Know-Center at SemEval-2016 Task 5: Using Word Vectors with Typed Dependencies for Opinion Target Expression Extraction

Proceedings of SemEval-2016, pages 266–270, San Diego, California, June 16-17, 2016. c⃝2016 Association for Computational Linguistics Know-Center at SemEval-2016 Task 5: Using Word Vectors with Typed Dependencies for Opinion Target Expression Extraction Stefan Falk Know-Center GmbH Inffeldgasse 13 Graz, 8010, Austria sfalk@know-center.at Andi Rexha Know-Center GmbH Inffeldgasse 13 Graz, 8010, Austria arexha@know-center.at Roman Kern Know-Center GmbH Inffeldgasse 13 Graz, 8010, Austria rkern@know-center.at Abstract This paper describes our participation in SemEval-2016 Task 5 for Subtask 1, Slot 2. The challenge demands to find domain spe- cific target expressions on sentence level that refer to reviewed entities. The detection of tar- get words is achieved by using word vectors and their grammatical dependency relation- ships to classify each word in a sentence into target or non-target. A heuristic based func- tion then expands the classified target words to the whole target phrase. Our system achieved an F1 score of 56.816% for this task. 1 Introduction Nowadays, modern technologies allow us to col- lect customer reviews and opinions in a way that changed the sheer amount of information avail- able to us. For that matter the requirement to ex- tract useful knowledge from this data rose up to a point where machine learning algorithms can help to accomplish this much faster and easier than hu- manly possible. Natural language processing (NLP) emerges as an interfacing tool between human natu- ral language and many technical fields such as ma- chine learning and information extraction. This article describes our approach towards Opin- ion Target Expression (OTE) extraction as defined by Task 5 for Subtask 1, Slot 2 of the SemEval- 2016 (Pontiki et al., 2016) challenge. The core goal behind Slot 2 in Subtask 1 of Task 5 is to extract consecutive words which, by means of a natural lan- guage, represent the opinion target expression. The opinion target expression is that part of a sentence which stands for the entity towards which an opin- ion is being expressed. An example could be the word “waitress” in the sentence “The waitress was very nice and courteous the entire evening.”. The evaluation for Slot 2 fell into evaluation phase A, where provided systems were tested in order to return a list of target expressions for each given sen- tence in a review text. Each target expression was an annotation composed of the index of the starting and end character of the particular expression as well as its corresponding character string. For our system we decided to used word vectors (Mikolov et al., 2013a; Mikolov et al., 2013b). Word vectors (Bengio et al., 2003) are distributed repre- sentations which are designed to carry contextual information of words if their training meets certain criteria. We also used typed grammatical dependen- cies to extract structural information from sentences. Furthermore we used a sentiment parser to deter- mine the polarity of words. 2 External Resources Our system uses Stanford dependencies (Chen and Manning, 2014) and utilizes the Stanford Sentiment Treebank (Socher et al., 2013) for sentiment word detection. 3 System …

Content truncated to 500 words