A Unified Framework for Multi-Language Sentiment Analysis

Abstract

The unified framework for multi-language sentiment analysis is a vital aspect of understanding customer opinions, emotions, and feedback. This paper presents a unified framework to increase the performance of the multi-language sentimental analysis. Two popular machine translation services, Google Translate, and Yandex Translate are employed to unify the sentiment analysis for the considered languages including English, Turkish, Arabic, and French. Our findings highlight the importance of machine translation services in facilitating and enhancing the performance of sentiment analysis algorithms for different languages. Our framework was evaluated on several datasets and showed promising results, with improvements in accuracy ranging from 1% to 22% depending on the language. Our approach outperforms language-specific models and demonstrates the effectiveness of the proposed translation-based multi-language framework. In addition, we found that the performance of sentiment analysis varies among the different languages, with Google Translate exhibiting better performance in sentiment analysis of Turkish and Arabic translations while Yandex Translate shows better results in sentiment analysis of English and French translations.