CN118018429A - Network service satisfaction analysis method, system, equipment and storage medium - Google Patents
Network service satisfaction analysis method, system, equipment and storage medium Download PDFInfo
- Publication number
- CN118018429A CN118018429A CN202410239037.3A CN202410239037A CN118018429A CN 118018429 A CN118018429 A CN 118018429A CN 202410239037 A CN202410239037 A CN 202410239037A CN 118018429 A CN118018429 A CN 118018429A
- Authority
- CN
- China
- Prior art keywords
- network service
- node
- matrix
- data matrix
- service satisfaction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 45
- 239000011159 matrix material Substances 0.000 claims abstract description 149
- 238000000034 method Methods 0.000 claims abstract description 43
- 238000012549 training Methods 0.000 claims description 25
- 238000004891 communication Methods 0.000 claims description 14
- 238000011478 gradient descent method Methods 0.000 claims description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 14
- 238000005457 optimization Methods 0.000 abstract description 7
- 230000006870 function Effects 0.000 description 24
- 230000008569 process Effects 0.000 description 14
- 238000005516 engineering process Methods 0.000 description 13
- 238000012545 processing Methods 0.000 description 11
- 230000001364 causal effect Effects 0.000 description 10
- 238000007476 Maximum Likelihood Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 5
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 230000008602 contraction Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011438 discrete method Methods 0.000 description 1
- 238000005315 distribution function Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/12—Discovery or management of network topologies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/142—Network analysis or design using statistical or mathematical methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Mathematical Optimization (AREA)
- Algebra (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application provides a network service satisfaction analysis method, a system, equipment and a storage medium, belonging to the technical field of artificial intelligence. According to the method, a node topological order is estimated through a maximum loop-free sub-graph algorithm and least square loss, network topology optimization efficiency is improved, a first data matrix output by a pre-trained Bayesian network model is ordered according to the estimated node order, and a weighted adjacent matrix of the model is estimated and optimized by using the ordered first data matrix.
Description
Technical Field
The present application relates to the field of artificial intelligence technologies, and in particular, to a method, a system, an apparatus, and a storage medium for analyzing network service satisfaction.
Background
With the development of internet technology, users have higher requirements on network communication quality, operators need to survey the satisfaction degree of network services for users in order to serve the users, the service quality is improved based on the network service satisfaction degree of the users, and factors affecting the network service satisfaction degree of the users are many, besides the network quality, the factors may be related to various factors such as network charging standards, accident handling efficiency and the like, so that the network service satisfaction degree needs to be analyzed by combining various factors, and the service quality can be improved more pertinently. At present, as for the causal relationship analysis mode, a bayesian network structure learning method can be adopted, and because the DAG (DIRECTED ACYCLIC GRAPH ) represents the structure of the bayesian network, the bayesian network structure learning is also called DAG learning, the DAG learning is an NP-hard problem, and the variable dimension related to the network service satisfaction degree is numerous, so that the network service satisfaction degree analysis efficiency is low.
Disclosure of Invention
The embodiment of the application mainly aims to provide a network service satisfaction analysis method, a system, equipment and a storage medium, aiming at improving the efficiency and accuracy of network service satisfaction analysis.
In order to achieve the above objective, an aspect of an embodiment of the present application provides a network service satisfaction analysis method, including the following steps:
Inputting the known satisfaction influence factor variable values into a pre-trained Bayesian network model to obtain a first data matrix, wherein the first data matrix comprises network service satisfaction variable values and various satisfaction influence factor variable values;
Iterating the node topology sequence of the first data matrix through least square loss and a maximum loop-free sub-graph algorithm to obtain an updated node topology sequence;
re-ordering the first data matrix according to the updated node topological order to obtain a second data matrix;
Performing Lasso estimation on non-zero elements of a weighted adjacent matrix of the Bayesian network model according to the second data matrix to obtain element weight values of the weighted adjacent matrix;
And constructing a directed acyclic graph of the satisfaction degree of the network service according to the weighted adjacency matrix.
In some embodiments, the bayesian network model is pre-trained by:
Acquiring a pre-training data set, wherein the pre-training data set comprises sample data of a plurality of users, and the sample data comprises network service satisfaction, user information and network information;
and carrying out structure learning and parameter learning of the Bayesian network according to the pre-training data set to obtain a pre-trained Bayesian network model.
In some embodiments, the performing structure learning and parameter learning of the bayesian network according to the pre-training data set to obtain a pre-trained bayesian network model includes the following steps:
Determining a plurality of variables and an order of the plurality of variables in the network according to the pre-training data set, wherein the plurality of variables comprise a network service satisfaction variable and a plurality of satisfaction influence factor variables;
initializing a Bayesian network structure by taking a plurality of variables as network nodes;
sequentially selecting one variable as a current node according to the sequence of the variables;
Constructing a plurality of candidate father node sets of the current node;
Determining a conditional probability distribution between the current node and the candidate parent node set according to the pre-training data set;
determining model scores under different candidate parent node sets according to conditional probability distribution between the current node and each candidate parent node set, and determining an optimal parent node set by each model score;
and updating the leaf network structure according to the optimal father node set to obtain a Bayesian network model.
