The complex nature of big data resources requires new structuring methods, especially for textual content. WordNet is a good knowledge source for the comprehensive abstraction of natural language as it offers good implementation for many languages. …
We propose a novel method for counting sentiment orientation that outperforms supervised learning approaches in time and memory complexity and is not statistically significantly different from them in accuracy. Our method consists of a novel approach …
A method for sentiment polarity assignment for textual content written in Polish using supervised machine learning approach with transfer learning scheme is proposed in the paper. It has been shown that performing simple natural language processing …
The paper examines two approaches to sentiment analysis: lexicon-based vs. supervised learning in the domain of movie reviews. In evaluation, the methods were compared using a standard movie review test collection. The results show that lexicon-based …
The main goal of the paper is to present that word's sentiment can be discovered from propagation through well-defined word networks such as Word-Net. Therefore a new method for propagation of sentiment from a given word seed - Micro-WNOp corpus over …