Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Method for detecting abnormal time sequence without class label

A technology of category labeling and time series, applied in the field of time series anomaly detection, can solve the problems of unsatisfactory segmentation effect of satellite telemetry data fixed point and manually setting the number of clusters for hierarchical clustering, so as to achieve compact and coupled segmentation results. high degree of effect

Active Publication Date: 2015-09-09
HARBIN INST OF TECH
View PDF5 Cites 27 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the unsatisfactory effect of segmenting fixed points for satellite telemetry data, the need to manually set the number of clusters for hierarchical clustering, and the fact that there is currently no directly available offline and label-free time series. The problem of online anomaly detection method framework, and propose a time series anomaly detection method without category labels

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
  • Method for detecting abnormal time sequence without class label
  • Method for detecting abnormal time sequence without class label
  • Method for detecting abnormal time sequence without class label

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0036] Specific implementation mode one: combine figure 1To illustrate this embodiment, a time series anomaly detection method without category labels is specifically carried out according to the following steps:

[0037] Step 1. Segment the historical satellite telemetry data according to the periodic characteristics of the satellite telemetry data, and obtain a time series without class labels X={x 1 ,x 2 ,...,x n}, where n is a positive integer greater than 0, indicating the number of time series, x 1 is the first time series in the time series without category labels, x 2 is the second time series in the time series without class labels, x n is the nth time series in the time series without class labels;

[0038] Step 2. For the time series without class labels obtained in step 1, X={x 1 ,x 2 ,...,x n} Carry out adaptive hierarchical clustering, and determine and delete the abnormal sequence in the time series without category labels, and obtain the time series wi...

specific Embodiment approach 2

[0040] Specific embodiment 2: The difference between this embodiment and specific embodiment 1 is that in the step 1, the satellite telemetry historical data is segmented according to the periodic characteristics of the satellite telemetry data, and the time series X={x without category labels is obtained 1 ,x 2 ,...,x n}, where n is a positive integer greater than 0, indicating the number of time series, x 1 is the first time series in the time series without category labels, x 2 is the second time series in the time series without class labels, x n is the nth time series in the time series without class labels; the specific process is: segment the historical data of satellite telemetry with the point of change in argument as the mark, and obtain the time series without class labels X={x 1 ,x 2 ,...,x n}.

[0041] Argument is one of the test parameters in satellite telemetry data. Its value changes in order from 0 to 360. When it reaches 360, it becomes 0 and starts to ...

specific Embodiment approach 3

[0043] Specific embodiment three: the difference between this embodiment and specific embodiment one or two is: in the step two, the time series without class labels obtained in step one X={x 1 ,x 2 ,...,x n} Carry out adaptive hierarchical clustering, and determine and delete the abnormal sequence in the time series without category labels, and obtain the time series with category labels under the normal operation mode of the satellite and category labels where n z is a positive integer greater than 0, indicating the number of normal time series, x' 1 is the first normal time series in the time series with category labels, x' 2 is the second normal time series in the time series with class labels, is the nth time series with category labels z a normal time series, l' 1 is the first normal time series in the category label, l' 2 is the second normal time series in the category labels, is the nth in the category label z A normal time series; the specific process i...

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 method for detecting an abnormal time sequence without a class label, and aims at solving the problems that ideal effect of segmenting fixed points based on satellite remote detecting data cannot be achieved, the clustering number is manually set during layer-based clustering, and offline and online abnormality detection methods for the label time sequence without the class label are currently not developed. According to the technical scheme, the method comprises the steps of 1, segmenting the satellite remote detecting historical data according to the cycle property of the satellite remote detecting data to obtain the time sequence without class label, namely, X={x1, x2..., xn}; 2, performing adaptive layer-based clustering for the X={x1, x2..., xn} obtained in step 1, and determining and deleting the abnormal sequence in the time sequence without the class label to obtain the formulas as shown in the specification; 3, adopting the formulas as shown in the specification as samples, performing mode matching for the formula shown in the specification by the nearest neighbor algorithm according to the matching threshold, so as to finish the abnormal satellite remote detecting data detection. The method is applied to the field of satellite data detection.

Description

technical field [0001] The invention relates to a time series anomaly detection method without category labels. Background technique [0002] By analyzing the yaw attitude angle in the satellite telemetry data, the overall change trend of the yaw attitude angle is as follows: figure 2 As shown, its details change as image 3 As shown, the satellite telemetry data has obvious periodicity, and this characteristic has been confirmed with the satellite telemetry data provider. By analyzing each period of the telemetry data, it can be concluded whether the satellite’s operating status within the period is normal, and the effect of segmenting the satellite telemetry data according to the fixed point is not ideal, such as Figure 4 As shown, the coupling degree between each sub-sequence is not high enough, there is a certain deviation, and this deviation will become more and more obvious as time goes on. [0003] At present, there are no clear reference materials for the normal ...

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
IPC IPC(8): G06F17/30G06K9/62G06K9/66
CPCG06F16/285G06V30/194G06V2201/10G06F18/22
Inventor 刘大同彭宇陈静张玉杰彭喜元
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
Eureka Blog
Learn More
PatSnap group products