Bielinskyi, AndriiSoloviev, VolodymyrSolovieva, ViktoriiaVelykoivanenko, HalynaВеликоіваненко, Галина ІванівнаВеликоиваненко, Галина Ивановна2024-04-012024-04-012022Fuzzy time series forecasting using semantic artificial intelligence tools / Andrii Bielinskyi, Vladimir Soloviev, Viktoria Solovieva, Halyna Velykoivanenko // Нейро-нечіткі технології моделювання в економіці : наук.-анал. журн. / М-во освіти і науки України, Київ. нац. екон. ун-т ім. Вадима Гетьмана ; [редкол.: А. В. Матвійчук (голов. ред.) та ін.]. – Київ : КНЕУ, 2022. – № 11. – С. 157–198.2306-3289https://ir.kneu.edu.ua/handle/2010/43397This study investigates the application of Fuzzy Time Series (FTS) methods in forecasting the Bitcoin market. FTS methods have gained significant attention due to their simplicity, adaptability, forecasting precision, and computational efficiency. They generate interpretable representations of time series patterns, enabling knowledge transfer, auditability, reusability, and upgradability. The study specifically focuses on time-invariant rule-based FTS techniques, namely the conventional First-Order FTS (Song and Chen) and Weighted First-Order FTS (Yu) models. The research rigorously evaluates and compares the predictive performance of these methods across a range of accuracy metrics. Additionally, the article expands the understanding and application of FTS methods in cryptocurrency forecasting. Through comprehensive experimental evaluations and statistical analyses, it uncovers insights into the strengths, limitations, and potential areas for improvement of these FTS approaches. By highlighting their comparative accuracy and computational efficiency, the research contributes to the existing studies and provides practical recommendations for researchers and practitioners in the cryptocurrency domain.enfuzzy time seriesfuzzy setfuzzy logical relationships groupfuzzy time series forecastingBitcoincryptocurrency marketArticle10.33111/nfmte.2022.157