Vector-based Sentiment Analysis of Movie Reviews
نویسندگان
چکیده
Sentiment analysis is an important step towards comprehension in natural language processing. Movie reviews are a convenient source of highly polarized sentences for use in sentiment analysis. Achieving a high level of accuracy in the sign of nonneutral sentiment is a challenge. When the problem is expanded to choosing one of 5 sentiment levels the problem becomes significantly harder. Among the models that we tested, softmax regression with a bag of phrases feature set provided the best 5-bin error rate, while SVM with the same bag of phrases features and a simple word sentiment sum model provided the best error rate on sign prediction.
منابع مشابه
Sentiment Analisis on Web-based Reviews using Data Mining and Support Vector Machine
This work aims to use sentiment analysis techniques, data mining, text mining and natural language processing to indicate the polarity of texts using support vector machine. Weka software and a movie review database from Internet Movie Database IMDb were used. This work uses preprocessing filters and WRAPPER techniques and Support Vector Machine (SVM) for classification. It presents better resu...
متن کاملMining and Summarizing Movie Reviews in Mobile Environment
In this paper, we design and develop various strategies required for sentiment analysis of movie domain in mobile environment. The main objective of review mining and summarization is extracting the features on which the reviewers express their opinions and determining whether the opinions are positive or negative. The sentiment classification is done by various classifiers such as maximum entr...
متن کاملOpinion Analysis on Web-based Reviews Using Support Vector Machine
This work aims to use sentiment analysis techniques, data mining, text mining and natural language processing to indicate the polarity of texts using SVM (support vector machine). Weka software and a movie review database from IMDb (internet movie database) were used. This work uses preprocessing filters and WRAPPER techniques and SVM for classification. It presents better results when compared...
متن کاملA Study on the Combined Approach of Sentiment Classification Based on Ontology
The text documents contain opinions or sentiments on some objects, such as movie reviews, book reviews, product reviews etc. Sentiment analysis is mining the sentiment or opinion words and identification or analysis of the opinion and arguments in text. Here this paper proposed an ontology based combination approach to improve the exits approaches of sentiment classifications and to use supervi...
متن کاملMovie reviews: do words add up to a sentiment?
Sentiment analysis, the automatic extraction of opinion from text, has been enjoying some attention in the media during the national elections. In this thesis, we will discuss the classification of movie reviews as ’thumbs up’ or ’thumbs down’. Movie reviews are interesting and difficult because of the wide range of topics in movies. The reviews are HTML web pages, which poses an interesting ch...
متن کامل