beni coefficient, an effectiveness index based on cluster compactness and separation. Therefore, the number of levels to which business quality should be divided can be adaptively determined.

[0053] Here, the Business Response Success Rate (BISR) is used as the input to the FCM algorithm. The purpose of clustering is to cluster the BISR, classify the business quality levels, and minimize the clustering objective function.

[0054] The objective function is as follows:

[0055]

[0056] Among them, u ij Let represent the membership degree of the j-th BISR belonging to the quality level of the i-th stage. The membership degree indicates the extent to which the j-th BISR belongs to the quality level of the i-th stage. Where 0 ≤ u ij ≤1, u ij =1 indicates that the j-th BISR is completely subordinate to level i, u ij =0 indicates that the j-th BISR does not belong to level i. The j-th BISR belongs to the service quality level corresponding to the highest membership degree. d ij This represents the distance from the j-th BISR to the i-th Business Quality Level, calculated using Euclidean distance. p is a fuzzy measure.

[0057] The process of classifying service quality levels based on FCM is as follows:

[0058] Input: Dataset D BISR ={BISR1, BISR2, ..., BISR N}

[0059] Output: Optimal number of clusters c * Quality grade D level ={Level i}, i = 1, 2, ..., N;

[0060] The specific steps are as follows:

[0061] Given a fuzzy measure p = 2, an initial number of clusters c = 2, an iteration count t = 0, and a convergence precision ε, set the optimal number of clusters to c. * Set the maximum number of clusters to c. max ;

[0062] Loop: for c <c max By using D BISR Subtractive clustering initialization of cluster centers V (0) ={v1,v2,…v c Each time, c (c = c + 1)

[0063] for||U (t+1) -U (t) ||>ε, each time t (t=t+1), calculate the membership degree u of BISR. ij And update the membership matrix.

[0064] U (t+1) ={u ij};

[0065] Calculate the cluster center v of BISR j And update the cluster center matrix V (t+1) ={v j} Calculate the value of the effectiveness index Xie_Beni, and output XB and c;

[0066] If XB is less than the latest XB setting c * =c,U * =U,V * =V;

[0067] Output c * and according to U * V * D level .

[0068] By classifying services into quality levels based on the aforementioned algorithm, the service quality parameters are consistent within the same level, while differing between different levels. Therefore, the service quality levels are determined, reflecting the quality of the services offered.

[0069] In another optional embodiment of the present invention, the business quality assessment model in step 12 is:

[0070] Where G(X) is the business quality assessment model, and I m (X) is the m-th decision tree, M is the number of decision trees, and I i (X) is the i-th decision tree.

[0071] In this embodiment, the business quality assessment model is implemented based on the XGBoost model, which is an ensemble algorithm based on tree or linear classifiers. It integrates several weak classifiers to form a strong classifier with good classification or regression performance; the regularization term of the XGBoost objective function includes the weights of the leaf nodes and the depth of the tree, which can control the complexity of the model and prevent overfitting; the XGBoost objective function is approximated by a second-order Taylor expansion, which optimizes the objective function and ensures prediction accuracy.

[0072] Business quality assessment is the process of determining the current business quality level based on selected business quality parameters, namely indicators that are highly correlated with BISR, such as FHRT, LCPT, and TAT.

[0073] Different quality parameters at different stages reflect the characteristics of those stages. Business quality levels directly reflect business quality. Therefore, the business quality assessment problem is transformed into a classification problem. A business quality assessment model is constructed based on the XGBoost classification algorithm, which leverages XGBoost's classification advantages. Business quality assessment is achieved through the mapping relationship between indicator parameters and business quality levels. The XGBoost model is used to obtain the business quality level at the current time.

