CN115225457B - Knowledge and data dual-drive-based network fault analysis method - Google Patents
Knowledge and data dual-drive-based network fault analysis method Download PDFInfo
- Publication number
- CN115225457B CN115225457B CN202210729564.3A CN202210729564A CN115225457B CN 115225457 B CN115225457 B CN 115225457B CN 202210729564 A CN202210729564 A CN 202210729564A CN 115225457 B CN115225457 B CN 115225457B
- Authority
- CN
- China
- Prior art keywords
- network fault
- network
- fault
- knowledge
- diagnosis
- 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.)
- Active
Links
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
- H04L41/06—Management of faults, events, alarms or notifications
- H04L41/0631—Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
- H04L41/06—Management of faults, events, alarms or notifications
- H04L41/0631—Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
- H04L41/065—Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis involving logical or physical relationship, e.g. grouping and hierarchies
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a network fault analysis method based on knowledge and data dual drive, which constructs a network fault knowledge map based on the step-by-step diagnosis of network faults by using machine learning, and performs network fault analysis by using knowledge retrieval and knowledge reasoning, so that the final network fault analysis result has high diagnosis accuracy and good interpretation.
Description
Technical Field
The invention relates to the technical field of communication, in particular to a network fault analysis method based on knowledge and data dual driving.
Background
Along with the development of technology and the progress of age, the development of mobile communication networks has undergone a change and innovation once and becomes an indispensable part of daily life. Along with the increase of the number of network users and the diversification of demands, the current network scene is very complex, and the management of the complex network environment is very challenging to ensure the normal operation of the network.
In recent years, with the development of big data mining and artificial intelligence technology, more and more intelligent fault diagnosis methods based on machine learning are favored. Currently, network fault diagnosis techniques based on machine learning are widely applied, and mainly comprise a fault diagnosis method based on a Support Vector Machine (SVM), a fault diagnosis method based on an Artificial Neural Network (ANN), a network fault diagnosis method based on deep learning, and the like. The method for diagnosing faults based on machine learning can make full use of big data to diagnose possible fault modes, fault reasons and the like, but the network fault diagnosis based on data only has two problems, namely, the method is firstly interpretable, the traditional model for machine learning and deep learning is a black box for a user, and the interpretation cannot be given for the output fault diagnosis result, so that the application of the model in actual engineering is reduced; secondly, the prior knowledge, such as unstructured knowledge, such as a fault isolation manual, cannot be effectively utilized, so that unstructured data resources are wasted.
The knowledge graph concept has been widely focused in the industry since 2012 by google, and has been widely applied in many fields in recent years. The knowledge graph can gather massive information, data and link relations into knowledge, so that information resources are easier to calculate, understand and evaluate, a set of semantic knowledge base is formed, a more effective mode is provided for massive, heterogeneous and dynamic big data expression, organization, management and utilization, the intelligent level of a network is higher, and the network is more similar to the cognitive thinking of human beings. In recent years, knowledge maps have been widely used in the fields of medical treatment, education, fault diagnosis, and the like.
In order to make up the defects of the fault diagnosis method based on machine learning only, the prior knowledge is fully utilized, and the network fault analysis method based on the combination of the knowledge graph and the machine learning is provided.
