Air quality anomaly detection method based on distributed online principal component analysis

Abstract

The invention discloses an air quality anomaly detection method based on distributed online principal component analysis. The method comprises acquiring known air pollutant concentration data of a city, designing a distributed online principal component analysis model, and processing air pollutant concentration data through the air quality anomaly detection method based on distributed online principal component analysis to obtain a judgment result so that the air quality anomaly is detected. The method can conveniently predict the overall air quality in a certain range, can suppress the influence caused by the abnormal training data, has the function of denoising, and has a fast processing speed and shorter time.

Description

technical field [0001] The invention belongs to the field of distributed signal processing, machine learning and anomaly detection, in particular to an air quality anomaly detection method based on distributed online principal component analysis. Background technique [0002] Principal Component Analysis (PCA) is a commonly used unsupervised dimensionality reduction method. The main direction of the data distribution can be obtained by PCA. Conventional PCA implements the principal component analysis process by constructing the data covariance matrix and calculating its eigenvectors. These eigenvectors are the more informative directions in the original data space and are therefore considered principal components or principal directions. For conventional PCA, it is generally necessary to obtain all the original data and generate the covariance matrix of the original data. Obviously, the principal component analysis method can be applied to the field of anomaly detection, ...

AI Technical Summary

Problems solved by technology

In this case, the data is scattered and collected by multiple sensor nodes, so the conventional PCA m

Images