A carregar...
Publicação
Cryptocurrency forecasting
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 950.33 KB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
This thesis aims to investigate the factors influencing daily cryptocurrency returns and
assesses the feasibility of forecasting these returns using a diverse set of 25 globally traded
assets as predictors between 2017 and 2022. Inspired by a New York Times article on
cryptocurrency bubbles and market volatility, the investigation undergoes several statistical
computations such as the Principal Component Analysis and the Complete Subset
Regression to observe the data from distinct angles. Notably, the FX-rate USD/CNY and the
Japanese NIKKEI225 index emerge as consistently influential predictors within the out-of sample forecasting periods, however, the significance of the Out-of-Sample R-squared
remains low. This research contributes to the understanding of cryptocurrency market
dynamics by examining the impact of a wide range of predictors and aligns with recent
academic findings regarding the challenges in forecasting cryptocurrencies. Further research
should focus on market sentiments regarding potential price movements in the crypto space.
Descrição
Palavras-chave
Forecasting Cryptocurrency market Complete subset regression Trading strategy Principal component analysis