Description
TL;DR: A forecasting experiment using roughly forty candidate features, feature selection, and five model families to estimate short-horizon Bitcoin price movement.
About This Document
Amine Batsi, Mohamed Biniz, and Rachid El Ayachi assemble market and technical variables, reduce them using correlation and mutual-information measures, and compare linear regression, XGBoost, support-vector regression, Prophet, and a bidirectional gated recurrent unit. They report error and fit metrics including MSE, MAE, MAPE, and R-squared for the selected datasets and prediction horizon.
Why Bitcoin People May Care
Model comparisons can show how methods behaved on one dataset without proving a durable trading edge. Bitcoin regimes change, correlated features can leak information, and low historical error does not remove execution cost or tail risk. This is a 2025 research result, not a live forecast, investment recommendation, or guarantee that any model will perform outside the authors' experiment.
What You Receive
Purchase reveals the verified preserved PDF location for this document.
Document Details
- Title: Forecasting Bitcoin Price Fluctuations with Time-Series Models
- Author / organization: Amine Batsi, Mohamed Biniz, and Rachid El Ayachi
- Year: 2025
- Language: English
- Document type: Financial machine-learning research paper
- Pages: 12
CacheRat provides researched source-location and document-identification information. CacheRat does not claim ownership of the underlying third-party documentation.
