Gathering abnormity detection method for single sensor data flow

An anomaly detection and single data stream technology, applied in the field of anomaly detection, can solve the problems that the anomaly detection method cannot meet the real-time requirements, and achieve the effect of reducing the amount of calculation and increasing the efficiency of algorithm execution

Active Publication Date: 2013-10-09
HARBIN INST OF TECH
View PDF0 Cites 23 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problem that the existing anomaly detection method cannot meet the real-time requirements, the present invention proposes an aggregation anomaly detection method oriented to a sensor single data stream

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
  • Gathering abnormity detection method for single sensor data flow
  • Gathering abnormity detection method for single sensor data flow
  • Gathering abnormity detection method for single sensor data flow

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0058] Specific implementation mode 1. Combination Figure 1 to Figure 3 This embodiment is specifically described. The sensor single data flow-oriented aggregation anomaly detection method described in this embodiment includes the following steps:

[0059] Step A. Offline aggregation anomaly detection:

[0060] Step 1. Determine the type of the mean function and covariance function of the Gaussian process regression model through the offline training data collected by the sensor for the data stream to be detected, and set the window size of the offline training data, and perform step 2;

[0061] Step 2. Set the sampling ratio of the uniform single-chain sampling method, and perform step 3;

[0062] Step 3. Use the window size of the offline training data in step 1 and the sampling ratio set in step 2 as the input parameters of the uniform single chain sampling method, and obtain the sampling of the data according to the offline data collected by the sensor in step 1 sample,...

specific Embodiment approach 2

[0162] Specific embodiment two, combine Figure 4 Describe this embodiment in detail. The difference between this embodiment and the sensor single data stream-oriented aggregation anomaly detection method described in Embodiment 1 is that the training data set described in Step 4 is used as the Gaussian process regression model in Step 6. For the training data pair, the initial value of the parameter set in step 5 is used as the initial value of the hyperparameter of the Gaussian process regression model to train the Gaussian process regression model, and the normalized index value in the training data set is used as the post-training The specific process of obtaining the mean value and variance output data corresponding to the predicted input data is as follows:

[0163] Step a. Obtain the training data of the Gaussian process regression model according to the set sampling ratio, including sampling samples and their corresponding index values;

[0164] Step b, determining th...

specific Embodiment approach 3

[0167] Specific embodiment three, combine Figure 5 Describe this embodiment in detail. The difference between this embodiment and the sensor single data stream-oriented aggregation anomaly detection method described in the first embodiment is that the uniform single-chain sampling method includes the following steps:

[0168] Step e, start, receive data stream; execute step f;

[0169] Step f, judging whether the received data element is the first data element, if so, execute step g, if not, execute i;

[0170] Step g, store the data element as a sample element, and execute step h;

[0171] Step h, randomly select an index value in the next small window, and execute step i;

[0172] Step i, judging whether the data element with the smallest index value among all the sampling sample elements obtained after being stored in step g is expired, if so, execute step j; if not, execute step k;

[0173] Step j, delete the data element with the smallest index among all the sampling ...

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 discloses a gathering abnormity detection method for single sensor data flow, which is used for solving the problem that a conventional abnormity detection method cannot meet the requirement on real-time performance. According to the gathering abnormity detection method for the single sensor data flow, a mode that off-line gathering abnormity detection and on-line gathering abnormity detection in real time are combined is employed, original data modeling is effectively carried out by combining a training subset, and the applicability of a regression forecast model to the gathering abnormity of the single sensor data flow in a sampling gaussian process is verified through real data. The method is applicable to the field of abnormity detection.

Description

technical field [0001] The invention relates to an anomaly detection method, in particular to an aggregation anomaly detection method oriented to a sensor single data stream. Background technique [0002] Since sensors are generally used in test equipment and their data is transmitted through a communication network, the data is easily corroded, and undetected errors will have a greater impact on the real-time analysis of data values. Therefore, NSF (National Science Foundation) has put forward clear requirements for self-improvement and control of data quality. Anomaly detection is used to identify patterns in data that deviate significantly from historical models. Abnormal data in the sensor is caused by errors in the sensor itself or in data transmission, or less frequently abnormal behavior of the system, and these anomalies are of great interest to users. For anomalies in sensors relative to historical patterns—when aggregated anomalies occur (aggregated anomalies gen...

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): G06F19/00
Inventor 彭宇庞景月潘大为刘大同彭喜元
Owner HARBIN INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap