TASS 2020: TWilBERT for Sentiment Analysis and Emotion Detection in Spanish Tweets

Abstract

使用Bert在西班牙语tweet中进行情感分类，双向Bert

Introduction

使用模型bert，基于对西班牙语tweet的Bert模型的预训练适应性的微调。介绍本文结构，第二节介绍解决了的任务，第三节提出了一些设想，和baseline model。第四节，对实验进行评估评价和实验的结果进行分析，最后第5节显示了一些结论和未来的工作。

对数据集进行分析

In Table 2 the tweet distribution for each

emotion in the training set of the task 2 is shown. In this case, there is a large bias towards the

class Others that acts like a sink of unconsidered emotions or combinations among emotions.

The less frequent class, by far, is the Fear class.

Task description

Task1：

Task2：The second task is also a single-label classification task

but with 7 different emotions ( joy , sadness , anger , surprise , disgust , fear and others ).

 

Systems

• Deep Averaging Networks：以前用过的版本作为baseline，该模型在averaging word embedding的基础上运用 feed-forward networks。
• TWilBERT：TWilBERT是在tweet领域中用于训练，评估，和微调的Bert模型，介绍一下TWilBERT，并与multi-lingual Bert进行比较，说明TWilBERT的优点。

Experimental Word

TwilBert与baseline(DAN)通过不同的设置进行尽可能公平的对比。