基于机器学习和LSTM深度学习模型的Mobile Legends应用评论情感分析
计算与语言
2026-05-05 v1
摘要
本文比较了机器学习和LSTM基深度学习方法，用于Mobile Legends应用评论的情感分析。使用包含10,000条标记为积极、消极和中性评论的数据集，研究评估了基于TF-IDF和PyCaret AutoML的传统模型，并将其与设计用于捕获序列文本依赖关系的LSTM模型进行比较。结果表明，LSTM模型在准确率达到92%和加权F1分数达到91%时优于经典机器学习基线。研究发现，深度学习在处理非正式且具有上下文依赖性的用户评论文本方面更有效。
关键词
引用
@article{arxiv.2605.01317,
title = {Sentiment Analysis of Mobile Legends App Reviews Using Machine Learning and LSTM-Based Deep Learning Models},
author = {Vira Putri Maharani and Kharisa Harvanny and Daris Samudra and Luluk Muthoharoh and Ardika Satria and Martin Clinton Tosima Manullang},
journal= {arXiv preprint arXiv:2605.01317},
year = {2026}
}
备注
8 pages, 3 figures, includes comparative evaluation of Machine Learning and LSTM models for sentiment analysis on Indonesian Mobile Legends app reviews, with dataset description, methodology, model architecture, results, discussion, acknowledgments, and references