[0074] The average values ​​of the preprocessed FHRT, LCPT, and TAT are combined to form a triple {μ(FHRT)} i ),μ(LCPT i ),μ(TAT i The dataset D, consisting of (i = 1, 2, ..., N) characters, serves as the input to the business quality assessment model. eva ={(X i ,Y i )}, (i = 1, 2, ..., N), where X i ={μ(FHRT) i ),μ(LCPT i ),μ(TATi )} is the input to the business quality assessment model, Y i ={Level i} is the output of the business quality assessment model.

[0075] Multiple decision trees are gradually added to the model; the objective function decreases relatively with each new decision tree added. Multiple decision trees are constructed by minimizing the objective function and then combined into a strong classifier. The leaf nodes of each decision tree are assigned weights corresponding to the evaluated business quality level. When business quality parameters are input into the classifier, the corresponding business quality level is determined based on the attributes.

[0076] The objective function includes a loss function and a regularization term. Since the business quality assessment problem is transformed into a classification problem, the Softmax function is used as the loss function.

[0077] The splitting process of a decision tree differs from that of a general decision tree. General decision trees do not consider complexity during splitting, relying on subsequent pruning operations to control complexity. XGB_QE adds a regularization term to the objective function to control tree complexity and avoid overfitting during splitting. Each decision tree is constructed iteratively based on previous decision trees. Each time a decision tree is constructed, the gain value of each leaf node is calculated, and the leaf node with the highest gain value is selected for splitting.

[0078] The decision tree will stop splitting when the gain value of a split leaf node is less than zero or the tree depth reaches the specified maximum depth. This completes the optimization of leaf node values ​​and the decision tree structure, resulting in the final business quality assessment model.

[0079] The business quality assessment model is developed by training m decision trees {I1(X), I2(X), ..., I...} M The formula for training the M-decision tree, obtained from (X)}, is: Where G(X) is the final strong classifier, i.e., the business quality assessment model, I m (X) is the m-th decision tree.

[0080] In this process, the maximum step size of the weight change for each tree during the iteration of the business quality assessment model was controlled, and the weights of each tree were adjusted. This avoids the influence of good business data on the classifier and reduces errors caused by imbalanced training data.

[0081] In another optional embodiment of the present invention, step 12 may include:

[0082] Step 121: Combine the average of the preprocessed first HTTP response time (FHRT), the last packet time, and the TCP ACK time to form a triple {μ(FHRT)}. i),μ(LCPT i ),μ(TAT i )}, (i=1,2,...,N), are input into the business quality assessment model for processing to obtain the target business quality level Y. i ={Level i}

[0083] In this embodiment, the average values ​​of FHRT LCPT and TAT are combined to form a triplet input service quality assessment model, thereby outputting a specific service quality level, which is one of the service quality levels obtained by the above-mentioned FCM algorithm.

[0084] In another optional embodiment of the present invention, step 13 may include:

[0085] Step 131: Based on the service quality level sequence XGB_Model{Level} constructed according to the target service quality level. i The time series sample set X consists of w service quality levels contained in the sliding time window (i = 1, 2, ..., N), where i = 1, 2, ..., N. i ={Level i Level i+1 ,…,Level i+w-1}

[0086] In this embodiment, service quality exhibits strong regularity in the short term. Therefore, XGB_Model{Level i The evaluation results of}, (i=1,2,…,N) are used to construct a sample set X of the time series using a sliding time window of size w. i ={Level i Level i+1 ,…,Level i+w-1}

[0087] In another optional embodiment of the present invention, the service quality prediction model is:

[0088]

[0089] Among them, f m This is the m-th regression tree, where M is the total number of regression trees. It is the i-th predicted business quality level. It is a business quality prediction model.

[0090] Here, the M-regression tree is constructed by progressively optimizing the objective function, and finally the M-regression trees are summed to obtain the business quality prediction model.

[0091] Definition of the objective function for the business quality prediction model:

[0092]

[0093] Ω(f m )=γT+1 / 2λ‖ω‖ 2 It is a regularization term that measures the complexity of the regression tree, including the number of leaf nodes and L2 regularization. T is the number of leaf nodes, ω represents the score of the leaf nodes, and γ and λ are weight parameters. It is the squared loss function.

