Ashchepkov V. Processing of liquid flow measurement results using machine learning

Aschepkov V.O. Processing of Liquid Flow Measurement Results Using Machine Learning. - Qualification scientific work as a manuscript. Dissertation for the degree of Candidate of Technical Sciences (Doctor of Philosophy) in specialty 152 “Metrology and Information-Measuring Technology” (15 - Automation and Instrumentation). - Kharkiv National University of Radio Electronics, Kharkiv, 2024. The dissertation focuses on the development of approaches for processing liquid flow measurement results using machine learning methods to detect outliers and improve the quality of metrological data. The main emphasis is on improving the processes of identifying and excluding outliers from measurement data, which allows increasing the accuracy and stability of results. The purpose of the dissertation is to enhance the stability of liquid flow measurements and the reliability of metrological data by developing and adapting an algorithm based on the isolation forest model. Eliminating excessive errors reduces standard uncertainty of type A, contributing to the reliability of obtained results. Object of the research: the process of measuring liquid flow and processing metrological data. Subject of the research: the detection of outliers in metrological data and their impact on the stability and reliability of measurement results.