RU
DonNTU   Master's Portal

Library of Materials

    Own Publications and Reports

  1. Forecasting Public Transport Passenger Flow Using Machine Learning Algorithms

    Authors: V. O. Savenkova, E. O. Savkova

    Description: This article explores the issue of forecasting public transport passenger flow using machine learning algorithms. Existing forecasting methods are analyzed. The stages of preparing input data, as well as algorithms for model training and development stages, are discussed.

    Source: Proceedings of the XV International Scientific and Technical Conference "Informatics, Control Systems, Mathematical and Computer Modeling - 2024 Donetsk", DonNTU 2024, Vol. 2, pp. 1156-1161

  2. Regional Features of Implementing Machine Learning in Public Transport Passenger Flow Forecasting Projects

    Authors: V. O. Savenkova, E. O. Savkova

    Description: This article investigates regional features of implementing machine learning (ML) for passenger flow forecasting in public transport systems.

    Source: IX International Scientific and Practical Conference "Business Engineering of Complex Systems: Models, Technologies, Innovations - BECS-2024"

    3. Thematic Articles

  3. Overview of Open-Source Libraries for Time Series Forecasting Tasks

    Authors: E.A. Svekolnikova, V.N. Panovskiy

    Description: This article provides an overview of various open-source Python libraries for time series analysis and forecasting. Tools such as Prophet, Kats, Merlion, as well as ARIMA, LSTM algorithms for studying seasonality, trends, and anomalies in time series data are covered. The capabilities, advantages, and application areas of each library in time data analysis are discussed in detail.

    Source: Svekolnikova E.A., Panovskiy V.N. Overview of Open-Source Libraries for Time Series Forecasting Tasks // Modeling and Data Analysis. 2024. Vol. 14. No. 2. P. 45–61. DOI: 10.17759/mda.2024140203 URL: https://psyjournals.ru/journals/mda/archive/2024_n2/Svekolnikova_Panovskiy

  4. Study of Time Series Forecasting Method in Transport Using Recurrent Neural Networks

    Authors: G. M. Lysov, F. N. Prikhodko, A. A. Konovalova, K. A. Timoshenko

    Description: For effective resource planning needed for transport sector projects and decisions, it is crucial to accurately determine the performance indicators of transport objects (passenger flow, freight flow, etc.). Accurate forecasts are made possible through time series forecasting using recurrent neural networks. This article considers three forecasting models: ARIMA, LSTM, PROPHET. The advantages and disadvantages of each model are defined.

    Source: Study of Time Series Forecasting Method in Transport Using Recurrent Neural Networks / G. M. Lysov, F. N. Prikhodko, A. A. Konovalova, K. A. Timoshenko // Science Diary. – 2023. – ¹ 1(73). – DOI 10.51691/2541-8327_2023_1_6. – EDN OAYFDH. URL: https://www.elibrary.ru/download/elibrary_50437580_42693612.pdf

  5. Comparative Testing of ARIMA and LSTM Models in Passenger Flow Forecasting Tasks

    Authors: M. A. Yakimov, K. V. Operaylo, E. N. Novikova

    Description: To meet the demand for public transport in real-time, bus operators need to adjust the schedule on time. Therefore, short-term passenger flow variations must be forecasted. This article examines the implementation and comparison of two forecasting methods, ARIMA and LSTM.

    Source: Yakimov M. A., Operaylo K. V., Novikova E. N. Comparative Testing of ARIMA and LSTM Models in Passenger Flow Forecasting Tasks // Science Symbol. 2022. No. 6-2. URL: https://cyberleninka.ru/article/n/sravnitelnoe-testirovanie-modeley-arima-i-ltsm-v-zadachah-prognozirovaniya-passazhiropotoka.