Using Natural Language Processing to Enhance Understandability of Financial Texts

EasyChair Preprint 9252, version 1

Abstract

Dealing with money has always been one of the basic skills one needs to live a comfortable life. However, financial literacy rates across the nations are extremely low. Furthermore, over the years the returns from traditional investment avenues like bank fixed deposits (FD), real estate, etc. have been diminishing. This entices new-age investors to trade and reap profits from the ever-growing stock markets. Nevertheless, in reality, only a handful of active traders are able to earn more than the FD rates. This is due to the lack of financial knowledge. The presence of complex concepts and jargons further reduces comprehensibility. In this paper, we present how financial texts can be demystified using Natural Language Processing (NLP). It consists of neural-based readability assessment and hypernym extraction tools to improve the readability of financial texts. Other modules include financial domain specific systems for automated claim detection, sustainability assessment, etc.

Keyphrases: Natural Language Processing, claim detection, financial text processing, hypernym detection, readability

@booklet{EasyChair:9252,
  author    = {Sohom Ghosh and Sudip Kumar Naskar},
  title     = {Using Natural Language Processing to Enhance Understandability of Financial Texts},
  howpublished = {EasyChair Preprint 9252},
  year      = {EasyChair, 2022}}