An easier way to describe sentiment analysis is by obtaining the text´s polarity, in other words, by establishing whether it has positive, neutral or negative connotations.
Many of the words we usually employ denote a sentiment connotation to the text. If we analyse these words and the relationship between them, it is certainly possible to obtain the user´s opinion regarding a topic.
For example, let´s assume you want to know the opinion held by Twitter users about the city of Salamanca (Spain). By doing a sentiment analysis over the tweets published regarding this topic, we can answer this question. Messages published by the users show a polarity, and as we have seen, it can be positive, neutral or negative. We can even know the exact reasons of these opinions by undertaking a deeper text analysis.
The BISITE Research Group has completed a project in this area in collaboration with the Advanced Services Reginal Centre (CSA) and ANPRO 21. The name of the project is “Network Threat Detection Platform (PIAR)” and its main objective was to develop a platform that enables the information retrieval and analysis coming from the Internet following a Big Data Schema.
The BISITE Research group is currently working on a research line regarding sentiment analysis. It is focused on developing a new platform to undertake multilingual sentiment analysis. These tools, called Esperanto, make it possible to apply a textual analysis of information in languages such as Spanish, English, French, German and Russian.
The motivation of this project relies on the increasing need of both companies and organizations to have the right tools to monitor the information hosted on the Internet, flowing through social networks and news sources. The aim of this project is to analyse the impact produced by this information in real life. Furthermore, this platform enables the identification of the most relevant users within a topic, community detection and storage, and Twitter profile searches.
The application of these analysis tools covers a wide range of possibilities: from security analysis to political or social issues and marketing. In the first case, it is possible to search for texts on terrorism and find users that either support or reject it. By searching terrorism supporters, security agencies can undertake a detailed investigation on these users. In the field of marketing, this technique makes it possible to obtain the opinion of a group of users with regard to a brand or advertisement campaign. Thus, brands can improve their reputation by listening to unhappy user requests.
