Create news_llm_analysis_prompt.md

Author: Jimmy2099

llm_prompt/news_llm_analysis_prompt.md

@@ -0,0 +1,18 @@
+# news_llm_analysis_prompt fast data cleaning through gemma3:1b
+```
+Below is a prompt template for analyzing news data and generating investment-related information. You can adjust the content based on specific needs:
+Prompt:
+Please analyze the following news content and generate a structured output in the following format:
+Article Sentiment: Determine the overall sentiment tendency of the news (positive, negative, or neutral).
+Affected Entities: List only the companies, entities, or individuals mentioned in the news. Use full names (no abbreviations) and limit to the top three most important entities.
+Categories: Based on the news content, select relevant category tags (e.g., management capability, market competitiveness, financial performance, customer satisfaction, etc.). Limit to the top six most important categories.
+Output in JSON format:
+{
+"Article Sentiment": "positive/negative/neutral",
+"Affected Entities": ["Company A", "Company B"],
+"Categories": ["Category 1", "Category 2"]
+}
+News Content:
+[{{#1711708591503.input_text#}}]
+```
+