Companies market their services and products on social media platforms with today's easy access to the internet. As result, they receive feedback and reviews from their users directly on their social media sites. Reading every text is time-consuming and resourcedemanding. With access to technology-based solutions, analyzing the sentiment of all these texts gives companies an overview of how positive or negative users are on specific subjects will minimize losses. In this paper, we propose a deep learning approach to perform sentiment analysis on reviews using a convolutional neural network model, because that they have proven remarkable results for text classification. We validate our convolutional neural network model using large-scale data sets: IMDB movie reviews and Reuters data sets with a final accuracy score of ~86% for both data sets.

中文翻译：

---

一种基于语料库的新卷积神经网络，用于大数据文本分析

随着当今互联网的便捷访问，公司可以在社交媒体平台上销售其服务和产品。结果，他们直接在其社交媒体网站上从用户那里收到反馈和评论。阅读每个文本既耗时又需要资源。通过使用基于技术的解决方案，分析所有这些文本的情绪可以使公司大致了解特定主题上正面或负面用户的使用情况，从而将损失降到最低。在本文中，我们提出了一种深度学习方法，以使用卷积神经网络模型对评论进行情感分析，因为它们已被证明对文本分类具有显着效果。我们使用大规模数据集验证了卷积神经网络模型：IMDB电影评论和路透社数据集的最终准确性得分均为这两个数据集的〜86％。