In some embodiments, the iteration of the node topology sequence of the first data matrix through the least squares loss and the maximum round-robin sub-graph algorithm to obtain an updated node topology sequence includes the following steps:
determining an initial node topological order of the first data matrix according to a maximum loop-free sub-graph algorithm;
Determining the least square loss of the node topological sequence according to the real satisfaction influence factor variable value, and updating the node topological sequence according to the least square loss by adopting a gradient descent method;
and repeatedly executing the step of determining the least square loss of the node topological sequence according to the sample data, and updating the node topological sequence according to the least square loss by adopting a gradient descent method until the least square loss is smaller than a preset value.
In some embodiments, the performing Lasso estimation on the non-zero elements of the weighted adjacency matrix of the bayesian network model according to the second data matrix to obtain element weight values of the weighted adjacency matrix includes the following steps:
extracting an upper triangle part from a weighted adjacency matrix of the Bayesian network model;
And performing Lasso estimation on the non-zero elements of the upper triangle part according to the second data matrix to obtain corresponding element weight values so as to update the element weight values of the weighted adjacent matrix.
In some embodiments, the performing Lasso estimation on the non-zero element of the upper triangle part according to the second data matrix obtains a corresponding element weight value, which includes the following steps:
determining a first weight of a non-zero element in the current column of the upper triangle part according to the initial values of the second data matrix and the weighted adjacency matrix based on a Lasso algorithm;
performing Lasso self-adaptive constraint on the first weight of the element according to a preset weight constraint formula to obtain a second weight of the element;
And determining element weight values according to the second data matrix, the second weight and the tuning parameters.
In some embodiments, before the step of constructing a directed acyclic graph of network service satisfaction from the weighted adjacency matrix, the network service satisfaction analysis method further comprises the steps of:
And reordering the weighted adjacency matrix in reverse order according to the updated node topology sequence.
To achieve the above object, another aspect of the embodiments of the present application provides a network service satisfaction analysis system, including:
the first module is used for inputting the known satisfaction influence factor variable values into a pre-trained Bayesian network model to obtain a first data matrix, wherein the first data matrix comprises network service satisfaction variable values and various satisfaction influence factor variable values;
The second module is used for iterating the node topological sequence of the first data matrix through the least square loss and the maximum loop-free sub-graph algorithm to obtain an updated node topological sequence;
a third module, configured to reorder the first data matrix according to the updated node topology sequence to obtain a second data matrix;
A fourth module, configured to perform Lasso estimation on non-zero elements of a weighted adjacency matrix of the bayesian network model according to the second data matrix, so as to obtain element weight values of the weighted adjacency matrix;
and a fifth module, configured to construct a directed acyclic graph of network service satisfaction according to the weighted adjacency matrix.
To achieve the above object, another aspect of the embodiments of the present application provides an electronic device including a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for implementing connection communication between the processor and the memory, the program implementing the network service satisfaction analysis method described in the above embodiments when executed by the processor.
To achieve the above object, another aspect of the embodiments of the present application provides a storage medium, which is a computer-readable storage medium, for computer-readable storage, where one or more programs are stored, and the one or more programs are executable by one or more processors to implement the network service satisfaction analysis method described in the above embodiments.
According to the network service satisfaction analysis method, system, equipment and storage medium, a node topological order is estimated through a maximum loop-free sub-graph algorithm and least square loss, network topology optimization efficiency is improved, a first data matrix output by a pre-trained Bayesian network model is ordered according to the estimated node order, and the weighted adjacent matrix of the model is estimated and optimized by using the ordered first data matrix.
Drawings
FIG. 1 is a flow chart of a network service satisfaction analysis method provided by an embodiment of the present application;
FIG. 2 is a flow chart of a Bayesian network model pre-training method in step S101 in FIG. 1;
fig. 3 is a flowchart of step S202 in fig. 2;
fig. 4 is a flowchart of step S102 in fig. 1;
fig. 5 is a flowchart of step S104 in fig. 1;
fig. 6 is a flowchart of step S502 in fig. 5;
FIG. 7 is a flow chart of a network service satisfaction analysis method provided by another embodiment of the present application;
FIG. 8 is a schematic diagram of a network service satisfaction analysis system according to an embodiment of the present application;
fig. 9 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present application;
Fig. 10 is a flowchart of a weighted adjacency matrix correction process according to an embodiment of the present application.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
It should be noted that although functional block diagrams are depicted as block diagrams, and logical sequences are shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the block diagrams in the system. The terms first, second and the like in the description and in the claims and in the above-described figures, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
Unless defined otherwise, 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 the application only and is not intended to be limiting of the application.
First, several nouns involved in the present application are parsed:
Artificial intelligence (ARTIFICIAL INTELLIGENCE, AI): is a new technical science for researching and developing theories, methods, technologies and application systems for simulating, extending and expanding the intelligence of people; artificial intelligence is a branch of computer science that attempts to understand the nature of intelligence and to produce a new intelligent machine that can react in a manner similar to human intelligence, research in this field including robotics, language recognition, image recognition, natural language processing, and expert systems. Artificial intelligence can simulate the information process of consciousness and thinking of people. Artificial intelligence is also a theory, method, technique, and application system that utilizes a digital computer or digital computer-controlled machine to simulate, extend, and expand human intelligence, sense the environment, acquire knowledge, and use knowledge to obtain optimal results.
A Bayesian network (Bayesian network), also known as a belief network (belief network), or directed acyclic graph model (DIRECTED ACYCLIC GRAPHICAL model, DAG), is a model of probability patterns. The directed acyclic graph is composed of directed edges representing variable nodes and connecting nodes. The nodes represent random variables, the directed edges among the nodes represent the interrelationships among the nodes (the father nodes point to the child nodes of the nodes), the relation strength is expressed by using conditional probability, and the information expression is carried out by using prior probability without the father nodes. The association relationship between nodes in the directed acyclic graph is called a node topology sequence.
