CN109343995A - Intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot - Google Patents
Intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot Download PDFInfo
- Publication number
- CN109343995A CN109343995A CN201811252658.6A CN201811252658A CN109343995A CN 109343995 A CN109343995 A CN 109343995A CN 201811252658 A CN201811252658 A CN 201811252658A CN 109343995 A CN109343995 A CN 109343995A
- Authority
- CN
- China
- Prior art keywords
- analysis
- machine learning
- data
- source heterogeneous
- customer service
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0751—Error or fault detection not based on redundancy
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0766—Error or fault reporting or storing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0766—Error or fault reporting or storing
- G06F11/0781—Error filtering or prioritizing based on a policy defined by the user or on a policy defined by a hardware/software module, e.g. according to a severity level
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0766—Error or fault reporting or storing
- G06F11/0787—Storage of error reports, e.g. persistent data storage, storage using memory protection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/079—Root cause analysis, i.e. error or fault diagnosis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0793—Remedial or corrective actions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3447—Performance evaluation by modeling
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Hardware Design (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
In a kind of intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot disclosed by the invention, analysis and processing module pre-processes real time data with multi-source heterogeneous Data fusion technique, is normalized, after correlation analysis, it is exported by Model Matching as a result, determining the basic reason of problem by root cause analysis；Model management module carries out machine learning for obtaining data, exports various inference patterns to analysis and processing module and recommends decision-making module；Recommend decision-making module, by context identification, and according to the corresponding subsequent processing behavior of strategy matching.The present invention is based on multi-source heterogeneous Data fusion techniques, the O&M mode of machine learning, the compatibility issue for having taken into account big data improves the degree of automation of O&M, guarantees network security and service quality, it is provided simultaneously with certain predictive ability, reduces related O&M cost；And the result output mode of customer service robot, further realize the intelligence of O&M.
Description
Technical field
The present invention relates to a kind of intelligent O&M analysis systems, and in particular to one kind is based on multi-source heterogeneous data fusion, machine
The intelligent O&M analysis system of study and customer service robot.
Background technique
During traditional IT system O&M, fault pre-alarming, malfunction elimination etc. are extremely important but time-consuming and laborious work, often
O&M mode can be usually 4 processes from higher level of abstraction: by O&M object modeling, performance alarm monitoring, for receiving
Data analyzed after decision carried out by expertise again, issue control finally by configuration script to by O&M object
On.During this, most automation measures and related tool are implemented by the way of rule, this rule similar first
For the rule in artificial intelligence, the automation O&M of management system is reached by it.Such as filtering rule, the performance pipe of alarm
The ANR (automatic cell neighboring relation configuration rule) in QoS rule, self-organizing network in reason.Here the rule being arranged all is
The accumulation of expertise, is set in software systems, for reaching the target of automation O&M.Rule herein is exactly not
O&M with scene is defined the logical relation between the various events for meeting this scene demand, behavior by maintenance personnel in the process,
Then software is run according to the logical relation of this scene definition, improves the degree of automation of O&M.
But there are several defects for the mode of rule setting: our this logical process cannot be too complicated first, otherwise only
It can be realized with code, and for flexibility, and hope directly defines at the scene, and this creates the terminal contradictions；Secondly this rule
Relatively good definition when a small amount of node, but if there is great deal of nodes, need to define tens of thousands of, hundreds of thousands rule, just
It is extremely difficult；Finally for large-scale open network, often there is centainly ecological, i.e., operation maintenance personnel is not known when
A new role can be generated in this network, what relationship this role can generate with other roles, so until generating
Problem, then come when rule monitoring is arranged, often loss is had occurred and that, the problem of this kind of typical case is exactly financial air control problem.
Intelligent O&M is the O&M mode based on machine learning and deep learning, passes through the study to operation/maintenance data, energy
Various rules are voluntarily generated for monitoring Network Abnormal automatically, network exception event reason is analyzed, selects the network optimization and healing
Behavior；And with the variation of network running quality and network topology, original various O&M rules can be carried out self evolution to fit
Close the situation of current network.Intelligent O&M be cloud computing, big data, artificial intelligence technology O&M field comprehensive application,
I.e. cloud computing provides acquisition, storage and computing capability for various log big datas, and artificial intelligence technology is provided in conjunction with industry
Knowledge converts O&M problem to the ability of big data analysis modeling.
Therefore, under big data scene, O&M to intelligent development be an important trend: based on to business operational system
Understanding, the algorithm for carrying out machine learning to a large amount of daily record datas of accumulation models, realization finds the problem automatically, problem analysis,
The multiple functions such as anticipation problem in advance play auxiliary operation maintenance personnel, the effect for finally reducing system cost, promoting O&M efficiency.
Meanwhile operation/maintenance data is isomery, multi-source, multimode, daily record data, user data, network data, text data,
A plurality of types of data such as image/video data and position data and distinct device do not have to business, different levels, different user
Data how to merge use, playing bigger effect is the significant challenge faced.
