Social networks such as twitter have emerged as social platforms that can impart a massive knowledge base for people to share their unique ideas and perspectives on various topics and issues with friends and families. Sentiment analysis based on machine learning has been successful in discovering the opinion of the people using redundantly available data. However, recent studies have pointed out that imbalanced data can have a negative impact on the results. In this paper, we propose a framework for improved sentiment analysis through various ordered preprocessing steps with the combination of resampling of minority classes to produce greater performance. The performance of the technique can vary depending on the dataset as its initial focus is on feature selection and feature combination. Multiple machine learning algorithms are utilized for the classification of tweets into positive, negative, or neutral. Results have revealed that random minority oversampling can provide improved performance and it can tackle the issue of class imbalance.

中文翻译：

---

通过对用户推文评论分类的随机少数过采样改进情感分析的框架

推特等社交网络已成为社交平台，可以为人们提供庞大的知识库，以便与朋友和家人分享他们对各种主题和问题的独特想法和观点。基于机器学习的情感分析已经成功地利用冗余可用数据发现人们的意见。然而，最近的研究指出，不平衡的数据会对结果产生负面影响。在本文中，我们提出了一个改进情感分析的框架，通过各种有序的预处理步骤结合少数类的重采样来产生更好的性能。该技术的性能可能因数据集而异，因为它最初的重点是特征选择和特征组合。多种机器学习算法用于将推文分类为正面、负面或中性。结果表明，随机少数过采样可以提供更好的性能，并且可以解决类别不平衡的问题。