Sentiment Analysis

A technology to automatically determine the tone of texts on social media. Detection of emotionally colored words in posts, comments, and responses.

This technology is used to reveal the author’s emotional assessment— negative, neutral, or positive. Text analysis determines +30 basic emotions, such as sadness, surprise, anger, etc.

About the algorithm

  • Machine learning methods (classification) are used to solve this task. As a rule, this involves three types of algorithms known as state-of-the-art solutions for different modifications of the task:
  • NB-SVM (state of the art 2012).
  • Convolutional Neural Networks (state of the art 2015).
  • LSTM/GRU (state of the art 2015).

In products

Other technologies

  • User Interests Mining

    Mining social media user interests based on their behavior. More than 250 different interests

  • Customer Profiling

    Determination of the social media user’s social and demographic profile

  • Social Network Search

    Search company clients on social media by telephone and/or e-mail for research and advertising purposes

  • Topic Categorization

    Text categorization by +200 topics such as sports, real estate, finance, etc.

  • Profile Matching

    Identification of a single user’s profiles on different social media and integration of their information

  • User Data Matching

    Used to match social media user profiles with user data in free form 

  • Look-alike

    Search social media users that look like business clients or a certain audience