Disclosure of Invention
The purpose of the invention is that: aiming at the problem that the network fault diagnosis process is only based on machine learning, the network fault analysis method based on the knowledge and data dual-drive is constructed by combining a knowledge graph, the fault diagnosis is carried out by using the machine learning, the diagnosis result is used as a known condition to carry out knowledge reasoning and knowledge retrieval in the knowledge graph, the information related to the fault is output, and the interpretability and the practicability of the fault analysis are improved;
in order to achieve the above functions, the invention designs a network fault analysis method based on knowledge and data dual drive, and the following steps S1-S3 are executed aiming at a network with network faults in a target area to finish network fault diagnosis:
s1, respectively taking a network fault name, a network fault reason, a network fault description, a network fault representation, a problem caused by a network fault, a network fault occurrence position, network fault solving measures and a network fault type as an upper-layer body of a network fault knowledge graph, respectively carrying out knowledge extraction on structured data, semi-structured data and unstructured data of the upper-layer body of each network fault knowledge graph to obtain a triplet corresponding to the upper-layer body, another upper-layer body related to the upper-layer body and a relation between the two upper-layer bodies, and constructing and storing the knowledge graph based on the triples corresponding to each upper-layer body;
s2, aiming at a network with network faults in a target area, collecting data samples with known network fault names and network fault reasons, firstly taking the data samples as input, taking the network fault names as output, training a machine learning network through a machine learning method to obtain a network fault name diagnosis model, then taking the data samples as input, taking the network fault reasons as output, training the machine learning network through the machine learning method to obtain a network fault reason diagnosis model, applying the obtained network fault name diagnosis model and the network fault reason diagnosis model, inputting actual data samples aiming at the target area to obtain the network fault names and the network fault reasons of the network faults in the target area, and constructing a network fault knowledge graph of the target area network based on the obtained network fault names and the network fault reasons of the target area and the knowledge graph obtained in the step S1;
and S3, based on the network fault knowledge graph of the target area network obtained in the step S2, aiming at network faults occurring in the target area, respectively diagnosing the fault name output by the model and the fault source output by the model as a core entity according to the fault name, searching and outputting an upper-layer body with the distance of 1 from the core entity in the knowledge graph based on a knowledge search method of sub-graph matching, and outputting the relation between the core entity and the upper-layer body to finish the diagnosis of the network faults occurring in the target area.
As a preferred technical scheme of the invention: in the step S1, aiming at the constructed knowledge graph, the knowledge graph is stored in a Neo4j graph database mode.
As a preferred technical scheme of the invention: in step S2, the network fault names include weak coverage and poor quality, the two network fault names respectively correspond to network fault reasons, the network fault reasons corresponding to the weak coverage include large station spacing, room leakage, abnormal switching threshold, missing neighbor cells and abnormal measurement threshold, and the network fault reasons corresponding to the poor quality include MOD3 interference, pilot pollution and overlapping coverage;
the specific steps of step S2 are as follows:
s21, aiming at a network with network faults in a target area, collecting a data sample comprising at least 8 characteristic values, carrying out standardized pretreatment on the data sample, and carrying out relevance descending order on each characteristic value of the data sample based on an XGboost algorithm characteristic value ordering function;
s22, based on the arrangement of the characteristic values of the data samples obtained in the step S21, selecting the first 8 characteristic values in the arrangement, taking the selected characteristic values as input, respectively based on 6 machine learning algorithms of logistic regression, linear discriminant analysis, K nearest neighbor, decision tree, naive Bayes and support vector machine, and training a machine learning model by using the network fault name and the network fault cause as output to obtain a network fault name diagnosis model and a network fault cause diagnosis model;
s23, comparing the diagnosis accuracy of the 6 machine learning algorithms in the step S22, respectively aiming at the network fault name diagnosis model and the network fault cause diagnosis model, matching the machine learning algorithm with the highest diagnosis accuracy, and outputting the diagnosis result of the machine learning algorithm with the highest diagnosis accuracy.
The beneficial effects are that: the advantages of the present invention over the prior art include:
1. the combination of machine learning and knowledge graph is applied to network fault diagnosis, so that the accuracy of fault diagnosis is ensured, and the interpretability of fault diagnosis is effectively improved.
2. The network fault diagnosis problem is divided into two sub-problems, and different machine learning algorithms are used for processing different problems, so that the accuracy of the network fault diagnosis is greatly improved.
3. The fault diagnosis result is displayed in the form of a knowledge graph sub-graph, so that the fault diagnosis result is clearer and the response is faster.
Drawings
FIG. 1 is a flow chart of a knowledge and data dual driven network failure analysis method provided in accordance with an embodiment of the present invention;
FIG. 2 is a diagram of a knowledge graph provided in accordance with an embodiment of the present invention;
fig. 3 is a network failure knowledge graph provided in accordance with an embodiment of the present invention;
FIG. 4 is a comparison of machine learning accuracy provided in accordance with an embodiment of the present invention;
fig. 5 is a knowledge subgraph of a weak coverage output as a core entity, provided according to an embodiment of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for more clearly illustrating the technical aspects of the present invention, and are not intended to limit the scope of the present invention.