Maximum likelihood estimation (Maximum Likelihood Estimation, MLE) is a parameter estimation method commonly used in statistics, the basic idea of which is to find a parameter value that maximizes the probability of the occurrence of sample data given the sample data. In bayesian network structure learning, maximum likelihood estimates may be used to determine parameters of the network structure.
Adaptive Lasso (minimum absolute contraction and selection operator) algorithm: the method is a process of automatically adjusting a processing method, a processing sequence, processing parameters, boundary conditions or constraint conditions according to the data characteristics of the processed data in the processing and analyzing processes to adapt to the statistical distribution characteristics and the structural characteristics of the processed data so as to obtain the optimal processing effect. The Lasso algorithm is a linear model for regression analysis, and the complexity of the model is constrained by adding a regularization term (i.e., L1 norm) to the loss function, so as to realize variable selection and parameter sparsification. In Lasso regression, regularization terms will encourage the model to use fewer predicted variables and accurately reduce the estimates of some coefficients to zero.
The maximum loop-free sub-graph algorithm, also known as the maximum loop-free sub-graph problem, aims to find the maximum loop-free sub-graph in a given directed graph, which contains the most nodes in the original graph. In other words, it is a largest subset of the original graph, where no loops exist between any two nodes.
The least squares loss (Least Squares Loss) is a loss function used in the least squares method to measure the difference between the model predictions and the actual observations. The goal of the least squares method is to minimize this loss function to find the best model parameters. In particular, the least squares loss function is generally defined as the sum of squared differences between the predicted value and the actual observed value. For a linear regression model, the least squares loss function can be expressed as: l (β) =Σ (Y-y_hat) ≡2; where Y represents the actual observations, Y_hat represents the model predictions, β is the model parameter vector, and Σ represents summing all observations. An intuitive interpretation of the least squares loss function is that it calculates the sum of squares of the "error" between the model predicted value and the actual observed value. By minimizing this loss function, a set of model parameters can be found such that the predicted value is as close as possible to the actual observed value, resulting in an optimal model fitting effect. The least squares Loss Function is a kind of Loss Function (Loss Function) used to measure the quality of the model. In machine learning and statistics, a loss function is typically used to optimize model parameters so that the model performs as well as possible on training data. Besides the least squares loss function, there are other common loss functions, such as cross entropy loss functions, mean square error loss functions, etc., which are applicable to different models and tasks, respectively. It should be noted that when using the least squares method for parameter estimation, some assumption conditions, such as the expectation of the error term being zero, the variances being the same and not being related to each other, etc., need to be satisfied. If these assumptions do not hold, the estimated result of the least squares method may be affected by deviations or instabilities. Therefore, in practical applications, appropriate preprocessing and inspection of the data is required to ensure that the least squares method can obtain reliable parameter estimation results.
With the advent of the big data age and the rise of artificial intelligence, bayesian networks became one of the most effective models for dealing with the problem of uncertainty reasoning. In the field of network service satisfaction analysis, more attention is required to causal relationships of network service satisfaction than to purely statistical correlations. Under causal bayesian network semantics, DAGs can be used to represent causal relationships between random variables in a graph model, and studies indicate that DAG structure learning is an NP-hard problem. Furthermore, two different DAGs may be markov equivalent, which also makes DAG learning difficult. Nevertheless, since DAG learning is a key element in discovering data generation mechanisms or discovering causal relationships, it is suitable for reasoning about network service satisfaction.
In the Bayesian network structure learning method, a learning method with high accuracy is easily affected by dimensions, and some methods which are not easily affected by dimensions or can process a high-dimensional structure have poor performance in accuracy, so that it is difficult to simultaneously compatible with analysis accuracy and computational complexity in reasoning of network service satisfaction causality.
Based on the above, the embodiment of the application provides a network service satisfaction analysis method, a system, a device and a storage medium, aiming at improving the efficiency and the accuracy of network service satisfaction analysis.
The method, the system, the device and the storage medium for analyzing the satisfaction of the network service provided by the embodiment of the application are specifically described through the following embodiments, and the method for analyzing the satisfaction of the network service in the embodiment of the application is described first.
The embodiment of the application can acquire and process the related data based on the artificial intelligence technology. Wherein artificial intelligence (ARTIFICIAL INTELLIGENCE, AI) is the theory, method, technique, and application system that uses a digital computer or a digital computer-controlled machine to simulate, extend, and expand human intelligence, sense the environment, acquire knowledge, and use knowledge to obtain optimal results.
Artificial intelligence infrastructure technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a robot technology, a biological recognition technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and other directions.
The embodiment of the application provides a network service satisfaction analysis method, which relates to the technical field of artificial intelligence. The network service satisfaction analysis method provided by the embodiment of the application can be applied to the terminal, can be applied to the server side, and can also be software running in the terminal or the server side. In some embodiments, the terminal may be a smart phone, tablet, notebook, desktop, etc.; the server side can be configured as an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, and a cloud server for providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, basic cloud computing services such as big data and artificial intelligent platforms and the like; the software may be an application or the like that implements the network service satisfaction analysis method, but is not limited to the above form.
