Aerospace Instrument-Making Annotation << Back Using AutoML Technologies for Automatic Time Series Analysis Sobolevskii V.A., Zaharov V.V., Mukhametov D.I. The analysis of time series data has emerged as a crucial concern across various commercial sectors, addressing tasks such as forecasting currency metrics and detecting industrial equipment malfunctions. This paper presents an AutoML approach aimed at automating the development of machine learning models for automatic time series analysis. The proposed system involves two primary models: one focused on time series forecasting and the other on classifi cation. This ensemble of models effectively tackles a broad spectrum of practical challenges in the fi eld. The automation of structural-parametric synthesis and hyperparameter optimization is showcased through these models, implemented within the AutoGenNet software platform. Emphasizing a No-Code development paradigm, AutoGenNet allows users to navigate complex model creation and training processes with ease, thus lowering entry barriers. Furthermore, the platform includes mechanisms for generating software wrappers for trained models. Collectively, these advancements enable scalable automation of model generation and training, facilitating effi cient solutions for diverse time series analysis tasks. Keywords: deep learning, time series, AutoML, XGBoost, linear regression, ResNet. DOI: 10.25791/aviakosmos.10.2024.1434 Pp. 10-18.