Continually developing and utilizing with communication software, client's consultation way that O&M client service center is faced is also continuous
Increase, such as QQ, wechat and short message consultation way are all gradually applied to by client in the consulting of O&M business, same with this
When, the problem of client seeks advice from, also gradually becomes specialized and objectifies, if a variety of high of O&M industry cannot effectively be established
Service mode is imitated, the problem of the problem of client is seeked advice from cannot get timely, accurate and efficient answer will be faced.
And the intelligent customer service machine of the modes such as text or voice interaction as core work and is passed through using domain knowledge base construction
People's system effective can then be integrated with Customer Service Center by all kinds of means, can be had while substantially reducing customer service cost
Effect reduces cost of labor, enhancing user experience, to promote the quality of service and the brand image of enterprise innovation.
Summary of the invention
The purpose of the present invention is overcoming the above-mentioned prior art, provide one kind can deep learning, there is pre- measurement of power
It can, be automatically repaired the intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot of function.
The present invention is achieved through the following technical solutions:
Intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot, including analysis
Processing module, recommends decision-making module, verification Application module at model management module, it is characterised in that:
The analysis and processing module is carried out in advance with real time data of the multi-source heterogeneous Data fusion technique to production environment
After processing, normalization, correlation analysis, exported by Model Matching as a result, the result includes the various of internal system
Abnormal, network key event various prediction results, the context of some special event is exported by root cause analysis, passes through institute
State the basic reason that context determines problem；The pretreatment, which refers to, to be parsed various isomery daily record datas, is converted, clearly
It washes, specification operation, the necessary processing and the quality of data before completing data use guarantee；The analysis processing, mainly includes streaming
Calculation processing frame Spark, offline batch processing MR frame, artificial intelligence Computational frame, data storage and search engine；It is described defeated
Result includes exporting result by customer service robot out.
The model management module, for obtaining historical data, progress machine learning exports various inference patterns, wherein
Output context identification, prediction, abnormality detection model are analyzed and processed to the analysis and processing module, are exported at context
Reason, output policy, rule model are used for automatic closed loop O&M to recommendation decision-making module；The machine learning includes offline machine
Device learning training platform, algorithm frame and model；
The recommendation decision-making module, by determining environment locating for current event based on the context identification of analysis processing
Later, according to the corresponding subsequent processing behavior of strategy matching, and behavior list is then that the maintenance behavior learning based on user is got,
In this way, reach automatic O&M；
The verification Application module carries out business monitorings to the results of automatic O&M by various rules, prevent it is entire from
Dynamic O&M sideslip.
Further, the multi-source heterogeneous Data fusion technique is in the storage and management mode for not changing initial data
Under, logic integration is carried out to multi-source heterogeneous data.
Further, the multi-source heterogeneous Data fusion technique is that the O&M achievement data of enormous amount is abstracted as timing
Data, and store to time series databases, realize the quick storage and inquiry of operation/maintenance data.
Further, the multi-source heterogeneous Data fusion technique is that the O&M achievement data of enormous amount is abstracted into the time
With two, space dimension, data matrix is formed, in favor of subsequent management and application.
Further, the machine learning is machine clustering learning, through machine clustering learning in some chance events
Sorted out same category of event is belonged to；Then by the correlation analysis between anomalous event, to find these events
Correlation；By the analysis of exception service and event contribution degree, the specific network event for leading to exception service is found；By complete
Link, which calls to excavate, finds that the relationship between different components, topology objects finds fault propagation chain in conjunction with the analysis of front；It will
Become a kind of means of event anomalies detection for the prediction of event, if it find that the value phase of detection is farther out from predicted value, then recognizes
To be abnormal.
Further, the algorithm of the machine learning is log intelligence aggregating algorithm, by clustering to operation/maintenance data
Analysis calculates the distance between log and the matching of regular pattern, data is carried out lossless compression, extract daily record data
Mode quickly removes log noise, and operation maintenance personnel is made quickly to navigate to exception information, carries out accident root-cause analysis；It extracts
Mode can be added in monitoring daily, hourly by recording its characteristic value, record each mode using log comparison algorithm
Exception information actual change, be applied to O&M change after, monitor the increase and decrease of each mode, in time find system exception feelings
Condition.
Further, the model management module is by being respectively set model parameter, and comprehensive fortune to every kind of capacity performance index
With ARMA and GARCH prediction model, short-term and long-term forecast is carried out to capacity performance index；Pass through prediction error analysis and parameters revision
Process, while the historical data of learning capacity index and new generation data, adjust the parameter combination of prediction model in time, adapt to refer to
The development and change of data are marked, model prediction accuracy is continuously improved.
Further, Dynamic Baseline Early-warning Model is introduced in the model management module, operation maintenance personnel need to only be unified to be arranged
Different threshold value of warning, so that it may take consecutive mean, dynamic on year-on-year basis automatically and dynamic by the Dynamic Baseline Early-warning Model
Ring reaches the uniform threshold early warning under different time sections, different flow scene, different system loads, realizes intelligence than algorithms of different
The early warning effect that can be automated.
