Performance

Our platform identifies, processes and analyses structured and unstructured data. The system constantly crawls millions of open-source documents and social media messages and extracts causal relationships between words/entities/concepts to visualise new knowledge. In order to ensure a high degree of performance the platform operates with high speed and precision. 

Speed

The platform is capable of crawling millions of multilingual sources with high precision in real time to detect the most significant events and developments. It has the capacity to process and evaluate high volumes of documents in short time and analyse underlying unstructured and structured data in high speed,  using cutting edge parsing technology with highest throughput.

 

 

Accuracy & Precision  

The semantic search function of our platform uses the result of linguistic processing (syntactic and semantic annotations) together with the immediate analysis of the user's request in order to ask more targeted questions. We have integrated such functions as auto-completion and partial correction of questions as well as the option to view correlated searches. Search queries can be applied to any document management system or as a search function within a web portal. The semantic advanced search is performed in two main phases: indexing and searching.

Indexing: While reading a text that is to be memorised, the platform automatically generates an index complete with metadata that subsequently allows for information retrieval more powerful than regular keyword searches. The content index is configured as a conceptual map that contains links between the analysed concepts, which allows for high data precision. 

Advanced Search: After entering or asking a specific question, the system interprets the text in real time, identifies the search information (e.g. names, places, events) and launches a query on the previously created concept map. As result, the system is capable of providing precise answers and not only a link to the most relevant documents.

 

Natural Language Processing: The core of the text analytics process is building Natural Language Processing (NLP) algorithms that use linguistic, statistical and machine learning knowledge to transform unstructured data into relevant information. Our semantic technology is the driver of our intelligence platform; complex queries generate accurate results via the Advanced Analytic’s Deep NLP Engine, which interprets grammatical, syntactical and semantical text structures to:

  • automatically identify entities, people, places and things as well as their roles and concepts in order to conduct a comprehensive sentiment and risk analysis, thus driving early warning systems and highly accurate reporting applications;
  • extract the most relevant information quickly and precise from a continuous stream of data (e.g. news, SMS, e-mail, RSS feeds) and highlight semantic similarities. It automatically creates connections between people, places, dates and other categories, thus replacing the manual tags selection;
  • facilitate a rapid development of new language modules through “training by reading” of numerous texts. The system automatically creates a complex parallel list of semantic/conceptual associations in different languages without any additional manual translation effort, and detects even fine-grained aspects of the language;
  • generate full-sentence analyses and comprehend the meaning of entire sentences and discourses rather than focusing on text portions alone;    
  • detect implicit information, e.g. entities referred to by expressions such as pronouns. 

Application Programming Interface: By applying an Application Programming Interface (API), the system provides the option of exchanging information and/or interacting with third-party software. It allows clients not only to use the graphical interface of our semantic application but also to export data to enrich their projects and applications with Web intelligence to enrich final results.