Rosoka Software develops and licenses software for the extraction and analysis of unstructured data, regardless of the source language or document type. Rosoka Software products were developed for PDA’s, Laptops, secured networks and Cloud Environments to be; 1) Multilingual with the ability to process over 230 languages, 2) Scalable to handle Big Data volume, variety, and velocity, and 3) Portable with the ability to process consistent results from most standard operating systems. The technology is based on natural language processing (NLP), linguistic rules and algorithms to provide named entity extraction, relationship analysis, enhanced geospatial mapping and text analytics. Product licenses are inclusive of all languages and API’s, easily customizable by customers, with improved accuracy obtained through supervised machine learning.
Rosoka software is primarily used; 1) for automated identification, discovery, monitoring and enforcement, 2) to assess sentiment and behavior or 3) to predict or persuade desired outcomes or results.
Rosoka multilingual entity extraction and analysis bundle includes 1 manager and 1 worker node for processing up to 1GB of unstructured data per hr. One Rosoka Studio included.
Rosoka Studio allows for modifying entity types, relationships, definitions, lexicons and rules for customized data management. Also provides regression testing to control quality.
Rosoka's worker nodes are the engines that enable data extraction and allow for scalability. Each worker node will process 0.5-1GB of unstructured data per hour.
Rosoka Manager resides within a cluster of Worker and Studio nodes to provide automated load balancing and centralized licensing.
Rosoka SDK offers multiple API's for integration of multilingual entity and relationship extraction and analysis. SDK instances each process 0.5-1GB of unstructured data per hr.
Rosoka Text Analytics plugin for i2 Analyst's Notebook allows laptop users the seamless ability to extract entities and relationships from text or unstructured data.
Enhanced geospatial tagging used with one of five gazetteers to return latitude and longitude coordinates or provides most likely location from extracted unstructured data.