Reed at SemEval-2020 Task 9: Fine-Tuning and Bag-of-Words Approaches to\n Code-Mixed Sentiment Analysis

2020/07/26 by Vinay Gopalan, Mark Hopkins, Gopalan, Vinay +1
Computer Science · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Natural Language Processing Techniques #Sentiment Analysis and Opinion Mining #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2007.13061

openalex publication_date 2020/07/26 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28

Abstract

We explore the task of sentiment analysis on Hinglish (code-mixed\nHindi-English) tweets as participants of Task 9 of the SemEval-2020\ncompetition, known as the SentiMix task. We had two main approaches: 1)\napplying transfer learning by fine-tuning pre-trained BERT models and 2)\ntraining feedforward neural networks on bag-of-words representations. During\nthe evaluation phase of the competition, we obtained an F-score of 71.3% with\nour best model, which placed 4^th out of 62 entries in the official system\nrankings.\n

Related

• SemEval-2020 Task 9: Overview of Sentiment Analysis of Code-Mixed Tweets
• Deep Learning Brasil -- NLP at SemEval-2020 Task 9: Overview of Sentiment Analysis of Code-Mixed Tweets
• LIMSI_UPV at SemEval-2020 Task 9: Recurrent Convolutional Neural Network for Code-mixed Sentiment Analysis
• IUST at SemEval-2020 Task 9: Sentiment Analysis for Code-Mixed Social\n Media Text using Deep Neural Networks and Linear Baselines
• HCMS at SemEval-2020 Task 9: A Neural Approach to Sentiment Analysis for\n Code-Mixed Texts