Referring to fig. 1, in the network fault analysis method based on knowledge and data dual driving provided by the embodiment of the invention, for a network where a network fault occurs in a target area, the following steps S1 to S3 are executed to complete network fault diagnosis:
s1, respectively taking a network fault name, a network fault reason, a network fault description, a network fault representation, a problem caused by a network fault, a network fault occurrence position, network fault solving measures and a network fault type as an upper-layer body of a network fault knowledge graph, respectively carrying out knowledge extraction on structured data, semi-structured data and unstructured data of the upper-layer body of each network fault knowledge graph to obtain a triplet corresponding to the upper-layer body, another upper-layer body related to the upper-layer body and a relation between the two upper-layer bodies, and constructing and storing the knowledge graph based on the triples corresponding to each upper-layer body; knowledge graph structure referring to fig. 2.
And storing the knowledge graph in a Neo4j graph database form aiming at the constructed knowledge graph.
S2, aiming at a network with network faults in a target area, collecting data samples with known network fault names and network fault reasons, firstly taking the data samples as input, taking the network fault names as output, training a machine learning network through a machine learning method to obtain a network fault name diagnosis model, then taking the data samples as input, taking the network fault reasons as output, training the machine learning network through the machine learning method to obtain a network fault reason diagnosis model, applying the obtained network fault name diagnosis model and the network fault reason diagnosis model, inputting actual data samples aiming at the target area to obtain the network fault names and the network fault reasons of the network faults in the target area, and constructing a network fault knowledge graph of the target area network based on the obtained network fault names and the network fault reasons of the target area and the knowledge graph obtained in the step S1; the network failure knowledge graph is referred to in fig. 3.
In step S2, the network fault names include weak coverage and poor quality, the two network fault names respectively correspond to network fault reasons, the network fault reasons corresponding to the weak coverage include large station spacing, room leakage, abnormal switching threshold, missing neighbor cells and abnormal measurement threshold, and the network fault reasons corresponding to the poor quality include MOD3 interference, pilot pollution and overlapping coverage;
the specific steps of step S2 are as follows:
s21, aiming at a network with network faults in a target area, collecting data samples comprising at least 8 eigenvalues, wherein the fault data samples comprise eigenvalues related to network fault states, and the eigenvalues comprise at least RSRP, RSRQ, RSSI and SINR; carrying out standardized pretreatment on the data sample, and carrying out relevance descending order arrangement on each characteristic value of the data sample based on the characteristic value ordering function of the XGboost algorithm;
the specific method for standardized pretreatment is as follows:
for k eigenvalue sets X 1 ,X 2 ,…,X i ,…,X k ++, wherein X i ＝*x i1 ,x i2 ,…,x in +，x i1 ,x i2 ,…,x in Respectively the ith characteristic value in n fault data samples;
the following formula is normalized for each eigenvalue:
wherein x is ij For the characteristic value set X i The j-th characteristic value, Y ij Is x ij Normalized values after pretreatment.
S23, selecting the first 8 characteristic values in the arrangement based on the characteristic value arrangement of the fault data sample obtained in the step S22, taking the selected characteristic values as input, respectively training a machine learning model based on 6 machine learning algorithms of logistic regression, linear discriminant analysis, K nearest neighbor, decision tree, naive Bayes and support vector machine by using a network fault mode and a network fault source as output;
s24, comparing the diagnosis accuracy of the 6 machine learning algorithms in the step S23, and respectively aiming at a network failure mode and a network failure cause, matching the machine learning algorithm with the highest diagnosis accuracy, and outputting the diagnosis result of the machine learning algorithm with the highest diagnosis accuracy; 6 machine learning accuracy contrast graphs refer to fig. 4.
The calculation formula of the diagnosis accuracy rate is as follows:
where Ac is the diagnostic accuracy, TP is the number of times positive samples are predicted as positive samples, TN is the number of times negative samples are predicted as negative samples, FP is the number of times negative samples are predicted as positive samples, i.e. the number of false positives, TN is the number of times positive samples are predicted as negative samples, i.e. the number of false negatives.
And S3, based on the network fault knowledge graph of the target area network obtained in the step S2, aiming at network faults occurring in the target area, respectively diagnosing the fault name output by the model and the fault source output by the model as a core entity according to the fault name, searching and outputting an upper-layer body with the distance of 1 from the core entity in the knowledge graph based on a knowledge search method of sub-graph matching, and outputting the relation between the core entity and the upper-layer body to finish the diagnosis of the network faults occurring in the target area.
