Divide (Text) and Conquer (Sentiment): Improved Sentiment Classification by Constituent Conflict Resolution

Jan Kościałkowski, Paweł Marcinkowski
5/8/2025
cs.LGcs.AIcs.CL

No official code found. Best implementation: jkoscialkowski/divide-and-conquer-sentiment (0 stars, MIT, last commit 2025-02-08).

Abstract

Sentiment classification, a complex task in natural language processing, becomes even more challenging when analyzing passages with multiple conflicting tones. Typically, longer passages exacerbate this issue, leading to decreased model performance. The aim of this paper is to introduce novel methodologies for isolating conflicting sentiments and aggregating them to effectively predict the overall sentiment of such passages. One of the aggregation strategies involves a Multi-Layer Perceptron (MLP) model which outperforms baseline models across various datasets, including Amazon, Twitter, and SST while costing $\sim$1/100 of what fine-tuning the baseline would take.

View on arXivView PDF

Code Implementations(1)

jkoscialkowski/divide-and-conquer-sentimentCommunity100%
MIT
00Python, Jupyter NotebookNov 29, 20241 years agoMIT

Similar Papers

Based on Data Balancing and Model Improvement for Multi-Label Sentiment Classification Performance Enhancement

83% match2025

Comprehensive Study on Sentiment Analysis: From Rule-based to modern LLM based system

82% match2024

Lexicon-Based Sentiment Analysis on Text Polarities with Evaluation of Classification Models

81% match2024

Dynamic Span Interaction and Graph-Aware Memory for Entity-Level Sentiment Classification

81% match2025

Multi-refined Feature Enhanced Sentiment Analysis Using Contextual Instruction

80% match2025

Multi-domain Multilingual Sentiment Analysis in Industry: Predicting Aspect-based Opinion Quadruples

80% match2025

Cite this paper

@article{kociakowski2025divide,
  title  = {Divide (Text) and Conquer (Sentiment): Improved Sentiment Classification by Constituent Conflict Resolution},
  author = {Jan Kościałkowski and Paweł Marcinkowski},
  year   = {2025},
  eprint = {2505.06320},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2505.06320v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2505.06320.svg)](https://codeofpaper.com/paper/2505.06320)

Discussion