Sentiment Analysis of Economic Text: A Lexicon-Based Approach

TL;DR

The use of the EL is illustrated in the context of a simple sentiment measure and the comparison to other lexicons shows that the EL is superior due to its wider coverage of domain relevant terms and its more accurate categorization of the word sentiment.

Abstract

We propose an Economic Lexicon (EL) specifically designed for textual applications in economics. We construct the dictionary with two important characteristics: 1) to have a wide coverage of terms used in documents discussing economic concepts, and 2) to provide a human-annotated sentiment score in the range [-1,1]. We illustrate the use of the EL in the context of a simple sentiment measure and consider several applications in economics. The comparison to other lexicons shows that the EL is superior due to its wider coverage of domain relevant terms and its more accurate categorization of the word sentiment.

Figures (9)

• Figure 1: Time series of the number of noun phrases extracted aggregated at the monthly frequency. The grey areas correspond to the recessions as indicated by National Bureau of Economic Research (NBER) for the US.
• Figure 2: Scatter plots of the sentiment scores of EL against LMD, REN and SSW dictionaries at the word level. We coded the negative/positive categories in LMD as -1 and 1. The points in the scatter plot represent only words that have a score in both lexicons. The red line represents the linear regression fit.
• Figure 3: Scatter plots of the sentiment scores of EL against LMD, REN and SSW dictionaries at the sentence level. Individual sentiment scores are coded as negative or positive depending on their sign. The red line represents the linear regression fit.
• Figure 4: Scatter plots of the sentiment scores of EL against REN and SSW dictionaries at the sentence level. The red line represents the linear regression fit.
• Figure 5: Time series of the number of positive and negative words for the EL, LMD, REN, and SSW dictionaries as a fraction of the total words included in a certain month. The fraction of negative terms is multiplied by -1.