[0094] The objective function is optimized step by step, and the main process is as follows:

[0095]

[0096]

[0097] …

[0098]

[0099] The objective function of the model at step t is as shown in equation [equation missing]

[0100]

[0101] Based on the second-order Taylor expansion, the first and second derivatives of the loss function, gi and hi, are replaced with the objective function. The objective function at step t is given by the following equation:

[0102]

[0103] Where C is a constant, I j ={i∣q(Y i Let ) = j} be the set of instances of leaf node j, representing the corresponding business quality value. By rewriting the above formula, we can use it as a scoring function to measure the quality of the tree structure:

[0104]

[0105] The formula corresponds to the minimum value, and the optimal weight of leaf j is calculated.

[0106]

[0107] The corresponding objective function is as follows:

[0108]

[0109] Next, a greedy algorithm can be used to iteratively add branches to the tree, starting from a single leaf. The loss function after node splitting is as follows:

[0110]

[0111] Among them, I L and I R It is the set of left and right node instances after the split, I = I L +I R Iterate until Gain is less than zero or the maximum depth value is reached.

[0112] In yet another optional embodiment of the invention, step 14 may include:

[0113] Step 141, the time series sample set X i ={Level i Level i+1 ,…,Level i+w-1 The input to the business quality prediction model is processed to obtain M prediction results; one regression tree corresponds to one prediction result.

[0114] Step 142: Add the M prediction results together to obtain the service quality level at the next moment.

[0115] In this embodiment, by using {Level i Level i+1 ,…,Level i+w-1 The data is input into the business quality prediction model, and the prediction results of the M-regression tree are added together to obtain the business quality level at the next moment.

[0116] In the above embodiments of the present invention, at least one indicator parameter related to the business response success rate is obtained; the at least one indicator parameter is input into a business quality assessment model for processing to obtain a target business quality level corresponding to the business response success rate. The target business quality level is one of multiple business quality levels, which are obtained by adaptively classifying business quality levels based on the business response success rate using a preset algorithm; a time series sample set consisting of multiple target business quality levels contained in a sliding time window is obtained; the time series sample set is input into a business quality prediction model for processing to obtain the business quality level at the next moment. An adaptive business quality perception model can be established to predict business quality over a period of time. This achieves adaptive classification of multiple business quality levels based on highly reliable sample data using an improved FCM algorithm, and can determine a specific business quality level based on the indicator parameter that is highly correlated with the business response success rate (BISR) through the business quality assessment model, and predict the business quality level at the next moment based on the determined business quality level, thereby improving the accuracy of business quality assessment.

[0117] Figure 3A schematic diagram of the structure of a service quality prediction device 30 provided in an embodiment of the present invention is shown. Figure 3 As shown, the device includes:

[0118] Module 31 is used to acquire at least one metric parameter related to the business response success rate;

[0119] Processing module 32 is used to input the at least one indicator parameter into a service quality assessment model for processing to obtain a target service quality level corresponding to the service response success rate. The target service quality level is one of multiple service quality levels, which are obtained by adaptively classifying service quality levels based on the service response success rate using a preset algorithm. It also acquires a time series sample set consisting of multiple target service quality levels contained in a sliding time window and inputs the time series sample set into a service quality prediction model for processing to obtain the service quality level at the next moment.

[0120] Optionally, at least one metric parameter related to the success rate of the service response includes: First HTTP Response Time (FHRT), Last Content Packet Time, and Transmission Control Protocol (TCP) ACK Time.

[0121] Optionally, the business quality assessment model is:

[0122] Where G(X) is the business quality assessment model, and I m (X) is the m-th decision tree, M is the number of decision trees, and I i (X) is the i-th decision tree.

