Clustering Big Data
We are proud to unleash Desert Iguana (2014-03), the latest release of the webLyzard platform with several fundamental improvements - including a new search and indexing strategy based on Elasticsearch, as well as the Recognyze module to detect named entities in text documents - locations, persons, organizations, etc. The new features significantly outperform the accuracy and scalability of competing systems and will be rolled out across all our projects and portals in April 2014.
In addition, we present the Cluster Map visualization as Feature of the Month to group search results by topic and interactively explore the resulting clusters.
|
|
|
Data Services
Recognyze | Named Entity Recognition. A new component that analyzes documents from online sources and corporate archives to identify named entities such as locations, events, persons and organizations. The system then aligns these entities with external knowledge sources – e.g., items contained in linked data archives. The result is a continuously evolving repository of actionable knowledge. This repository enables decision support services and helps to better understand business networks and the relations among their actors.
[read more] |
|
Visualization
Cluster Map | Feature of the Month. Data clustering is a great way to quickly get an overview of the main topics covered in the list of search results. To position documents by similarity, the system offers a choice of either k-means or hierarchical clustering. The size of a node represents the reach of the source, its color the sentiment of the document. Interactive controls let users restrict the search to a specific cluster or explore individual documents within a cluster.
[read more] |
|
|
