| Summary: | 碩士 === 國立雲林科技大學 === 資訊工程系 === 107 === Natural Language Processing enables computers to understand human natural languages and assists humans to perform many tasks, such as information retrieval, question answering, automatic summarization, text categorization, machine translation, etc. To our knowledge, there is still no NLP employed on the application of translating negative sentimental statements into positive sentimental statements with similar semantics although it is an important application in communications between humans. In this paper, we propose the concepts and implementations of translating negative sentimental statements into positive sentimental statements. First, we build negative-positive sentimental statement data sets for a sentiment translation model. Then, the sentiment translation model is trained using deep learning techniques. In the experiments, we use perplexity, BLEU, and subjective human assessments to evaluate the model, and the experimental results are satisfactory. Finally, we think the methods employed in translating sentimental statements are feasible if the trained data sets can be built as expected, and more advanced models are developed.
|
|---|
