Method for generating training data for a machine learning model

Abstract

According to various embodiments, a method for generating training data for a machine learning model (104), which generates a textual description for a time series supplied to it, is described, comprising, for each example time series (201) from a set of example time series, extracting values ​​(206) of time series features from the example time series (201), filling a text template (208), which has gaps for at least some of the time series features, with the extracted values ​​(206) of the time series features, generating a reformulation (210) of the filled text template by supplying the filled text template as part of a prompt to a large-language model, and generating a training example for the machine learning model (104) with the time series as training input and the reformulation (210) of the filled text template as target output for the training input.

Images