Further, the customer service robot include by machine learning constantly build and improve robot knowledge base, into
The analysis of row back-end data and feedback.
Further, the customer service robot has the speech processing system based on machine learning.
The present invention is the O&M mode based on machine learning, by the study to operation/maintenance data, can voluntarily generate various rule
Then for monitoring Network Abnormal automatically, network exception event reason is analyzed, the network optimization and healing behavior are selected；And with network
The variation of running quality and network topology, original various O&M rules can be carried out self evolution to be suitble to the shape of current network
Condition；The degree of automation of network O&M is promoted, guarantees network security and service quality, is provided simultaneously with certain predictive ability, is dropped
Low correlation O&M cost.
The present invention is based on big datas, by using multi-source heterogeneous Data fusion technique, according to the cardinal trait of operation/maintenance data
And service requirement, it realizes the storage, calculating and retrieval to magnanimity isomery operation/maintenance data, had both taken into account polymorphic type operation/maintenance data in O&M
Requirement in terms of format compatibilities, and can guarantee the timeliness of mass data inquiry and processing.
The present invention further realizes the intelligence of O&M by the result output mode in conjunction with customer service robot, improves intelligence
The efficiency of energy O&M.
Detailed description of the invention
Fig. 1 is intelligent O&M analysis system frame diagram of the invention.
Fig. 2 is intelligent O&M scenarios figure of the invention.
Specific embodiment
Such as Fig. 1, the intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot, packet
It includes analysis and processing module, model management module, recommend decision-making module, verification Application module, it is characterised in that:
The analysis and processing module is carried out in advance with real time data of the multi-source heterogeneous Data fusion technique to production environment
After processing, normalization, correlation analysis, exported by Model Matching as a result, the result includes the various of internal system
Abnormal, network key event various prediction results, the context of some special event is exported by root cause analysis, passes through institute
State the basic reason that context determines problem；The pretreatment, which refers to, to be parsed various isomery daily record datas, is converted, clearly
It washes, specification operation, the necessary processing and the quality of data before completing data use guarantee；The analysis processing, mainly includes streaming
Calculation processing frame Spark, offline batch processing MR frame, artificial intelligence Computational frame, data storage and search engine；It is described defeated
Result includes exporting result by customer service robot out.
The model management module, for obtaining historical data, progress machine learning exports various inference patterns, wherein
Output context identification, prediction, abnormality detection model are analyzed and processed to the analysis and processing module, are exported at context
Reason, output policy, rule model are used for automatic closed loop O&M to recommendation decision-making module；The machine learning includes offline machine
Device learning training platform, algorithm frame and model；
The recommendation decision-making module, by determining environment locating for current event based on the context identification of analysis processing
Later, according to the corresponding subsequent processing behavior of strategy matching, and behavior list is then that the maintenance behavior learning based on user is got,
In this way, reach automatic O&M；
The verification Application module carries out business monitorings to the results of automatic O&M by various rules, prevent it is entire from
Dynamic O&M sideslip.
The multi-source heterogeneous Data fusion technique be in the case where not changing the storage and management mode of initial data, it is different to multi-source
Structure data carry out logic integration.
The multi-source heterogeneous Data fusion technique is the O&M achievement data of enormous amount to be abstracted as time series data, and deposit
Storage realizes the quick storage and inquiry of operation/maintenance data to time series databases.
The multi-source heterogeneous Data fusion technique is that the O&M achievement data of enormous amount is abstracted into time and space two
A dimension forms data matrix, in favor of subsequent management and application.
The machine learning is machine clustering learning, same belonging in some chance events by machine clustering learning
The event of classification is sorted out；Then by the correlation analysis between anomalous event, to find the correlation of these events；It is logical
The specific network event for leading to exception service is found in the analysis for crossing exception service and event contribution degree；It is called and is dug by full link
Pick finds that the relationship between different components, topology objects finds fault propagation chain in conjunction with the analysis of front；It will be for event
Prediction becomes a kind of means of event anomalies detection, if it find that the value phase of detection is farther out from predicted value, then it is assumed that be abnormal.
The algorithm of the machine learning is log intelligence aggregating algorithm, by carrying out clustering to operation/maintenance data, is calculated
Data are carried out lossless compression, extract the mode of daily record data, quickly by the distance between log and the matching of regular pattern
Log noise is removed, operation maintenance personnel is made quickly to navigate to exception information, carries out accident root-cause analysis；The mode extracted can lead to
Its characteristic value of overwriting is added in monitoring daily, hourly, the exception information of each mode is recorded using log comparison algorithm
Actual change, be applied to O&M change after, monitor the increase and decrease of each mode, in time find system exception situation.
The model management module by being respectively set model parameter to every kind of capacity performance index, and integrated use ARMA and
GARCH prediction model carries out short-term and long-term forecast to capacity performance index；By prediction error analysis and parameters revision process, together
When learning capacity index historical data and it is new generate data, adjust the parameter combination of prediction model in time, adapt to achievement data
Development and change, be continuously improved model prediction accuracy.
Dynamic Baseline Early-warning Model is introduced in the model management module, different early warning only need to be uniformly arranged in operation maintenance personnel
Threshold value, so that it may consecutive mean, dynamic be taken on year-on-year basis by the Dynamic Baseline Early-warning Model automatically and dynamic ring is than different calculations
Method reaches the uniform threshold early warning under different time sections, different flow scene, different system loads, realizes intelligent automation
Early warning effect.
The customer service robot includes constantly building and improving robot knowledge base by machine learning, carry out back-end data
Analysis and feedback.
The customer service robot has the speech processing system based on machine learning.
Such as Fig. 2, intelligence O&M is divided into intelligent measurement and intelligent monitoring two from the angle of scene by the above O&M analysis system
A dimension, there are three relevant big scene, such dimension classification and networks to influence feature phase big, that range is wide for each dimension
It closes.Even due to the result of artificial intelligence be often some probabilistic as a result, high probability as a result, the still thing of low probability
Once feelings are still very huge for the influence of whole network generation, so the inevitable first development intelligence from the perspective of operation
Monitor this read-only scene.In intellectual monitoring dimension, the big change of the process of the relatively traditional O&M prediction field that has been more
Scape is predicted the various events of network；Modified version in followed by traditional O&M scenarios, intelligent event monitoring and root
Because of analysis.And in intelligent monitoring dimension, first scene is the improvement of original SON (self-organizing network), intelligent optimization, passes through net
Network event prediction and personalized network feature extraction keep the network optimization much sooner and more targeted；Intelligent arranging is then needle
Modification to customer service is equipped with suitable Internet resources according to study to meet the new quality of service requirement of user；Self-healing is then
It is gradually to be evolved to from reparation behavior recommendation by the accumulation for O&M knowledge base and be automatically repaired network software failure.
In whole system design, the corresponding above-mentioned intellectual monitoring closed loop of analysis processing recommends decision to correspond to above-mentioned intelligence
It can control closed loop, two closed loops can distinguish evolution；Secondly there are a deviation-rectifying systems, this is by a large amount of business rule group
At, caused by this is also due to the importance of network, it is also due to network and generates (equipment vendor) and manage (operator) separation
Characteristic caused by；There are super managers for this last framework, entire automatic intelligent O&M process can be cut off；In addition
It can be seen that the inference pattern of whole network machine learning is divided into demand, prediction, mode, recommends four kinds, user couple is respectively represented
In the requirement of result, predictive behavior, the context identification behavior of problem, self-healing behavior.
Above-listed detailed description is illustrating for possible embodiments of the present invention, and the embodiment is not to limit this hair
Bright the scope of the patents, all equivalence enforcements or change without departing from carried out by the present invention, is intended to be limited solely by the scope of the patents of this case.
Claims (10)
1. the intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot, which is characterized in that
Including analysis and processing module, model management module, recommend decision-making module, verification Application module: the analysis and processing module, uses
Multi-source heterogeneous Data fusion technique pre-processes the real time data of production environment, normalized, correlation analysis；Pass through mould again
Type matches to export as a result, the result includes various exceptions, the various prediction results of network key event of internal system；So
The context for exporting some special event by root cause analysis afterwards, the basic reason of problem is determined by the context；Institute
State pretreatment refer to various isomery daily record datas are parsed, are converted, are cleaned, specification operation, complete data use before must
It handles and the quality of data guarantees；The analysis processing mainly includes that streaming computing handles frame Spark, offline batch processing MR
Frame, artificial intelligence Computational frame, data storage and search engine；
The output result includes exporting result by customer service robot；
The model management module carries out machine learning, exports various inference patterns for obtaining historical data, wherein output
Context identification, prediction, abnormality detection model are analyzed and processed to the analysis and processing module, and output context handles, is defeated
Strategy, rule model are used for automatic closed loop O&M to recommendation decision-making module out；The machine learning includes offline engineering
Practise training platform, algorithm frame and model；
The recommendation decision-making module, by based on analysis processing context identification come determine environment locating for current event it
Afterwards, according to the corresponding subsequent processing behavior of strategy matching, and behavior list is then that the maintenance behavior learning based on user is got, and is led to
Such mode is crossed, automatic O&M is reached；
The verification Application module carries out business monitoring by result of the various rules to automatic O&M, prevents entire automatic fortune
Tie up sideslip.
2. the intelligent O&M according to claim 1 based on multi-source heterogeneous data fusion, machine learning and customer service robot
Analysis system, which is characterized in that the multi-source heterogeneous Data fusion technique is in the storage and management side for not changing initial data
Under formula, logic integration is carried out to multi-source heterogeneous data.
3. the intelligence fortune according to claim 1 or 2 based on multi-source heterogeneous data fusion, machine learning and customer service robot
Tie up analysis system, which is characterized in that the multi-source heterogeneous Data fusion technique is to be abstracted the O&M achievement data of enormous amount
It for time series data, and stores to time series databases, realizes the quick storage and inquiry of operation/maintenance data.
