Trading on AI Insights: Evaluating Forex Predictions from Central Bank Statements [Slides]

Public Deposited

Creator
Contributors
Abstract
  • My research objective is to explore AI's capability, specifically OpenAI's GPT4, to make accurate forex market trade predictions based on monetary policy statements from the G10 Central Banks. This group represents the world's most traded foreign currency exchange markets. Using a backtesting methodology, I deployed GPT4 to interpret the nuances within each policy statement and predict currency pair movements, such as the EUR/USD forex pair, when comparing the US Federal Reserve with the European Central Bank. GPT4's role as a currency trader involved classifying the policy stance of each bank and making a trade recommendation. These recommendations were tested against historical price data to gauge profitability. The backtest yielded a remarkable 65-70% success rate in profitable trade predictions, suggesting the potential for AI to significantly impact financial trading markets. This study underscores AI's transformative potential within the currency trading and broader investment sphere. Despite successful backtesting, it's important to recognize that past performance may not predict future results.

Language
Keyword
Date created
Resource type
Rights statement

MLA citation style (9th ed.)

Jessup, Matthew. Trading On Ai Insights: Evaluating Forex Predictions From Central Bank Statements [slides]. Dunaway, Eric.. 2024. wabash.hykucommons.org/concern/generic_works/4353bb00-49e0-49ce-827b-316d95b06e19?locale=en.

APA citation style (7th ed.)

J. Matthew. (2024). Trading on AI Insights: Evaluating Forex Predictions from Central Bank Statements [Slides]. https://wabash.hykucommons.org/concern/generic_works/4353bb00-49e0-49ce-827b-316d95b06e19?locale=en

Chicago citation style (CMOS 17, author-date)

Jessup, Matthew. Trading On Ai Insights: Evaluating Forex Predictions From Central Bank Statements [slides]. 2024. https://wabash.hykucommons.org/concern/generic_works/4353bb00-49e0-49ce-827b-316d95b06e19?locale=en.

Note: These citations are programmatically generated and may be incomplete.