# FEUP at SemEval-2017 Task 5: Predicting Sentiment Polarity and Intensity with Financial Word Embeddings

## TL;DR

This paper describes a regression-based approach using financial domain-specific word embeddings to predict sentiment polarity and intensity in financial microblogs and news, achieving competitive scores in SemEval 2017.

## Contribution

It introduces the use of financial word embeddings combined with traditional features for fine-grained sentiment analysis in financial texts.

## Key findings

- Achieved cosine similarity scores of 0.69 and 0.68 in two sub-tasks.
- Demonstrated the effectiveness of financial word embeddings for sentiment prediction.
- Combined lexical features with domain-specific embeddings for improved results.

## Abstract

This paper presents the approach developed at the Faculty of Engineering of University of Porto, to participate in SemEval 2017, Task 5: Fine-grained Sentiment Analysis on Financial Microblogs and News. The task consisted in predicting a real continuous variable from -1.0 to +1.0 representing the polarity and intensity of sentiment concerning companies/stocks mentioned in short texts. We modeled the task as a regression analysis problem and combined traditional techniques such as pre-processing short texts, bag-of-words representations and lexical-based features with enhanced financial specific bag-of-embeddings. We used an external collection of tweets and news headlines mentioning companies/stocks from S\&P 500 to create financial word embeddings which are able to capture domain-specific syntactic and semantic similarities. The resulting approach obtained a cosine similarity score of 0.69 in sub-task 5.1 - Microblogs and 0.68 in sub-task 5.2 - News Headlines.