4. the intelligence fortune according to claim 1 or 2 based on multi-source heterogeneous data fusion, machine learning and customer service robot
Tie up analysis system, which is characterized in that the multi-source heterogeneous Data fusion technique is to be abstracted the O&M achievement data of enormous amount
At time and two, space dimension, data matrix is formed, in favor of subsequent management and application.
5. the intelligent O&M according to claim 1 based on multi-source heterogeneous data fusion, machine learning and customer service robot
Analysis system, which is characterized in that the machine learning is machine clustering learning, by machine clustering learning in some chance events
Middle belongs to same category of event and is sorted out；Then by the correlation analysis between anomalous event, to find these things
The correlation of part；By the analysis of exception service and event contribution degree, the specific network event for leading to exception service is found；Pass through
Full link calls the relationship excavated and found between different components, topology objects, finds fault propagation chain；By the prediction for event
As a kind of means of event anomalies detection, if it find that the value phase of detection is farther out from predicted value, then it is assumed that be abnormal.
6. the intelligent O&M according to claim 1 based on multi-source heterogeneous data fusion, machine learning and customer service robot
Analysis system, which is characterized in that the algorithm of the machine learning is log intelligence aggregating algorithm, by gathering to operation/maintenance data
Alanysis calculates the distance between log and the matching of regular pattern, data is carried out lossless compression, extract daily record data
Mode, quickly remove log noise, operation maintenance personnel made quickly to navigate to exception information, carry out accident root-cause analysis；It extracts
Mode by recording its characteristic value, be added in monitoring daily, hourly, record each mode using log comparison algorithm
The actual change of exception information monitors the increase and decrease of each mode after being applied to O&M change, finds system exception situation in time.
7. the intelligent O&M according to claim 3 based on multi-source heterogeneous data fusion, machine learning and customer service robot
Analysis system, which is characterized in that the model management module is by being respectively set model parameter, and synthesis to every kind of capacity performance index
With ARMA and GARCH prediction model, short-term and long-term forecast is carried out to capacity performance index；It is repaired by prediction error analysis and parameter
Positive process, while the historical data of learning capacity index and new generation data, adjust the parameter combination of prediction model in time, adapt to
Model prediction accuracy is continuously improved in the development and change of achievement data.
8. the intelligent O&M according to claim 1 based on multi-source heterogeneous data fusion, machine learning and customer service robot
Analysis system, which is characterized in that Dynamic Baseline Early-warning Model is introduced in the model management module, operation maintenance personnel only need to uniformly be set
Set different threshold value of warning, so that it may take consecutive mean, dynamic year-on-year and dynamic automatically by the Dynamic Baseline Early-warning Model
State ring reaches the uniform threshold early warning under different time sections, different flow scene, different system loads, realizes than algorithms of different
The early warning effect of intelligent automation.
9. the intelligent O&M according to claim 1 based on multi-source heterogeneous data fusion, machine learning and customer service robot
Analysis system, which is characterized in that the customer service robot include by machine learning constantly build and improve robot knowledge base,
Carry out back-end data analysis and feedback.
10. the intelligent O&M according to claim 1 based on multi-source heterogeneous data fusion, machine learning and customer service robot
Analysis system, which is characterized in that the customer service robot has the speech processing system based on machine learning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811252658.6A CN109343995A (en) | 2018-10-25 | 2018-10-25 | Intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811252658.6A CN109343995A (en) | 2018-10-25 | 2018-10-25 | Intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109343995A true CN109343995A (en) | 2019-02-15 |
Family
ID=65312283
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811252658.6A Pending CN109343995A (en) | 2018-10-25 | 2018-10-25 | Intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109343995A (en) |
Cited By (35)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109787846A (en) * | 2019-03-27 | 2019-05-21 | 湖北大学 | A kind of 5G network service quality exception monitoring and prediction technique and system |
| CN110806961A (en) * | 2019-11-05 | 2020-02-18 | 上海七牛信息技术有限公司 | Intelligent early warning method and system and recommendation system |
| CN110807085A (en) * | 2019-09-12 | 2020-02-18 | 口碑(上海)信息技术有限公司 | Fault information query method and device, storage medium and electronic device |
| CN110891283A (en) * | 2019-11-22 | 2020-03-17 | 超讯通信股份有限公司 | Small base station monitoring device and method based on edge calculation model |
| CN110942155A (en) * | 2019-11-29 | 2020-03-31 | 广西电网有限责任公司 | Research method of machine learning engine |
| CN111143447A (en) * | 2019-11-28 | 2020-05-12 | 国网山东省电力公司经济技术研究院 | Power grid weak link dynamic monitoring and early warning decision system and method |
| CN111178537A (en) * | 2019-12-09 | 2020-05-19 | 华为技术有限公司 | Feature extraction model training method and device |
| CN111400284A (en) * | 2020-03-20 | 2020-07-10 | 广州咨元信息科技有限公司 | Method for establishing dynamic anomaly detection model based on performance data |
| CN111475682A (en) * | 2020-04-06 | 2020-07-31 | 武汉智领云科技有限公司 | Intelligent operation and maintenance platform based on super-large-scale data system |
| CN111639497A (en) * | 2020-05-27 | 2020-09-08 | 北京东方通科技股份有限公司 | Abnormal behavior discovery method based on big data machine learning |
| CN111666270A (en) * | 2020-06-03 | 2020-09-15 | 北京软通智慧城市科技有限公司 | Event analysis system and event analysis method |
| CN111767202A (en) * | 2020-07-08 | 2020-10-13 | 中国工商银行股份有限公司 | Abnormality detection method, abnormality detection device, electronic apparatus, and medium |
| CN111984499A (en) * | 2020-08-04 | 2020-11-24 | 中国建设银行股份有限公司 | Fault detection method and device for big data cluster |
| CN112116123A (en) * | 2019-08-05 | 2020-12-22 | 云智慧（北京）科技有限公司 | Intelligent alarm method and system based on dynamic baseline |
| CN112329928A (en) * | 2020-12-30 | 2021-02-05 | 四川新网银行股份有限公司 | Heterogeneous model-based user satisfaction analysis method |
| CN112363891A (en) * | 2020-11-18 | 2021-02-12 | 西安交通大学 | Exception reason obtaining method based on fine-grained event and KPIs analysis |
| CN112446031A (en) * | 2020-10-26 | 2021-03-05 | 国网安徽省电力有限公司信息通信分公司 | Operation and maintenance data display platform based on artificial intelligence |
| CN112711508A (en) * | 2020-12-21 | 2021-04-27 | 航天信息股份有限公司 | Intelligent operation and maintenance service system facing large-scale client system |
| CN112884452A (en) * | 2021-03-17 | 2021-06-01 | 北京幂数科技有限公司 | Intelligent operation and maintenance multi-source data acquisition visualization analysis system |
| CN112926749A (en) * | 2020-12-30 | 2021-06-08 | 国网宁夏电力有限公司信息通信公司 | Intelligent power grid information equipment monitoring system and method |
| CN112966860A (en) * | 2021-02-23 | 2021-06-15 | 武汉育知联信息科技有限公司 | Data analysis and prediction system based on attached pre-chain |
| CN113256180A (en) * | 2021-07-07 | 2021-08-13 | 杭州实在智能科技有限公司 | Customer service work order information intelligent dynamic loading method and system based on machine learning |
| CN113328878A (en) * | 2021-05-13 | 2021-08-31 | 上海励能信息技术有限公司 | Intelligent operation and maintenance monitoring system |
| CN113516447A (en) * | 2021-05-21 | 2021-10-19 | 陕西迅税通智能科技有限公司 | Electronic device and method for outputting fiscal reasoning matching result based on computer |
| CN113609201A (en) * | 2021-08-10 | 2021-11-05 | 珍岛信息技术（上海）股份有限公司 | Service data processing method and system |
| CN113934780A (en) * | 2021-12-15 | 2022-01-14 | 南京迪塔维数据技术有限公司 | Asset management system and method based on data middleboxes |
| CN114048882A (en) * | 2022-01-12 | 2022-02-15 | 北京鼎兴达信息科技股份有限公司 | Railway fault handling operation and maintenance decision suggestion method |
| CN114401398A (en) * | 2022-03-24 | 2022-04-26 | 北京华创方舟科技集团有限公司 | Intelligent video operation and maintenance management system |
| CN114638373A (en) * | 2020-12-15 | 2022-06-17 | Aptiv技术有限公司 | Managing machine learning environment |
| CN114662803A (en) * | 2022-05-23 | 2022-06-24 | 成都普惠道智慧能源科技有限公司 | Distributed energy storage safety monitoring method and Internet of things system |
| CN114676978A (en) * | 2022-03-03 | 2022-06-28 | 北京中科智上科技有限公司 | Production intelligent decision-making system and method for oil and gas field |
| CN114721336A (en) * | 2022-03-03 | 2022-07-08 | 上海核工程研究设计院有限公司 | Information security event early warning method for technological parameters of instrument control system |
| US11770307B2 (en) | 2021-10-29 | 2023-09-26 | T-Mobile Usa, Inc. | Recommendation engine with machine learning for guided service management, such as for use with events related to telecommunications subscribers |
| US11842301B1 (en) | 2022-05-23 | 2023-12-12 | Chengdu Puhuidao Smart Energy Technology Co., Ltd. | Methods for monitoring distributed energy storage safety and internet of things systems thereof |
