An online anomaly detection method for large-scale high-dimensional sensor data

Abstract

The invention discloses an online anomaly detection method for large-scale high-dimensional sensor data, comprising the following steps: (101), acquiring historical data; (102), establishing a deep belief network-1 / 4 spherical support vector machine hybrid model Dimensionality reduction and detection of data; (103), using historical data to train the hybrid model; (104), collecting sensor data; (105), creating a sliding window to realize online detection technology; (106), using the above trained The hybrid model detects the data collected by the sensor; (107), outputting all abnormal data after detection. The present invention improves the relevant algorithms and processes of the prior art, and proposes a method for realizing online detection technology when processing high-dimensional data, which greatly improves the accuracy rate of abnormal data detection and greatly reduces the detection time.

Description

technical field [0001] The invention relates to the field of sensor network abnormal data processing methods, in particular to an online abnormal detection method for large-scale high-dimensional sensor data. Background technique [0002] With the popularity of the Internet of Things, wireless sensor networks have been widely used in various fields. By analyzing and mining the data collected and reported by sensors, it can provide valuable and effective information for various industries. However, the complex deployment environment and the sensor's own memory, CPU, and energy conditions are very easy to cause software and hardware failures in the sensor, resulting in abnormal data, and the analysis of mixed abnormal data sets will seriously affect the mining of effective information and key decision-making. formulate. Therefore, it becomes more and more important to detect abnormal data collected by wireless sensor network accurately in real time. On the one hand, timely d...

Images