Create README.md

README.md

@@ -0,0 +1,17 @@
+This dataset contains 50,000 samples of user financial queries paired with their corresponding structured data requests (context). It was created to facilitate the fine-tuning of Large Language Model (LLM)-based Financial Agents for accurate data extraction and query answering.
+
+## Dataset Construction
+
+# Diverse Query Sources
+The dataset was built using a multi-method query sampling approach, incorporating:
+
+- Real-world queries from retail customer technology channels (e.g., email, web forms).
+
+- Consultations with 10 financial advisors on common retail investor inquiries.
+
+- Interviews with 20 retail investors about their pre-investment questions.
+
+# Template-Based Expansion:
+5,000 query templates were generated (4,000 from technology channels, 500 from advisors, and 500 from investors). These templates were then scaled up to 50,000 samples by randomly varying elements within each template, ensuring a broad and varied representation of financial queries.
+
+This dataset provides a valuable resource for developing and evaluating LLMs in the financial domain, particularly for applications requiring accurate understanding and response to user financial queries.