Problem Description

Time series forecasting plays a critically important in many economic contexts, with stock price predictions and sales forecasts being prominent examples. Currently, however, it remains a domain for experts. The project aims to investigate the suitability of Large Language Models (LLMs) as tutors for applying time series forecasting in economic scenarios.

Project Definition

Despite the development of various libraries for forecasting algorithms in recent years, employing these techniques in business remains challenging without the involvement of experts from this field. With the public breakthrough of ChatGPT, there is hope that such expert could eventually be fulfilled by LLMs, enabling non-specialists to successfully implement forecasting applications through dialogue with LLMs.

This project aims to examine whether the current state of LLMs is sufficient to meet these expectations. To do so, a laboratory experiment will be conducted using a comprehensive dataset from the economic sector. The task for the subjects is to develop a forecasting application with LLM support and evaluate its forecasting quality and applicability. The of the lab supervisors is to prepare the lab setup, assess the subjects knowledge of time series forecasting at the beginning and end of the project, and document their learning progress.

Learning Outcome

• Understanding of Large Language Models (LLMs) and their application in the context of time series forecasting in economic scenarios.
• Practical experience in handling time series data and prompting LLMs.
• Experimental thinking and critical evaluation of models for time series forecasting.
• Methods of scientific research, particularly in experimental lab settings, and effectively communicate and present the results.

Participation Requirements

Interest in time series forecasting, artificial intelligence, and scientific research.

External Partner

If there is an external partner, give the contact person(s) and a brief description here; otherwise leave blank