Using the Robust Principal Component Analysis to Identify Incorrect Aerological Data

The “Middle Atmosphere” Regional Information and Analytic Center (Central Aerological Observatory) works out algorithms for analyzing the quality of aerological data based on machine learning methods. Different approaches to the data preparation are described, the examples of data that were rejected using standard approaches are given, the ways to develop and improve the quality of aerological information transmitted to the WMO international network are outlined.