Prompt Engineering for Time-Series Analysis with Large Language Models

Prompt engineering can significantly enhance time-series analysis when using Large Language Models (LLMs). The guide outlines core strategies like patch-based, zero-shot, and neighbor-augmented prompting for forecasting tasks. It provides specific, structured prompts for essential preprocessing steps, including stationarity testing, autocorrelation analysis, and seasonal decomposition. Furthermore, the content details methods for anomaly detection and automated feature engineering, complete with Python code examples to streamline implementation.