Aster → TimesFM Forecasting Pipeline

Overview

End-to-end crypto price forecasting: pull historical klines from Aster DEX via MCP, transform to numpy arrays, run through TimesFM 2.5 for zero-shot point forecasts with quantile prediction intervals.

When to Use

• Forecasting crypto pair prices (BTC/USDT, ETH/USDT, SOL/USDT, etc.)
• Generating trading signals with confidence intervals
• Risk assessment via quantile spreads (10th-90th percentile)
• Comparing forecast vs actual for strategy validation

Pipeline Steps

Step 1: Pull Market Data via Aster MCP