Description
TL;DR: A BTC/USDT forecasting experiment that changes the LSTM lookback window and compares the results with SMA, WMA, and EMA baselines.
About This Document
The study uses price data from March through April 2022 and tests several time-parameter settings for a long short-term memory model. It compares mean absolute percentage error against simple, weighted, and exponential moving averages. The authors report their lowest error when the model uses the previous three days of price data.
Why Bitcoin People May Care
Price-prediction papers age fast, which is part of their value as records. This one preserves a specific dataset window, model choice, baseline comparison, and reported 0.0927% MAPE result. It is research evidence, not a trading signal beamed backward from the future.
What You Receive
Purchase reveals the verified preserved PDF location for this document.
Document Details
- Title: Cryptocurrency Price Forecasting with Time-Varying LSTM Parameters
- Author / organization: Laor Boongasame, Panida Songram
- Year: 2023
- Document type: Machine-learning research paper
- Pages: 9
CacheRat provides researched source-location and document-identification information. CacheRat does not claim ownership of the underlying third-party documentation.
