Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Semantic Approach to Negation Detection and Word Disambiguation with Natural Language Processing

Izunna Okpala, Guillermo Romera Rodriguez, Andrea Tapia, Shane Halse, Jess Kropczynski

TL;DR

The paper tackles how negation hinders sentiment analysis and proposes a semantic framework that detects negation through word-sense disambiguation and applies antonymization to negated terms using an NLP knowledge base (WordHoard). The method integrates text decontraction, negation cue detection, and POS-aware sequence labeling to disambiguate negation and adjust polarity contextually. Empirical results on the Stanford Contradiction Corpora demonstrate substantial improvements across SentiWordNet (+35%), Vader (+20%), and TextBlob (+6%), indicating stronger alignment with reference sentiment. This approach enhances text classification reliability in negation-heavy text and holds practical value for domains like crisis informatics and other real-world NLP tasks.

Abstract

This study aims to demonstrate the methods for detecting negations in a sentence by uniquely evaluating the lexical structure of the text via word-sense disambiguation. The proposed framework examines all the unique features in the various expressions within a text to resolve the contextual usage of all tokens and decipher the effect of negation on sentiment analysis. The application of popular expression detectors skips this important step, thereby neglecting the root words caught in the web of negation and making text classification difficult for machine learning and sentiment analysis. This study adopts the Natural Language Processing (NLP) approach to discover and antonimize words that were negated for better accuracy in text classification using a knowledge base provided by an NLP library called WordHoard. Early results show that our initial analysis improved on traditional sentiment analysis, which sometimes neglects negations or assigns an inverse polarity score. The SentiWordNet analyzer was improved by 35%, the Vader analyzer by 20% and the TextBlob by 6%.

A Semantic Approach to Negation Detection and Word Disambiguation with Natural Language Processing

TL;DR

The paper tackles how negation hinders sentiment analysis and proposes a semantic framework that detects negation through word-sense disambiguation and applies antonymization to negated terms using an NLP knowledge base (WordHoard). The method integrates text decontraction, negation cue detection, and POS-aware sequence labeling to disambiguate negation and adjust polarity contextually. Empirical results on the Stanford Contradiction Corpora demonstrate substantial improvements across SentiWordNet (+35%), Vader (+20%), and TextBlob (+6%), indicating stronger alignment with reference sentiment. This approach enhances text classification reliability in negation-heavy text and holds practical value for domains like crisis informatics and other real-world NLP tasks.

Abstract

This study aims to demonstrate the methods for detecting negations in a sentence by uniquely evaluating the lexical structure of the text via word-sense disambiguation. The proposed framework examines all the unique features in the various expressions within a text to resolve the contextual usage of all tokens and decipher the effect of negation on sentiment analysis. The application of popular expression detectors skips this important step, thereby neglecting the root words caught in the web of negation and making text classification difficult for machine learning and sentiment analysis. This study adopts the Natural Language Processing (NLP) approach to discover and antonimize words that were negated for better accuracy in text classification using a knowledge base provided by an NLP library called WordHoard. Early results show that our initial analysis improved on traditional sentiment analysis, which sometimes neglects negations or assigns an inverse polarity score. The SentiWordNet analyzer was improved by 35%, the Vader analyzer by 20% and the TextBlob by 6%.
Paper Structure (18 sections, 5 figures, 4 algorithms)

This paper contains 18 sections, 5 figures, 4 algorithms.

Table of Contents

  1. Introduction
  2. Background
  3. Previous Work
  4. The Theory of Double Negatives
  5. Lexicon-Based Analysis for Dealing with Negations
  6. Natural Language Processing (NLP)
  7. Word Negation and Sequence Labeling
  8. Methodology
  9. Data Collection
  10. Data Pipeline
  11. Text Decontraction
  12. Negation Detection
  13. Text Antonymization
  14. Sequence Labeling, Parts of Speech (POS) Tagging, and Negation Disambiguation
  15. Results
  16. ...and 3 more sections

Figures (5)

  • Figure 1: Data Pipeline
  • Figure 2: Vader test on the original and antonymized text
  • Figure 3: TextBlob test on the original and antonymized text
  • Figure 4: Sentiwordnet test on the original the antonymized text
  • Figure 5: Correlation Heatmap of all the sentiments