Wherein, the knowledge subgraph output by taking weak coverage as a core entity refers to fig. 5.
The embodiments of the present invention have been described in detail with reference to the drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the spirit of the present invention.
Claims (3)
1. A network fault analysis method based on knowledge and data dual drive is characterized in that the following steps S1-S3 are executed aiming at a network with network faults in a target area to finish network fault diagnosis:
s1, respectively taking a network fault name, a network fault reason, a network fault description, a network fault representation, a problem caused by a network fault, a network fault occurrence position, network fault solving measures and a network fault type as an upper-layer body of a network fault knowledge graph, respectively carrying out knowledge extraction on structured data, semi-structured data and unstructured data of the upper-layer body of each network fault knowledge graph to obtain a triplet corresponding to the upper-layer body, another upper-layer body related to the upper-layer body and a relation between the two upper-layer bodies, and constructing and storing the knowledge graph based on the triples corresponding to each upper-layer body;
s2, aiming at a network with network faults in a target area, collecting data samples with known network fault names and network fault reasons, firstly taking the data samples as input, taking the network fault names as output, training a machine learning network through a machine learning method to obtain a network fault name diagnosis model, then taking the data samples as input, taking the network fault reasons as output, training the machine learning network through the machine learning method to obtain a network fault reason diagnosis model, applying the obtained network fault name diagnosis model and the network fault reason diagnosis model, inputting actual data samples aiming at the target area to obtain the network fault names and the network fault reasons of the network faults in the target area, and constructing a network fault knowledge graph of the target area network based on the obtained network fault names and the network fault reasons of the target area and the knowledge graph obtained in the step S1;
and S3, based on the network fault knowledge graph of the target area network obtained in the step S2, aiming at network faults occurring in the target area, respectively diagnosing the fault name output by the model and the fault source output by the model as a core entity according to the fault name, searching and outputting an upper-layer body with the distance of 1 from the core entity in the knowledge graph based on a knowledge search method of sub-graph matching, and outputting the relation between the core entity and the upper-layer body to finish the diagnosis of the network faults occurring in the target area.
2. The network fault analysis method based on knowledge and data dual driving as claimed in claim 1, wherein in step S1, the knowledge graph is stored in the form of Neo4j graph database for the constructed knowledge graph.
3. The method for analyzing network faults based on knowledge and data dual driving as claimed in claim 1, wherein in step S2, the network fault names include weak coverage and poor quality, the two network fault names correspond to network fault reasons respectively, the network fault reasons corresponding to the weak coverage include large station spacing, room separation leakage, abnormal switching threshold, missing neighbor cells and abnormal measurement threshold, and the network fault reasons corresponding to the poor quality include MOD3 interference, pilot pollution and overlapping coverage;
the specific steps of step S2 are as follows:
s21, aiming at a network with network faults in a target area, collecting a data sample comprising at least 8 characteristic values, carrying out standardized pretreatment on the data sample, and carrying out relevance descending order on each characteristic value of the data sample based on an XGboost algorithm characteristic value ordering function;
s22, based on the arrangement of the characteristic values of the data samples obtained in the step S21, selecting the first 8 characteristic values in the arrangement, taking the selected characteristic values as input, respectively based on 6 machine learning algorithms of logistic regression, linear discriminant analysis, K nearest neighbor, decision tree, naive Bayes and support vector machine, and training a machine learning model by using the network fault name and the network fault cause as output to obtain a network fault name diagnosis model and a network fault cause diagnosis model;
