Aspect extrAction Using conditionAl

This paper describes the aspect extraction system that was presented at SentiRuEval-2015: aspect-based sentiment analysis of users’ reviews in Russian. The proposed system uses a conditional random field algorithm for extracting aspects mentioned in the text. We used a set of morphological and syntactic features for machine learning and demonstrated that using lemmas as a feature can improve aspect extraction results. The system was used to perform two subtasks, Task A—automatic extraction of explicit aspects and Task B—automatic extraction of all aspects (explicit, implicit and sentiment facts), and tested on two domains—restaurants and cars. Both subtasks, A and B, in both domains have been completed with quite a high level of precision which meant that the system was capable of rather accurate recognition of aspect terms. But lower recall results implied that the system found enough aspect terms that could not be treated as aspects according to the gold standard. Our systems performed competitively and showed the results comparable to those of the other 10 participants.