Sentiment Analysis in Czech Social Media Using Supervised Machine Learning

摘要: This article provides an in-depth research of machine learning methods for sentiment analysis Czech social media. Whereas in English, Chinese, or Spanish this field has a long history and evaluation datasets various domains are widely available, case language there not yet been any systematical conducted. We tackle issue establish common ground further by providing large humanannotated media corpus. Furthermore, we evaluate state-of-the-art supervised analysis. explore different pre-processing techniques employ features classifiers. Moreover, addition to our newly created dataset, also report results on other popular domains, such as movie product reviews. believe that will only extend the current another family languages, but encourage competition which potentially leads production high-end commercial solutions.