Our Content Analytics technology based on Text and Media Data Analytics, which are able to extract insights from the unstructured data. What is unstructured data sources? There are:
- Any kind of text – web content, word, excel, pdf documents, emails, sms. So this is the any text created by human or machine;
- Media data – mostly but without limiting, it is audio sources (phone calls, audio sensors data) and video sources (YouTube casts, television and cctv records).
What we can do with this data? Features:
- Structuring unstructured and semi-structured data
It is the first step in text mining. As a result we have the ability to analyze data that previously was only storable by our or any another solutions. - Knowledge extraction and answering to business questions
We can make your data “query supportable”. For example we can develop for you the reports and real-time dashboards witch will answer to predefined questions: “What does this text about?”, “Is this text about my goods/services/assets and which one of them exactly?”, “If this text about my rent service that author think about it?”, “Does this text have trade secret information?” and many other depends on your business needs. - Identify hidden insights
You can also ask us to explore your data. Our researchers will make a complete report that will help you more accurate understand your data specific and show you hidden possibilities and the way to push up your business using already existing resources. - Present results effective way
Text mining and its result is not so easy to understand. This is why we engage top-professional ux/ui disinters to generate well-understandable and high-effective visual user interface to analyze reports and interact with dashboards.
How can we do this? There are three stages:
- First – Entity and Conceptual Extraction. We use the Natural Language Processing (NLP) Annotation Query Language (AQL) technologies to for building extractors that extract structured information from unstructured or semi structured data and mark significant words into it. We can processing texts in English, German, Russian, Kazakh and Hebrew languages. Sometime it is useful to normalize text – convert words into their first form.
- Second – Relational Extraction. Link analysis. Since we have the text structure and its building blocks (significant words) we are able to understand text semantic (analyze significant words in their context).
- Third – Sentiment analysis. One of the most useful feature – is to understand how author refers to the text ideas. We define following structure Object, Subject and Sentiment. In other words, it means: Who are author, What about he talks and How he refers to it.