s23, comparing the diagnosis accuracy of the 6 machine learning algorithms in the step S22, respectively aiming at the network fault name diagnosis model and the network fault cause diagnosis model, matching the machine learning algorithm with the highest diagnosis accuracy, and outputting the diagnosis result of the machine learning algorithm with the highest diagnosis accuracy.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210729564.3A CN115225457B (en) | 2022-06-24 | 2022-06-24 | Knowledge and data dual-drive-based network fault analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210729564.3A CN115225457B (en) | 2022-06-24 | 2022-06-24 | Knowledge and data dual-drive-based network fault analysis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115225457A CN115225457A (en) | 2022-10-21 |
| CN115225457B true CN115225457B (en) | 2023-09-26 |
Family
ID=83609821
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210729564.3A Active CN115225457B (en) | 2022-06-24 | 2022-06-24 | Knowledge and data dual-drive-based network fault analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115225457B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110008288A (en) * | 2019-02-19 | 2019-07-12 | 武汉烽火技术服务有限公司 | The construction method in the knowledge mapping library for Analysis of Network Malfunction and its application |
| CN111311059A (en) * | 2020-01-16 | 2020-06-19 | 成都大汇物联科技有限公司 | Knowledge graph-based water mill room fault diagnosis method |
| KR20210068687A (en) * | 2019-12-02 | 2021-06-10 | 대우조선해양 주식회사 | Method for diagnosing failure and serching based on knowledge in complex facility |
| CN113312411A (en) * | 2021-06-22 | 2021-08-27 | 中国电子科技集团公司第十四研究所 | Equipment fault cause diagnosis method based on knowledge graph and rule constraint |
-
2022
- 2022-06-24 CN CN202210729564.3A patent/CN115225457B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110008288A (en) * | 2019-02-19 | 2019-07-12 | 武汉烽火技术服务有限公司 | The construction method in the knowledge mapping library for Analysis of Network Malfunction and its application |
| KR20210068687A (en) * | 2019-12-02 | 2021-06-10 | 대우조선해양 주식회사 | Method for diagnosing failure and serching based on knowledge in complex facility |
| CN111311059A (en) * | 2020-01-16 | 2020-06-19 | 成都大汇物联科技有限公司 | Knowledge graph-based water mill room fault diagnosis method |
| CN113312411A (en) * | 2021-06-22 | 2021-08-27 | 中国电子科技集团公司第十四研究所 | Equipment fault cause diagnosis method based on knowledge graph and rule constraint |
Non-Patent Citations (1)
| Title |
|---|
| 基于神经网络的局域网故障诊断系统设计;刘凌雁;杨新华;李伟;;微计算机信息(第28期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115225457A (en) | 2022-10-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yang et al. | Learn to explain efficiently via neural logic inductive learning | |
| CN108363478B (en) | For wearable device deep learning application model load sharing system and method | |
| CN109543043A (en) | A kind of automatic monitoring method of the power communication big data mistake of knowledge based map reasoning | |
| He et al. | Fault diagnosis and location based on graph neural network in telecom networks | |
| Maatouk et al. | Large language models for telecom: Forthcoming impact on the industry | |
| Zhang et al. | Cause-aware failure detection using an interpretable XGBoost for optical networks | |
| CN111586728B (en) | Small sample characteristic-oriented heterogeneous wireless network fault detection and diagnosis method | |
| CN103793589B (en) | High-speed train fault handling method | |
| Yan et al. | Graph-based knowledge acquisition with convolutional networks for distribution network patrol robots | |
| Li et al. | A hybrid coevolutionary algorithm for designing fuzzy classifiers | |
| CN115225457B (en) | Knowledge and data dual-drive-based network fault analysis method | |
| CN114116957A (en) | Electromagnetic information leakage intelligent analysis method based on knowledge graph | |
| Cross et al. | Patterns of fuzzy rule-based inference | |
| Kaur et al. | Machine learning empowered green task offloading for mobile edge computing in 5G networks | |
| Chandiok et al. | Cognitive decision support system for medical diagnosis | |
| He et al. | Representation learning of knowledge graph for wireless communication networks | |
| CN115130620A (en) | Power consumption mode identification model generation method and device for power equipment | |
| Eremeev et al. | Methods and tools for reasoning by analogy in intelligent decision support systems | |
| Chernov et al. | Intelligent decision support by means of dynamic description logic | |
| Sandhu et al. | A comparative analysis of fuzzy, neuro-fuzzy and fuzzy-GA based approaches for software reusability evaluation | |
| Wang et al. | A Knowledge Graph Reasoning Approach Integrating Attention-based LSTM and Multi-Agent Reinforcement Learning | |
| CN113530921A (en) | Hydraulic machine fault diagnosis method based on ES-MLSTM | |
| Sohail et al. | Interpretable and Adaptable Early Warning Learning Analytics Model. | |
| Ahmed et al. | A Comprehensive Review on Machine Learning Models for Medical Data Classification | |
| Vagin et al. | Modelling human reasoning in intelligent decision support systems |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |