Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Real-time KPI data exception detection method and device based on machine learning

A data anomaly and machine learning technology, applied in the field of intelligent operation and maintenance, can solve problems such as system abnormalities and software bugs, and achieve the effects of guaranteeing service quality, improving operation and maintenance management capabilities, good versatility and practicability

Active Publication Date: 2021-01-05
上海蒙帕智能科技股份有限公司 +1
View PDF10 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Abnormal points in real-time monitored KPI data are often detected through traditional threshold settings, but this detection needs to rely on the rich experience of operation and maintenance personnel for support, but as time goes by, actual system abnormalities or software There may be a lot of bugs, and it is almost impossible to detect by manually setting thresholds

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Real-time KPI data exception detection method and device based on machine learning
  • Real-time KPI data exception detection method and device based on machine learning
  • Real-time KPI data exception detection method and device based on machine learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0058] figure 1 It is a schematic flowchart showing a real-time data anomaly detection method based on machine learning KPI indicators according to an embodiment of the present invention, see figure 1 As shown, the method may include the following steps:

[0059] Online real-time monitoring to obtain KPI data;

[0060] Substituting the KPI data into multiple unsupervised models for screening;

[0061] Putting the data points that are suspiciously abnormal after being screened by all unsupervised models into the first data set;

[0062] Substituting the data points of the first data set into the first detection model for judgment, and storing all unsupervised model screening results as normal and the first detection model judgment result as normal KPI data into the historical database;

[0063] Output judgment result.

Embodiment 2

[0065] figure 2 It is a schematic flowchart showing the establishment of the first detection model of a real-time data anomaly detection method based on machine learning KPI indicators according to an embodiment of the present invention. On the basis of Embodiment 1, this embodiment also has the following content:

[0066] see figure 2 As shown, the establishment of the first detection model in the method may include the following steps:

[0067] Obtain historical monitoring KPI data in the data center operation and maintenance system;

[0068] Perform quality analysis and preprocessing on the acquired KPI data;

[0069] Perform anomaly detection on the preprocessed KPI data according to multiple unsupervised models. When all unsupervised models judge that the current data point is normal, mark the current data point as normal, otherwise mark it as suspected abnormal;

[0070] Establish a training model, and construct an independent forest anomaly detection model (the fir...

Embodiment 3

[0075] image 3 It is a schematic flow diagram showing a data anomaly detection method based on machine learning KPI indicators according to an embodiment of the present invention, see image 3 As shown, on the basis of Embodiment 2, this embodiment also includes the following content:

[0076] Acquisition of KPI data, in chronological order, centrally collects and obtains KPI data monitored by the system.

[0077] The quality analysis and preprocessing of KPI data belong to the data exploration and preparation stage: explore the collected KPI data, better understand its data characteristics, perform quality checks; then preprocess it, that is, convert it into structured data , filter out dirty data, fill missing values ​​and remove noise in KPI data.

[0078] Anomaly detection model training, based on multiple unsupervised models to filter out normal data points in the above preprocessed data, and use the normal data points in these historical data to build an independent f...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a real-time KPI data exception detection method and device based on machine learning. The real-time KPI data exception detection method comprises the following steps: acquiringKPI data in real time; substituting the KPI data into a plurality of unsupervised models for screening; putting the data points which are screened by all the unsupervised models and have suspected anomalies into a first data set; substituting the data points of the first data set into a first detection model for judgment; and outputting judgment results. According to the method, the defects that in a traditional operation and maintenance mode of an enterprise, a large number of thresholds are manually set, and high participation of a user is needed are overcome, exceptions can be found in timefrom massive KPI data, rapid exception screening can be manually completed, and the emergency response capacity of a detection system is improved.

Description

technical field [0001] The invention relates to the technical field of intelligent operation and maintenance, in particular to a machine learning-based real-time KPI data anomaly detection method and device. Background technique [0002] With the development of the Internet and mobile Internet, the scale of enterprise information technology infrastructure construction continues to expand. How to ensure the stability and security of various services and systems of online products, as well as how to efficiently diagnose and locate problems will become an issue for enterprises. The core problems faced, and the traditional technical architecture and operation and maintenance methods have been unable to effectively solve the existing problems. In order to make up for the deficiencies in the traditional technical architecture and operation and maintenance methods, the concept of AIOps (Artificial Intelligence for IT Operations, intelligent operation and maintenance) has been propo...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): H04L12/24H04L12/26G06N20/00G06K9/62
CPCH04L41/145H04L41/069H04L43/08G06N20/00G06F18/24323
Inventor 韩丹东虎
Owner 上海蒙帕智能科技股份有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com