The application is operational with numerous general purpose or special purpose computer system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like. The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
In the embodiments of the present application, when related processing is performed according to user information, user behavior data, user history data, user location information, and other data related to user identity or characteristics, permission or consent of the user is obtained first, and the collection, use, processing, and the like of the data comply with related laws and regulations and standards of related countries and regions. In addition, when the embodiment of the application needs to acquire the sensitive personal information of the user, the independent permission or independent consent of the user is acquired through popup or jump to a confirmation page and the like, and after the independent permission or independent consent of the user is definitely acquired, the necessary relevant data of the user for enabling the embodiment of the application to normally operate is acquired.
Fig. 1 is an optional flowchart of a network service satisfaction analysis method according to an embodiment of the present application, where the method in fig. 1 may include, but is not limited to, steps S101 to S105.
Step S101, inputting a known satisfaction influence factor variable value into a pre-trained Bayesian network model to obtain a first data matrix, wherein the first data matrix comprises network service satisfaction variable values and various satisfaction influence factor variable values;
Step S102, iterating the node topological sequence of the first data matrix through least square loss and a maximum loop-free sub-graph algorithm to obtain an updated node topological sequence;
step S103, reordering the first data matrix according to the updated node topology sequence to obtain a second data matrix;
Step S104, performing Lasso estimation on non-zero elements of a weighted adjacent matrix of the Bayesian network model according to the second data matrix to obtain element weight values of the weighted adjacent matrix;
Step S105, constructing a directed acyclic graph of network service satisfaction according to the weighted adjacency matrix.
In the steps S101 to S105 shown in the embodiment of the application, a node topological order is estimated through a maximum loop-free sub-graph algorithm and least square loss, the network topology optimization efficiency is improved, a first data matrix output by a pre-trained Bayesian network model is ordered according to the estimated node order, and the weighted adjacent matrix of the model is estimated and optimized by using the ordered first data matrix.
In step S101 of some embodiments, the pre-trained bayesian network model refers to a model that is trained based on a given training data set to obtain a conditional probability between variables in the characterization network, where the bayesian network model may be represented by a weighted adjacency matrix, and where the element weight values of the weighted adjacency matrix at the (i, j) locations may represent the probability that the variable j occurs if the variable j has occurred at the variable i. The pre-trained Bayesian network model can be constructed in advance based on the general data set, the operation amount of the causal relationship of the network service satisfaction can be reduced to a certain extent by using the pre-trained Bayesian network model, but the pre-trained Bayesian network model is usually under the condition of over fitting, so that the model generalization is low, and the accuracy of the output analysis result is poor. In this embodiment, a first data matrix is predicted from a bayesian network model with a known satisfaction impact factor variable value based on a pre-trained bayesian network model, the first data matrix characterizes each network service satisfaction variable value and each satisfaction impact factor variable value, and a weighted adjacency matrix is corrected based on the first data matrix, so that the weighted adjacency matrix can accurately characterize the satisfaction and the causal relationship between the satisfaction and the impact factors. It is understood that satisfaction impact factors may include, but are not limited to, web speed, package price, user occupation, network stability, and the like.
Referring to fig. 2, in some embodiments, the bayesian network model in step S101 may be pre-trained by:
Step S201, a pre-training data set is obtained, wherein the pre-training data set comprises sample data of a plurality of users, and the sample data comprises network service satisfaction, user information and network information;
Step S202, structure learning and parameter learning of the Bayesian network are carried out according to the pre-training data set, and a pre-trained Bayesian network model is obtained.
In this embodiment, the pre-training process of the bayesian network can be divided into two main steps: structure learning and parameter learning. The goal of structure learning is to learn the structure of the directed acyclic graph from a given dataset, i.e., to determine the dependency between variables, which can be accomplished by different algorithms, such as the PC algorithm, the K2 algorithm, the BIC algorithm, the Hill-Climbing algorithm, and so forth. The algorithm builds the network structure step by step based on a conditional independence test, bayesian theorem, maximum a posteriori probability principle, or enumerating all possible structures. After determining the structure of the bayesian network, an estimate of the conditional probability between each node is required. Common parameter learning algorithms include maximum likelihood estimation, bayesian estimation, and the like. The embodiment of the application can use a maximum likelihood estimation algorithm to perform parameter learning, and estimate the parameters by maximizing likelihood functions so as to maximize the probability of occurrence of sample data under given parameters.
Referring to fig. 3, in some embodiments, in step S202, the steps of performing structure learning and parameter learning of the bayesian network according to the pre-training data set to obtain a pre-trained bayesian network model may include, but is not limited to, the following steps:
Step S301, determining a plurality of variables and the sequence of the plurality of variables in the network according to the pre-training data set, wherein the plurality of variables comprise a network service satisfaction variable and a plurality of satisfaction influence factor variables;
Step S302, a plurality of variables are used as network nodes, and a Bayesian network structure is initialized;
Step S303, sequentially selecting one variable as a current node according to the sequence of the variables;
Step S304, constructing a plurality of candidate father node sets of the current node;
step S305, determining a conditional probability distribution between the current node and the candidate father node set according to the pre-training data set;
step S306, determining model scores under different candidate father node sets according to conditional probability distribution between the current node and each candidate father node set, and determining an optimal father node set by each model score;
step S307, the Bayesian network structure is updated according to the optimal father node set, and a Bayesian network model is obtained.
