Model Card for Sentiment Analysis on Financial News
Overview
This dataset contains sentiments for financial news headlines from the perspective of a retail investor. The data is derived from the research by Malo et al. (2014), which focuses on detecting semantic orientations in economic texts.
Dataset Details
- Source: Malo, P., Sinha, A., Takala, P., Korhonen, P., and Wallenius, J. (2014). “Good debt or bad debt: Detecting semantic orientations in economic texts.” Journal of the American Society for Information Science and Technology.
- Sentiment Distribution:
- Neutral: 59%
- Positive: 28%
- Negative: 12%
Example Headlines
- Neutral: "According to Gran, the company has no plans to move all production to Russia, although that is where the company is growing."
- Negative: "The international electronic industry company Elcoteq has laid off tens of employees from its Tallin office."
- Positive: "With the new production plant, the company would increase its capacity to meet the expected increase in demand."
Additional Information
- The dataset includes various financial news headlines categorized by sentiment, which can be useful for training sentiment analysis models in the finance domain.
- The headlines reflect real-world events and their perceived impact on the market, making this dataset valuable for research and practical applications in financial analytics.