[0123] Optionally, the processing module 32 is further configured to combine the average of the preprocessed first HTTP response time (FHRT), the last packet time, and the TCP ACK time of the transmission control protocol response to form a triple {μ(FHRT)}. i ),μ(LCPT i ),μ(TAT i )}, (i=1,2,...,N), are input into the business quality assessment model for processing to obtain the target business quality level Y. i ={Level i}

[0124] Optionally, the processing module 32 is further configured to perform a service quality level sequence XGB_Model{Level} based on the target service quality level. i The time series sample set X consists of w service quality levels contained in the sliding time window (i = 1, 2, ..., N), where i = 1, 2, ..., N. i ={Level i Leveli+1 ,…,Level i+w-1}

[0125] Optionally, the business quality prediction model is:

[0126]

[0127] Among them, f m This is the m-th regression tree, where M is the total number of regression trees. It is the i-th predicted business quality level. It is a business quality prediction model.

[0128] Optionally, the processing module 32 is further configured to process the time series sample set X i ={Level i Level i+1 ,…,Level i+w-1 The input to the business quality prediction model is processed to obtain M prediction results; one regression tree corresponds to one prediction result.

[0129] The service quality level for the next moment is obtained by summing the M prediction results.

[0130] Figure 4 A schematic diagram of another service quality prediction device provided in an embodiment of the present invention is shown, as follows: Figure 4 As shown, the business quality prediction device further includes: a data acquisition module, used to acquire data from the data source and perform effective preprocessing operations. First, it determines whether there is missing, non-numerical, or other invalid data in the sample data and the data required by the algorithms for each category and indicator. After obtaining the basic data and removing invalid data, it inputs the data into the modeling module, which makes the results obtained by the model more accurate.

[0131] The above processing module may include a modeling module and a prediction module, and the functions of the modeling module and the prediction module are implemented through the methods in the above method embodiments.

[0132] It should be noted that this embodiment is a device embodiment corresponding to the above method embodiment. All implementation methods in the above method embodiment are applicable to this device embodiment and can achieve the same technical effect.

[0133] This invention provides a non-volatile computer storage medium storing at least one executable instruction that can execute the service quality prediction method in any of the above method embodiments.

[0134] Figure 5The diagram shows a structural schematic of a computing device provided in an embodiment of the present invention. The specific embodiments of the present invention do not limit the specific implementation of the computing device.

[0135] like Figure 5 As shown, the computing device may include a processor, a communications interface, memory, and a communications bus.

[0136] The processor, communication interface, and memory communicate with each other via a communication bus. The communication interface is used to communicate with other network elements, such as clients or other servers. The processor executes programs, specifically the relevant steps in the above-described embodiment of the service quality prediction method for computing devices.

[0137] Specifically, the program may include program code, which includes computer operation instructions.

[0138] The processor may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The computing device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0139] Memory is used to store programs. Memory may include high-speed RAM, and may also include non-volatile memory, such as at least one disk drive.

[0140] Specifically, the program can be used to cause the processor to execute the service quality prediction method in any of the above method embodiments. The specific implementation of each step in the program can be found in the corresponding descriptions of the steps and units in the above service quality prediction method embodiments, and will not be repeated here. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the devices and modules described above can be referred to the corresponding process descriptions in the foregoing method embodiments, and will not be repeated here.

[0141] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, the embodiments of the present invention are not directed to any particular programming language. It should be understood that the embodiments of the present invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the embodiments of the present invention.

[0142] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0143] Similarly, it should be understood that, in order to streamline the embodiments of the invention and aid in understanding one or more of the various inventive aspects, features of the embodiments of the invention are sometimes grouped together in a single embodiment, figure, or description thereof in the above description of exemplary embodiments of the invention. However, this disclosure should not be construed as reflecting an intention that the claimed embodiments of the invention require more features than are expressly recited in each claim. Rather, as reflected in the following claims, inventive aspects lie in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of the invention.

[0144] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0145] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0146] The various component embodiments of the present invention can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components according to the embodiments of the present invention. The embodiments of the present invention can also be implemented as device or apparatus programs (e.g., computer programs and computer program products) for performing part or all of the methods described herein. Such programs implementing the embodiments of the present invention can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0147] It should be noted that the above embodiments are illustrative of the present invention and not restrictive of the invention, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. Embodiments of the present invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A method for predicting business quality, characterized in that, The method includes: Obtain at least one metric parameter related to the business response success rate; wherein, the at least one metric parameter related to the business response success rate includes: First HTTP Response Time (FHRT), Last Packet Time, and Transmission Control Protocol Response (TCPACK) Time; The at least one indicator parameter is input into the business quality assessment model for processing to obtain the target business quality level corresponding to the business response success rate. The target business quality level is one of multiple business quality levels, which are obtained by adaptively classifying business quality levels based on the business response success rate using a preset algorithm. The number of clusters is adaptively determined by the Xie_Beni coefficient, an effectiveness indicator based on cluster compactness and separation, so as to adaptively determine how many levels of business quality are divided into. Obtain a time series sample set consisting of multiple target service quality levels contained within a sliding time window; The time series sample set is input into the service quality prediction model for processing to obtain the service quality level at the next time step.

2. The business quality prediction method according to claim 1, characterized in that, The business quality assessment model is as follows: ; in, It is a business quality assessment model. It is the m-th decision tree, where M is the number of decision trees.

3. The business quality prediction method according to claim 1, characterized in that, The at least one indicator parameter is input into the business quality assessment model for processing to obtain the target business quality level corresponding to the business response success rate, including: The average of the preprocessed First HTTP Response Time (FHRT), the Last Packet Time, and the Transmission Control Protocol (TCP) ACK Time is combined to form a triple. For each i=1,2,...,N, the input is processed by the business quality assessment model to obtain the target business quality level corresponding to the business response success rate. .

4. The business quality prediction method according to claim 1, characterized in that, Obtain a time series sample set consisting of multiple target service quality levels contained within a sliding time window, including: A service quality level sequence is constructed based on the target service quality level corresponding to the service response success rate. Given i = 1, 2, ..., N, obtain the time series sample set consisting of w service quality levels contained in the sliding time window. .

5. The business quality prediction method according to claim 1, characterized in that, The business quality prediction model is as follows: ； in, This is the m-th regression tree, where M is the total number of regression trees. It is the i-th predicted business quality level. It is a business quality prediction model.

6. The business quality prediction method according to claim 5, characterized in that, The time series sample set is input into the service quality prediction model for processing to obtain the service quality level at the next time step, including: The time series sample set The input to the business quality prediction model is processed to obtain M prediction results; one regression tree corresponds to one prediction result. The service quality level for the next moment is obtained by summing the M prediction results.

7. A business quality prediction device, characterized in that, include: The acquisition module is used to acquire at least one indicator parameter related to the business response success rate; wherein, the at least one indicator parameter related to the business response success rate includes: the first HTTP response time (FHRT), the last packet time, and the TCP ACK time of the transmission control protocol response; The processing module is used to input the at least one indicator parameter into a service quality assessment model for processing, to obtain a target service quality level corresponding to the service response success rate. The target service quality level is one of multiple service quality levels, which are obtained by adaptively classifying service quality levels based on the service response success rate using a preset algorithm. Specifically, the number of clusters is adaptively determined using the Xie_Beni coefficient, an effectiveness indicator based on cluster compactness and separation, to adaptively determine how many levels of service quality to classify. A time series sample set consisting of multiple target service quality levels contained within a sliding time window is obtained. The time series sample set is then input into a service quality prediction model for processing to obtain the service quality level at the next time step.

8. A computing device, comprising: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform an operation corresponding to the service quality prediction method as described in any one of claims 1-6.

9. A computer storage medium storing at least one executable instruction that causes a processor to perform an operation corresponding to the service quality prediction method as described in any one of claims 1-6.

Images