In this embodiment, the pre-training data set may be subjected to data analysis in combination with a priori knowledge, a plurality of network service related variables may be selected from the pre-training data set according to the data analysis structure, and the variable order may be primarily determined according to the network service correlations. Initializing the network, wherein no node and edge exist in the initialized network, all variables are regarded as root nodes without any father node, and the conditional probability of the node can be expressed by prior probability. According to the variable sequence, selecting a variable node as the current node to gradually learn parameters. For the current node, consider all possible parent node combinations thereof, resulting in a candidate parent node set. The conditional probability distribution between the current node and its set of candidate parent nodes is evaluated using maximum likelihood estimation, i.e. the parameters are estimated using existing data sets. And selecting one candidate parent node set with the maximum score of the model as the parent node set of the current node according to the evaluation criterion. Adding the current node to its optimal parent node set and updating the overall network structure, including adding new edges in the network to represent dependencies between nodes. The next node is selected and the process is repeated until all nodes have been added to the network.
In step S102 of some embodiments, the least squares loss and maximum circle less sub-graph algorithm may be referred to as a NOE algorithm, which utilizes the least squares loss and the maximum circle less sub-graph for node order estimation. The optimization function used by the NOE algorithm has no complex calculation in the optimization process, so that the process efficiency of obtaining the circled solution is very high, and the circled solution can be obtained quickly even under the high-dimensional condition.
Referring to fig. 4, in some embodiments, in step S102, the step of iterating the node topology sequence of the first data matrix through the least squares loss and the maximum loop-free sub-graph algorithm to obtain an updated node topology sequence may include, but is not limited to, the following steps:
Step S401, determining an initial node topological order of a first data matrix according to a maximum loop-free sub-graph algorithm;
Step S402, determining the least square loss of the node topological sequence according to the real satisfaction influence factor variable value, and updating the node topological sequence according to the least square loss by adopting a gradient descent method;
step S403, repeating step S402 until the least squares loss is smaller than a preset value.
In this embodiment, the maximum circle sub-graph algorithm is used to extract an initial node topology sequence from the first data matrix, where the node topology sequence is a linear sequence that characterizes causal relationships between network service satisfaction variables and related impact factors. And determining the least square loss of the node topological sequence according to the real satisfaction influence factor variable value, adopting a gradient descent method and updating the node topological sequence according to the calculated least square loss, thereby obtaining more accurate causal relationship. The topological order is updated through least square estimation and a maximum loop-free sub-graph algorithm, so that the node topological order is prevented from being solved by a discrete method, and the applicability in a large Bayesian network is improved.
In step S103 of some embodiments, the first data matrix is reordered according to the node topology order, that is, the node topology order performs column-row transformation on the first data matrix to obtain a new data matrix, and by ordering the first data matrix, the weighted adjacency matrix can be conveniently and more efficiently estimated later.
In step S104 of some embodiments, after reordering the first data matrix to obtain a second data matrix, lasso estimation is performed on the upper triangle part of the weighted adjacency matrix of the bayesian network model using the second data matrix to obtain the element weight value. The solving process of the element weight value mainly comprises two parts, wherein the first part is to solve the weight value of the self-adaptive Lasso by using a Lasso-based maximum likelihood function; the second part is based on the first part, and the directed acyclic graph learning is performed by using a maximum likelihood function based on the adaptive Lasso. The maximum likelihood estimation based on Lasso can restrict model parameters through the L1 regularization term while maximizing likelihood functions, so that the estimated parameters can be sparse as much as possible while meeting data fitting, namely, a plurality of parameter estimated values can be zero, the complexity of the model can be reduced, the generalization capability of the model is improved, and the risk of overfitting is reduced.
Referring to fig. 5, in some embodiments, in step S104, the step of performing Lasso estimation on non-zero elements of the weighted adjacency matrix of the bayesian network model according to the second data matrix to obtain element weight values of the weighted adjacency matrix may include, but is not limited to, the following steps:
Step S501, extracting an upper triangle part from a weighted adjacency matrix of the Bayesian network model;
Step S502, performing Lasso estimation on the non-zero elements of the upper triangle part according to the second data matrix to obtain corresponding element weight values so as to update the element weight values of the weighted adjacent matrix.
In this embodiment, the weighted adjacency matrix is a square matrix, i.e. the number of rows and columns are equal, and are equal to the number of vertices in the figure. The upper triangular portion of the matrix refers to the portion above the main diagonal, i.e., the elements of row i and column j, where i < j. And performing Lasso estimation on the non-zero elements of the upper triangle part according to the second data matrix, so as to realize continuous optimization and improve the DAG learning efficiency.
Referring to fig. 6, in some embodiments, in step S502, the step of performing Lasso estimation on the non-zero elements of the upper triangle part according to the second data matrix to obtain the corresponding element weight values may include, but is not limited to, the following steps:
Step S601, determining a first weight of a non-zero element in the current column of the upper triangle part according to initial values of a second data matrix and a weighted adjacent matrix based on a Lasso algorithm;
Step S602, performing Lasso self-adaptive constraint on a first weight of an element according to a preset weight constraint formula to obtain a second weight of the element;
step S603, determining element weight values according to the second data matrix, the second weight and the tuning parameters.
In this embodiment, the initial value is used to solve Lasso estimates for elements other than 0 in the current column of the upper triangle portion of the weighted adjacency matrixI.e. the first weight, is as follows:
Wherein, Representing a second data matrix,/>Elements representing the ith row and jth column of the second data matrix, W representing the weighted adjacency matrix.
