Sentiment analyzer expands Linguakit

  • 17 June, 2014
  • Linguakit

Linguakit started his career on last Wednesday with verbal conjugators but as we had already announced the idea is to expand the package of tools and their features. So today we begin to enlarge, incorporating a new linguistic instrument: a sentiment analyzer in Galician, Spanish, Portuguese and English. This is the first sentiment analyzer available in Galician!

Communication is one of the most noteworthy human characteristics and as a result of increase social networks the possibilities of communication have been expanded. But can we know what other people is saying about us, our company or anything in which we are interested? Can we measure sentiments on twitter? With sentiment analyzer we can do it automatically and even apply it to thousands and millions of tweets using CilenisAPI.

How does sentiment analyzer works?

We not only want to provide you tools, we also want to explain you how we do them. In this case when you put a phrase in sentiment analyzer the system will assign a value between – 1 and 1 so this way you can measure the degree of positivity or negativity, but if the value assigned by the system is 0 it is because you have introduced a neutral phrase. The system is, therefore, a classifier that is based on a model built on two lexical resources. On the one hand, the system uses polarized lexicon, it means, a vocabulary already classified as positive, negative or neutral and on the other hand, it uses the training corpus which consists of a system’s learning which use information compiled from other collections of statements, mainly from tweets. In addition, linguakit’s sentiment analyzer incorporates grammatical information, it means, take into account the inclusion of negative adverbs that can change the polarity of a word which are syntactically linked.

A sentiment analyzer recognized in scientific competitions

Behind the creation of the opnion analyzer is once again, Cilenis team and the USC’s ProLNat research group. This group, also linked with other research centers such as the USC’s CITIUS, attended several competitions and contests with this linguistic tool. Thus, sentiment analyzer obtained 2nd position in TweetNorm 2013 for the Spanish language, a tweet normalization workshop, but also aroses to national events like Tass 2013, a systems competition organized by Daedalus for the analysis of feelings in social networks and structured in a series of tasks. In this case, the opinion analyzer for the Spanish language got the 2nd position in task one and 1st position in task three. Also, ProLNat research group led to the scanner at the international contest SemEval obtaining an intermediate position between the proposed top 50. None of these contests were awards as such, but do not cease to be a representation of the good results of sentiment analyzer which is available from today in Linguakit.

Sentiment analyzer’s application cases

Nowadays sentiment analyzers are a very demanded instrument in the field of communication as a result of the simple but worthy information that it provides. As consequence of the growing importance of Social Media, it is essential to know the feedback of a brand or product, it means, the appreciation that users are doing through comments on social networks. This will largely determine the direction of an advertising strategy and even a political campaign but, above all, it will define the results of the campaign and the ability to have an impact on our target audience. Also, knowing the valuation of what we say and what is said about us as individual users is something that we must not ignore, because the impact that may have on our career professional can be really important. So, what do you expect?

Don’t waste more time and discover the benefits of employing Linguakit’s sentiment