| CN117376108A (en) * | 2023-12-07 | 2024-01-09 | 深圳市亲邻科技有限公司 | Intelligent operation and maintenance method and system for Internet of things equipment |
-
2018
- 2018-10-25 CN CN201811252658.6A patent/CN109343995A/en active Pending
Cited By (45)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109787846A (en) * | 2019-03-27 | 2019-05-21 | 湖北大学 | A kind of 5G network service quality exception monitoring and prediction technique and system |
| CN112116123A (en) * | 2019-08-05 | 2020-12-22 | 云智慧（北京）科技有限公司 | Intelligent alarm method and system based on dynamic baseline |
| CN110807085A (en) * | 2019-09-12 | 2020-02-18 | 口碑(上海)信息技术有限公司 | Fault information query method and device, storage medium and electronic device |
| CN110806961A (en) * | 2019-11-05 | 2020-02-18 | 上海七牛信息技术有限公司 | Intelligent early warning method and system and recommendation system |
| CN110891283A (en) * | 2019-11-22 | 2020-03-17 | 超讯通信股份有限公司 | Small base station monitoring device and method based on edge calculation model |
| CN111143447A (en) * | 2019-11-28 | 2020-05-12 | 国网山东省电力公司经济技术研究院 | Power grid weak link dynamic monitoring and early warning decision system and method |
| CN111143447B (en) * | 2019-11-28 | 2023-10-31 | 国网山东省电力公司经济技术研究院 | Dynamic monitoring early warning decision system and method for weak links of power grid |
| CN110942155A (en) * | 2019-11-29 | 2020-03-31 | 广西电网有限责任公司 | Research method of machine learning engine |
| CN111178537A (en) * | 2019-12-09 | 2020-05-19 | 华为技术有限公司 | Feature extraction model training method and device |
| CN111178537B (en) * | 2019-12-09 | 2023-11-17 | 华为云计算技术有限公司 | Feature extraction model training method and device |
| CN111400284B (en) * | 2020-03-20 | 2023-09-12 | 广州咨元信息科技有限公司 | Method for establishing dynamic anomaly detection model based on performance data |
| CN111400284A (en) * | 2020-03-20 | 2020-07-10 | 广州咨元信息科技有限公司 | Method for establishing dynamic anomaly detection model based on performance data |
| CN111475682A (en) * | 2020-04-06 | 2020-07-31 | 武汉智领云科技有限公司 | Intelligent operation and maintenance platform based on super-large-scale data system |
| CN111639497A (en) * | 2020-05-27 | 2020-09-08 | 北京东方通科技股份有限公司 | Abnormal behavior discovery method based on big data machine learning |
| CN111666270A (en) * | 2020-06-03 | 2020-09-15 | 北京软通智慧城市科技有限公司 | Event analysis system and event analysis method |
| CN111767202A (en) * | 2020-07-08 | 2020-10-13 | 中国工商银行股份有限公司 | Abnormality detection method, abnormality detection device, electronic apparatus, and medium |
| CN111984499A (en) * | 2020-08-04 | 2020-11-24 | 中国建设银行股份有限公司 | Fault detection method and device for big data cluster |
| CN111984499B (en) * | 2020-08-04 | 2024-05-28 | 中国建设银行股份有限公司 | Fault detection method and device for big data cluster |
| CN112446031A (en) * | 2020-10-26 | 2021-03-05 | 国网安徽省电力有限公司信息通信分公司 | Operation and maintenance data display platform based on artificial intelligence |
| CN112363891A (en) * | 2020-11-18 | 2021-02-12 | 西安交通大学 | Exception reason obtaining method based on fine-grained event and KPIs analysis |
| CN112363891B (en) * | 2020-11-18 | 2022-10-25 | 西安交通大学 | Method for obtaining abnormal reasons based on fine-grained events and KPIs (Key Performance indicators) analysis |
| CN114638373A (en) * | 2020-12-15 | 2022-06-17 | Aptiv技术有限公司 | Managing machine learning environment |
| CN112711508A (en) * | 2020-12-21 | 2021-04-27 | 航天信息股份有限公司 | Intelligent operation and maintenance service system facing large-scale client system |
| CN112926749A (en) * | 2020-12-30 | 2021-06-08 | 国网宁夏电力有限公司信息通信公司 | Intelligent power grid information equipment monitoring system and method |
| CN112329928A (en) * | 2020-12-30 | 2021-02-05 | 四川新网银行股份有限公司 | Heterogeneous model-based user satisfaction analysis method |
| CN112966860A (en) * | 2021-02-23 | 2021-06-15 | 武汉育知联信息科技有限公司 | Data analysis and prediction system based on attached pre-chain |
| CN112884452A (en) * | 2021-03-17 | 2021-06-01 | 北京幂数科技有限公司 | Intelligent operation and maintenance multi-source data acquisition visualization analysis system |
| CN113328878A (en) * | 2021-05-13 | 2021-08-31 | 上海励能信息技术有限公司 | Intelligent operation and maintenance monitoring system |
| CN113516447B (en) * | 2021-05-21 | 2024-04-23 | 陕西迅税通智能科技有限公司 | Electronic device and method for outputting financial tax reasoning matching result based on computer |
| CN113516447A (en) * | 2021-05-21 | 2021-10-19 | 陕西迅税通智能科技有限公司 | Electronic device and method for outputting fiscal reasoning matching result based on computer |
| CN113256180A (en) * | 2021-07-07 | 2021-08-13 | 杭州实在智能科技有限公司 | Customer service work order information intelligent dynamic loading method and system based on machine learning |
| CN113609201A (en) * | 2021-08-10 | 2021-11-05 | 珍岛信息技术（上海）股份有限公司 | Service data processing method and system |
| US11770307B2 (en) | 2021-10-29 | 2023-09-26 | T-Mobile Usa, Inc. | Recommendation engine with machine learning for guided service management, such as for use with events related to telecommunications subscribers |
| CN113934780B (en) * | 2021-12-15 | 2022-04-05 | 南京迪塔维数据技术有限公司 | Asset management system and method based on data middleboxes |
| CN113934780A (en) * | 2021-12-15 | 2022-01-14 | 南京迪塔维数据技术有限公司 | Asset management system and method based on data middleboxes |
| CN114048882A (en) * | 2022-01-12 | 2022-02-15 | 北京鼎兴达信息科技股份有限公司 | Railway fault handling operation and maintenance decision suggestion method |
| CN114721336A (en) * | 2022-03-03 | 2022-07-08 | 上海核工程研究设计院有限公司 | Information security event early warning method for technological parameters of instrument control system |
| CN114676978A (en) * | 2022-03-03 | 2022-06-28 | 北京中科智上科技有限公司 | Production intelligent decision-making system and method for oil and gas field |
| CN114721336B (en) * | 2022-03-03 | 2024-05-03 | 上海核工程研究设计院股份有限公司 | Information security event early warning method for technological parameters of instrument control system |
| CN114401398A (en) * | 2022-03-24 | 2022-04-26 | 北京华创方舟科技集团有限公司 | Intelligent video operation and maintenance management system |
| CN114662803B (en) * | 2022-05-23 | 2022-08-26 | 成都普惠道智慧能源科技有限公司 | Distributed energy storage safety monitoring method and Internet of things system |
| US11842301B1 (en) | 2022-05-23 | 2023-12-12 | Chengdu Puhuidao Smart Energy Technology Co., Ltd. | Methods for monitoring distributed energy storage safety and internet of things systems thereof |
| CN114662803A (en) * | 2022-05-23 | 2022-06-24 | 成都普惠道智慧能源科技有限公司 | Distributed energy storage safety monitoring method and Internet of things system |
| CN117376108A (en) * | 2023-12-07 | 2024-01-09 | 深圳市亲邻科技有限公司 | Intelligent operation and maintenance method and system for Internet of things equipment |
| CN117376108B (en) * | 2023-12-07 | 2024-03-01 | 深圳市亲邻科技有限公司 | Intelligent operation and maintenance method and system for Internet of things equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109343995A (en) | Intelligent O&M analysis system based on multi-source heterogeneous data fusion, machine learning and customer service robot | |
| CN106168799B (en) | A method of batteries of electric automobile predictive maintenance is carried out based on big data machine learning | |
| CN117234785B (en) | Centralized control platform error analysis system based on artificial intelligence self-query | |
| CN117391641B (en) | Pilatory production flow management method and system | |
| CN113627032B (en) | Intelligent decision-making method for equipment design/maintenance scheme based on digital twinning | |
| CN110555058A (en) | Power communication equipment state prediction method based on improved decision tree | |
| Xie et al. | Logm: Log analysis for multiple components of hadoop platform | |
| CN113268370A (en) | Root cause alarm analysis method, system, equipment and storage medium | |
| CN116992346A (en) | Enterprise production data processing system based on artificial intelligence big data analysis | |
| CN117460129B (en) | Energy-saving street lamp control method and system based on Internet of things driving | |
| CN117630758B (en) | Method and system for monitoring health state of power station transformer | |
| CN107818523B (en) | Electric power communication system data truth value distinguishing and deducing method based on unstable frequency distribution and frequency factor learning | |
| CN117933660A (en) | Energy management platform based on Internet of things | |
| CN117521498A (en) | Charging pile guide type fault diagnosis prediction method and system | |
| CN117216702A (en) | Power transformation equipment parameter processing method and device and electronic equipment | |
| CN115297016B (en) | Power network liveness evaluation and prediction method based on deep learning | |
| Zhao et al. | On-board artificial intelligence based on edge computing in optical transport networks | |
| Larrinaga et al. | Implementation of a reference architecture for cyber physical systems to support condition based maintenance | |
| CN112801815B (en) | Power communication network fault early warning method based on federal learning | |
| CN113807462A (en) | AI-based network equipment fault reason positioning method and system | |
| Wang | Intelligent quality management using knowledge discovery in databases | |
| CN102231081A (en) | Energy utilization state diagnosis method for process industrial equipment | |
| CN117527622B (en) | Data processing method and system of network switch | |
| CN117973904B (en) | Intelligent manufacturing capacity analysis method and system | |
| CN117354251A (en) | Automatic extraction method for electric power Internet of things terminal characteristics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190215 |
|
| WD01 | Invention patent application deemed withdrawn after publication |