Performing Lasso self-adaptive constraint on the first weight of the element according to a preset weight constraint formula to obtain a second weight of the element, wherein the weight constraint formula is as follows:
Wherein, Representing the second weight,/>L n =min (1/N ), N can take the value 104,A threshold representing the upper bound of the weight disappears when n goes to infinity.
The tuning parameters are calculated as follows:
Wherein alpha is the significance level, Is a standard normal distribution function.
According to the calculated self-adaptive weight and tuning parameters, solving the self-adaptive Lasso estimation to obtain a final element weight value, wherein the final element weight value is as follows:
referring to fig. 7, in some embodiments, before step S105, the network service satisfaction analysis method according to the embodiment of the present application may further include, but is not limited to, the following steps:
Step S701, the weighted adjacency matrix is reordered in reverse order according to the updated node topology order.
In this embodiment, the first data matrix is rearranged according to the node topology sequence to obtain the second data matrix, and when the weighted adjacency matrix is subjected to weight estimation by using the second data matrix, the order of the weighted adjacency matrix is changed, that is, the node order of the weighted adjacency matrix estimated by Lasso is different from that of the original weighted adjacency matrix, so that the weighted adjacency matrix obtained in step S104 needs to be subjected to row-column transformation according to the node order of the updated node topology sequence, so that the node order of the weighted adjacency matrix is consistent with that of the first data matrix.
In some embodiments of step S105, a directed acyclic graph is drawn according to the modified weighted adjacency matrix, so that the causal relationship between the user network service satisfaction correlation factors can be clearly shown, and the network service correlation factors can be improved pertinently, so that the network service can be better provided to the user.
Referring to fig. 10, a process of correcting a weighted adjacency matrix according to an embodiment of the present application is described as follows:
Step S10, starting a NOE algorithm, and iterating by using a least square loss and a maximum loop-free sub-graph algorithm to obtain node topological order estimation of the DAG;
Step S20, reordering the original data matrix output by the pre-training model by using the node topological order obtained by the NOE algorithm;
Step S30, estimating elements of each column of the weighted adjacency matrix that may not be zero by using the ordered data matrix, specifically including:
step S31, defining a loss function estimated by Lasso, and setting an optimizer parameter;
step S32, calculating element weight values by using a loss function estimated by Lasso;
Step S33, parameter tuning is carried out by using an optimizer, and element weight values estimated by the self-adaptive Lasso are obtained;
step S40, reordering and restoring the weighted adjacency matrix according to the node topological order.
Referring to fig. 8, an embodiment of the present application further provides a system for analyzing satisfaction of network services, including:
The first module is used for inputting the known satisfaction influence factor variable values into a pre-trained Bayesian network model to obtain a first data matrix, wherein the first data matrix comprises network service satisfaction variable values and various satisfaction influence factor variable values;
The second module is used for iterating the node topological sequence of the first data matrix through the least square loss and the maximum loop-free sub-graph algorithm to obtain an updated node topological sequence;
A third module, configured to reorder the first data matrix according to the updated node topology sequence to obtain a second data matrix;
A fourth module, configured to perform Lasso estimation on non-zero elements of a weighted adjacency matrix of the bayesian network model according to the second data matrix, so as to obtain element weight values of the weighted adjacency matrix;
And a fifth module for constructing a directed acyclic graph of network service satisfaction according to the weighted adjacency matrix.
It can be understood that the content in the above embodiment of the network service satisfaction analysis method is applicable to the embodiment of the present system, and the functions specifically implemented by the embodiment of the present system are the same as those in the embodiment of the above embodiment of the network service satisfaction analysis method, and the beneficial effects achieved by the embodiment of the present system are the same as those achieved by the embodiment of the above embodiment of the network service satisfaction analysis method.
The embodiment of the application also provides electronic equipment, which comprises: the network service satisfaction analysis system comprises a memory, a processor, a program stored on the memory and capable of running on the processor, and a data bus for realizing connection communication between the processor and the memory, wherein the program is executed by the processor to realize the network service satisfaction analysis method. The electronic equipment can be any intelligent terminal including a tablet personal computer, a vehicle-mounted computer and the like.
Referring to fig. 9, fig. 9 illustrates a hardware structure of an electronic device according to another embodiment, the electronic device includes:
The processor 901 may be implemented by a general purpose CPU (central processing unit), a microprocessor, an application specific integrated circuit (ApplicationSpecificIntegratedCircuit, ASIC), or one or more integrated circuits, etc. for executing related programs, so as to implement the technical solution provided by the embodiments of the present application;
The memory 902 may be implemented in the form of read-only memory (ReadOnlyMemory, ROM), static storage, dynamic storage, or random access memory (RandomAccessMemory, RAM), among others. The memory 902 may store an operating system and other application programs, and when the technical solutions provided in the embodiments of the present disclosure are implemented by software or firmware, relevant program codes are stored in the memory 902, and the processor 901 invokes the network service satisfaction analysis method for executing the embodiments of the present disclosure;
An input/output interface 903 for inputting and outputting information;
The communication interface 904 is configured to implement communication interaction between the device and other devices, and may implement communication in a wired manner (e.g. USB, network cable, etc.), or may implement communication in a wireless manner (e.g. mobile network, WIFI, bluetooth, etc.);
a bus 905 that transfers information between the various components of the device (e.g., the processor 901, the memory 902, the input/output interface 903, and the communication interface 904);
wherein the processor 901, the memory 902, the input/output interface 903 and the communication interface 904 are communicatively coupled to each other within the device via a bus 905.
