The paper considers approaches to sentiment analysis towards a specific entity and its characteristics (aspects). To solve the aspect-oriented sentiment analysis task, it is necessary to extract aspect terms from texts, to classify or cluster aspect terms into aspect categories, to determine the sentiment expressed towards the specfic aspect. The paper also briefly presents SentiRuEval-2015 evaluation of aspect-oriented sentiment analysis systems in Russian.

The article investigates the problem of aspect-based sentiment analysis. Such version of analysis is more challenging compared to general task of sentiment detection problem. It implies the solutions to the number of related subtasks such as aspect term extraction, aspect term polarity detection and aspect category polarity detection. The solution of aspect-based sentiment analysis problem significantly extends the capabilities of natural language processing systems.
The article gives the overview of previous works in the field and describes the train and test data from the Russian evaluation workshop SentiRuEval. For the task of aspect term extraction the vector space of distributed representations of words was used. Aspect term detection is based on mutual information method and semantic similarity. The paper contains the number of experimental results. At the end the final conclusions are drawn.

Sentiment analysis and opinion mining technologies are growing fast. This is mostly due to a rapid grow of the data sources consisting a vast amount of user opinions and reviews on a wide set of topics. In this paper we describe methods for sentiment analysis of reviews and short messages (tweets), as well as evaluation of results obtained during SentiRuEval-2015.

The paper describes our approach to the task of sentiment analysis of tweets within SentiRuEval – an open evaluation of sentiment analysis systems for the Russian language. We took part in the task of sentiment analysis of Russian tweets concerning two types of organizations: banks and telecommunications companies. On both datasets, the participants were required to perform a three-way classification of tweets: positive, negative or neutral.

We used various statistical methods as basis for our machine learning algorithms. Linguistic features produced by our morpho-syntactic analyzer are applied to the classification. Syntactic relations proved to be a crucial feature for any statistical method evaluated, and SVM-based classification performed better than the others. Normalized words are another important feature for the algorithm.

The evaluation revealed that our method proved to be rather successful: we scored the first in three out of four evaluation measures.

This paper focuses on the use of a linguistics-based method for automatic object-oriented sentiment analyses. The study was conducted as part of SentiRuEval automatic sentiment analysis system testing cycle. The original task was to extract users’ opinions (positive, negative, neutral) about telecom companies, expressed in tweets and news. In this study news was excluded from the dataset because, being formal texts, news significantly differs from informal ones in its structure and vocabulary and therefore demands a different approach. Only linguistic approach based on syntactic and semantic analysis was used. In this approach, a sentiment-bearing word or expression is linked to its target object at either of two stages, which perform successively. The first stage includes usage of semantic templates matching the dependence tree, and the second stage involves heuristics for linking sentiment expressions and their target objects when syntactic relations between them do not exist. No machine learning was used. The method showed a very high quality, which roughly coincides with the best results of machine learning methods and hybrid approaches.

This paper describes the Information 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 to extract aspect terms mentioned in the text. A set of morphological features was used for machine learning. The system intent 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 automobiles. Our systems performed competitively and showed the results comparable to those of the other 10 participants.