NRC-Canada-2014: Detecting Aspects and Sentiment in Customer Reviews

Kiritchenko, Svetlana; Zhu, Xiaodan; Cherry, Colin; Mohammad, Saif M.

doi:10.3115/v1/s14-2076

Cited by 570 publications

(322 citation statements)

References 15 publications

Supporting

Mentioning

293

Contrasting

Unclassified

Order By: Relevance

“…Clearly, this sentence is negative, but without negation, the presence of the word 'best,' a typically positive word, might lead this tweet to be classified as positive, not negative. If however, a tag is added (in this case 'NOT ') to any words following a negation key, those words will be more likely to be classified appropriately, as 'NOT best' will more often be seen in negative contexts (Kiritchenko et al, 2014).…”

Section: Negationmentioning

confidence: 99%

“…The lexicon consists of words that humans have tagged as having either strongly negative or strongly positive sentiment. If a word in a tweet is preidentified as highly positive or negative, we add a special feature to the tweet's features to indicate that the tweet included a highly positive word or a highly negative word (Kiritchenko et al, 2014). Although multiple lexicons exist, e.g.…”

Section: Sentiment Lexiconmentioning

confidence: 99%

See 1 more Smart Citation

SWASH: A Naive Bayes Classifier for Tweet Sentiment Identification

Talbot¹,

Acheampong²,

Wicentowski

2015

Proceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015)

View full text Add to dashboard Cite

This paper describes a sentiment classification system designed for SemEval-2015, Task 10, Subtask B. The system employs a constrained, supervised text categorization approach. Firstly, since thorough preprocessing of tweet data was shown to be effective in previous SemEval sentiment classification tasks, various preprocessessing steps were introduced to enhance the quality of lexical information. Secondly, a Naive Bayes classifier is used to detect tweet sentiment. The classifier is trained only on the training data provided by the task organizers. The system makes use of external human-generated lists of positive and negative words at several steps throughout classification. The system produced an overall F-score of 59.26 on the official test set.

show abstract

Section: Negationmentioning

confidence: 99%

Section: Sentiment Lexiconmentioning

confidence: 99%

SWASH: A Naive Bayes Classifier for Tweet Sentiment Identification

Talbot¹,

Acheampong²,

Wicentowski

2015

Proceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015)

View full text Add to dashboard Cite

show abstract

“…Using sentiment lexicons in Sentiment Analysis has been a common and rewarding practice (Mohammad et al, 2013;Kiritchenko et al, 2014). The characterisation of the sentiment associated to words in tweets is important for two reasons: to detect the global sentiment (e.g.…”

Section: Sentiments and Emotional Lexiconsmentioning

confidence: 99%

UPF-taln: SemEval 2015 Tasks 10 and 11. Sentiment Analysis of Literal and Figurative Language in Twitter

Barbieri

Ronzano

Saggion

2015

Proceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015)

View full text Add to dashboard Cite

In this paper, we describe the approach used by the UPF-taln team for tasks 10 and 11 of SemEval 2015 that respectively focused on "Sentiment Analysis in Twitter" and "Sentiment Analysis of Figurative Language in Twitter". Our approach achieved satisfactory results in the figurative language analysis task, obtaining the second best result. In task 10, our approach obtained acceptable performances. We experimented with both wordbased features and domain-independent intrinsic word features. We exploited two machine learning methods: the supervised algorithm Support Vector Machines for task 10, and Random-Sub-Space with M5P as base algorithm for task 11. MotivationDuring the last decade the study and characterisation of sentiments and emotions in on-line usergenerated content has attracted more and more interest. Since 2013 several tasks dealing with Sentiment Analysis have been organised in the context of SemEval. These tasks have been mainly focused on the analysis of short texts like SMS or tweets. In this paper we describe the approach adopted by UPF-taln team for tasks 10 and 11 of SemEval 2015, both dealing with the analysis of English tweets. Task 10 concerned "Sentiment Analysis in Twitter" * The research described in this paper is partially funded by the Spanish fellowship RYC-2009-04291, the SKATER-TALN UPF project (TIN2012-38584-C06-03), and the EU project Dr. Inventor (n. 611383).and included different subtasks. We participated in the subtask B, named "Sentiment Polarity Classification". Given a message, we were asked to classify whether the message was of positive, negative, or neutral sentiment. In Task 11 the participants were asked to determine the polarity score (between -5 to +5) of tweets rich in metaphor and irony. Our model reaches satisfactory results in the figurative language task 11, however it has suboptimal performance in task 10.We exploited an extended version of the tweet classification features and approach described in . In particular, we experimented the use of intrinsic word features, characterising each word in a tweet to try to model and thus automatically determine its polarity. Thanks to intrinsic word features, we aimed to detect two aspects of tweets: the style used (e.g. register used, frequent or rare words, positive or negative words, etc.) and the unexpectedness in the use of words, particularly important for figurative language. We also exploited textual features (like word occurrences, bigrams, skipgrams or other word patterns) in order to capture the way words are used in positive and negative tweets. As machine learning approach we choose the supervised method Support Vector Machines (Platt, 1999) for task 10 and the regression algorithm Random-Sub-Space (Ho, 1998) with M5P (Quinlan, 2014) as base algorithm for task 11.In Section 2 and 3 we describe the dataset used and the tools we employed to process the tweets. In Section 4 we introduce the features we built our model on. In Section 5 we discuss the performance of our model in SemEval 2015 and in Section 6 ...

show abstract

“…Following (Kiritchenko et al, 2014), we manually filtered out categories not corresponding to food related businesses (173 out of 720 were finally selected). A total of 997,721 reviews (117.1M tokens) comprise what we henceforth call the Yelp food corpus (C Y elp ).…”

Section: Corporamentioning

confidence: 99%

EliXa: A Modular and Flexible ABSA Platform

Vicente¹,

Saralegi²,

Agerri

2015

Proceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015)

View full text Add to dashboard Cite

This paper presents a supervised Aspect Based Sentiment Analysis (ABSA) system. Our aim is to develop a modular platform which allows to easily conduct experiments by replacing the modules or adding new features. We obtain the best result in the Opinion Target Extraction (OTE) task (slot 2) using an off-the-shelf sequence labeler. The target polarity classification (slot 3) is addressed by means of a multiclass SVM algorithm which includes lexical based features such as the polarity values obtained from domain and open polarity lexicons. The system obtains accuracies of 0.70 and 0.73 for the restaurant and laptop domain respectively, and performs second best in the out-of-domain hotel, achieving an accuracy of 0.80.

show abstract

NRC-Canada-2014: Detecting Aspects and Sentiment in Customer Reviews

Cited by 570 publications

References 15 publications

SWASH: A Naive Bayes Classifier for Tweet Sentiment Identification

SWASH: A Naive Bayes Classifier for Tweet Sentiment Identification

UPF-taln: SemEval 2015 Tasks 10 and 11. Sentiment Analysis of Literal and Figurative Language in Twitter

EliXa: A Modular and Flexible ABSA Platform

Contact Info

Product

Resources

About