The embodiment of the application also provides a storage medium, which is a computer readable storage medium and is used for computer readable storage, the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to realize the network service satisfaction analysis method.
The memory, as a non-transitory computer readable storage medium, may be used to store non-transitory software programs as well as non-transitory computer executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory optionally includes memory remotely located relative to the processor, the remote memory being connectable to the processor through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The network service satisfaction analysis method, system, equipment and storage medium provided by the embodiment of the application estimate a node topological order through a maximum loop-free sub-graph algorithm and least square loss, improve network topology optimization efficiency, sort a first data matrix output by a pre-trained Bayesian network model according to the estimated node order, and estimate and optimize a weighted adjacent matrix of the model by using the sorted first data matrix. .
The embodiments described in the embodiments of the present application are for more clearly describing the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application, and those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.
It will be appreciated by persons skilled in the art that the embodiments of the application are not limited by the illustrations, and that more or fewer steps than those shown may be included, or certain steps may be combined, or different steps may be included.
The system embodiments described above are merely illustrative, in that the units illustrated as separate components may or may not be physically separate, i.e., may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
Those of ordinary skill in the art will appreciate that all or some of the steps of the methods, systems, functional modules/units in the devices disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof.
The terms "first," "second," "third," "fourth," and the like in the description of the application and in the above figures, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the application described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It should be understood that in the present application, "at least one (item)" means one or more, and "a plurality" means two or more. "and/or" for describing the association relationship of the association object, the representation may have three relationships, for example, "a and/or B" may represent: only a, only B and both a and B are present, wherein a, B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of" or the like means any combination of these items, including any combination of single item(s) or plural items(s). For example, at least one (one) of a, b or c may represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, c may be single or plural.
In the several embodiments provided by the present application, it should be understood that the disclosed systems and methods may be implemented in other ways. For example, the system embodiments described above are merely illustrative, e.g., the division of the above elements is merely a logical functional division, and there may be additional divisions in actual implementation, e.g., multiple elements or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some interface, system or unit indirect coupling or communication connection, which may be in electrical, mechanical or other form.
The units described above as separate components may or may not be physically separate, and components shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including multiple instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method of the various embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory RAM), a magnetic disk, or an optical disk, or other various media capable of storing a program.
The preferred embodiments of the present application have been described above with reference to the accompanying drawings, and are not thereby limiting the scope of the claims of the embodiments of the present application. Any modifications, equivalent substitutions and improvements made by those skilled in the art without departing from the scope and spirit of the embodiments of the present application shall fall within the scope of the claims of the embodiments of the present application.
Claims (10)
1. A network service satisfaction analysis method, comprising the steps of:
Inputting the known satisfaction influence factor variable values into a pre-trained Bayesian network model to obtain a first data matrix, wherein the first data matrix comprises network service satisfaction variable values and various satisfaction influence factor variable values;
Iterating the node topology sequence of the first data matrix through least square loss and a maximum loop-free sub-graph algorithm to obtain an updated node topology sequence;
re-ordering the first data matrix according to the updated node topological order to obtain a second data matrix;
Performing Lasso estimation on non-zero elements of a weighted adjacent matrix of the Bayesian network model according to the second data matrix to obtain element weight values of the weighted adjacent matrix;
And constructing a directed acyclic graph of the satisfaction degree of the network service according to the weighted adjacency matrix.
2. The network service satisfaction analysis method of claim 1 wherein said bayesian network model is pre-trained by:
Acquiring a pre-training data set, wherein the pre-training data set comprises sample data of a plurality of users, and the sample data comprises network service satisfaction, user information and network information;
and carrying out structure learning and parameter learning of the Bayesian network according to the pre-training data set to obtain a pre-trained Bayesian network model.
3. The network service satisfaction analysis method according to claim 2, wherein the performing structure learning and parameter learning of the bayesian network according to the pre-training data set to obtain a pre-trained bayesian network model comprises the following steps:
Determining a plurality of variables and an order of the plurality of variables in the network according to the pre-training data set, wherein the plurality of variables comprise a network service satisfaction variable and a plurality of satisfaction influence factor variables;
initializing a Bayesian network structure by taking a plurality of variables as network nodes;
sequentially selecting one variable as a current node according to the sequence of the variables;
Constructing a plurality of candidate father node sets of the current node;
Determining a conditional probability distribution between the current node and the candidate parent node set according to the pre-training data set;
determining model scores under different candidate parent node sets according to conditional probability distribution between the current node and each candidate parent node set, and determining an optimal parent node set by each model score;
and updating the leaf network structure according to the optimal father node set to obtain a Bayesian network model.
4. The network service satisfaction analysis method according to claim 1, wherein the node topology sequence of the first data matrix is iterated through least squares loss and a maximum loop-free sub-graph algorithm to obtain an updated node topology sequence, and the method comprises the following steps:
determining an initial node topological order of the first data matrix according to a maximum loop-free sub-graph algorithm;
Determining the least square loss of the node topological sequence according to the real satisfaction influence factor variable value, and updating the node topological sequence according to the least square loss by adopting a gradient descent method;
and repeatedly executing the step of determining the least square loss of the node topological sequence according to the sample data, and updating the node topological sequence according to the least square loss by adopting a gradient descent method until the least square loss is smaller than a preset value.
