A cor das palavras

Our paper relies on stock price reactions to colour words, in order to provide new dictionaries of positive and negative words in a finance context. We extend the machine learning algorithm of Taddy (2013), adding a cross-validation layer to avoid over-fitting. In head-to-head comparisons, our dictionaries outperform the standard bag-of-words approach (Loughran and McDonald, 2011) when predicting stock price movements out-of-sample. By comparing their composition, word-by-word, our method refines and expands the sentiment dictionaries in the literature. The breadth of our dictionaries and their ability to disambiguate words using bigrams both help to colour finance discourse better.

Aqui

O primeiro parágrafo é bem esclarecedor:

Since Tetlock (2007), the literature in Finance and Accounting studying different types of textual data has flourished.1 The current state of the art to measure sentiment is to use a “bag-of-words” approach, counting words in dictionaries that are specialized to Finance and Accounting jargon, namely those developed by Loughran and McDonald (2011) (LM dictionaries). This approach has been criticized as potentially having low power in comparison to more sophisticated machine learning techniques (Gentzkow et al., 2019). Our paper contributes to this debate by constructing new dictionaries using techniques from the natural language processing literature (NLP) in Computer Science, explicitly comparing their composition and predictive power relative to the LM dictionaries.

Foto: Sharon Pittaway