5. The network service satisfaction analysis method according to claim 1, wherein the step of performing Lasso estimation on non-zero elements of a weighted adjacency matrix of a bayesian network model according to the second data matrix to obtain element weight values of the weighted adjacency matrix comprises the steps of:
extracting an upper triangle part from a weighted adjacency matrix of the Bayesian network model;
And performing Lasso estimation on the non-zero elements of the upper triangle part according to the second data matrix to obtain corresponding element weight values so as to update the element weight values of the weighted adjacent matrix.
6. The network service satisfaction analysis method according to claim 5, wherein the performing Lasso estimation on the non-zero element of the upper triangle part according to the second data matrix obtains a corresponding element weight value, and the method comprises the following steps:
determining a first weight of a non-zero element in the current column of the upper triangle part according to the initial values of the second data matrix and the weighted adjacency matrix based on a Lasso algorithm;
performing Lasso self-adaptive constraint on the first weight of the element according to a preset weight constraint formula to obtain a second weight of the element;
And determining element weight values according to the second data matrix, the second weight and the tuning parameters.
7. The network service satisfaction analysis method of claim 1 wherein prior to said step of constructing a directed acyclic graph of network service satisfaction from said weighted adjacency matrix, said network service satisfaction analysis method further comprises the step of:
And reordering the weighted adjacency matrix in reverse order according to the updated node topology sequence.
8. A network service satisfaction analysis system, comprising:
the first module is used for inputting the known satisfaction influence factor variable values into a pre-trained Bayesian network model to obtain a first data matrix, wherein the first data matrix comprises network service satisfaction variable values and various satisfaction influence factor variable values;
The second module is used for iterating the node topological sequence of the first data matrix through the least square loss and the maximum loop-free sub-graph algorithm to obtain an updated node topological sequence;
a third module, configured to reorder the first data matrix according to the updated node topology sequence to obtain a second data matrix;
A fourth module, configured to perform Lasso estimation on non-zero elements of a weighted adjacency matrix of the bayesian network model according to the second data matrix, so as to obtain element weight values of the weighted adjacency matrix;
and a fifth module, configured to construct a directed acyclic graph of network service satisfaction according to the weighted adjacency matrix.
9. An electronic device comprising a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for enabling a connection communication between the processor and the memory, the program when executed by the processor implementing the steps of the network service satisfaction analysis method of any of claims 1 to 7.
10. A storage medium, which is a computer-readable storage medium, for computer-readable storage, characterized in that the storage medium stores one or more programs executable by one or more processors to implement the steps of the network service satisfaction analysis method of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410239037.3A CN118018429A (en) | 2024-03-01 | 2024-03-01 | Network service satisfaction analysis method, system, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410239037.3A CN118018429A (en) | 2024-03-01 | 2024-03-01 | Network service satisfaction analysis method, system, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118018429A true CN118018429A (en) | 2024-05-10 |
Family
ID=90945032
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410239037.3A Pending CN118018429A (en) | 2024-03-01 | 2024-03-01 | Network service satisfaction analysis method, system, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118018429A (en) |
-
2024
- 2024-03-01 CN CN202410239037.3A patent/CN118018429A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US12014257B2 (en) | Domain specific language for generation of recurrent neural network architectures | |
| US20190340533A1 (en) | Systems and methods for preparing data for use by machine learning algorithms | |
| Farias et al. | Exploring multimodal data fusion through joint decompositions with flexible couplings | |
| CN114330713B (en) | Convolutional neural network model pruning method and device, electronic equipment and storage medium | |
| Esnaashari et al. | Irregular cellular learning automata | |
| CN109120462A (en) | Prediction technique, device and the readable storage medium storing program for executing of opportunistic network link | |
| CN112085615B (en) | Training method and device for graphic neural network | |
| CN112598111B (en) | Abnormal data identification method and device | |
| CN108171010B (en) | Protein complex detection method and device based on semi-supervised network embedded model | |
| Mall et al. | Representative subsets for big data learning using k-NN graphs | |
| CN114072809A (en) | Small and fast video processing network via neural architectural search | |
| CN112418320A (en) | Enterprise association relation identification method and device and storage medium | |
| CN113592593A (en) | Training and application method, device, equipment and storage medium of sequence recommendation model | |
| CN117061322A (en) | Internet of things flow pool management method and system | |
| CN115221396A (en) | Information recommendation method and device based on artificial intelligence and electronic equipment | |
| CN114219562A (en) | Model training method, enterprise credit evaluation method and device, equipment and medium | |
| CN115496144A (en) | Power distribution network operation scene determining method and device, computer equipment and storage medium | |
| CN115374950A (en) | Sample detection method, sample detection device, electronic apparatus, and storage medium | |
| CN114492669A (en) | Keyword recommendation model training method, recommendation method and device, equipment and medium | |
| CN110347916B (en) | Cross-scene item recommendation method and device, electronic equipment and storage medium | |
| CN109952742B (en) | Graph structure processing method, system, network device and storage medium | |
| CN118018429A (en) | Network service satisfaction analysis method, system, equipment and storage medium | |
| US11609936B2 (en) | Graph data processing method, device, and computer program product | |
| Wang et al. | Randomized block coordinate descendant STRONG for large-scale stochastic optimization | |
| CN114973377A (en) | Face beauty prediction method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |