Shaking up content categorization
Claiming to deliver the most accurate content categorization, Quiver has introduced QKS Classifier, the first component of the Quiver Knowledge Suite. The company reports its hybrid taxonomy platform combines classification algorithms with workflow and directory management tools to capture human input and expertise.
The QKS thin-client Directory Management Toolset provides an intuitive user interface into distributed workflow management, Quiver says, giving each “information manager” the ability to create or modify a custom taxonomy, choose appropriate training set examples to increase accuracy of content, view and control what documents are being retrieved and implement business rules to automate appropriate publishing steps.
Further, the Workflow Memory function is said to eliminate redundant effort with the ability to recall what actions the information manager last took relative to each document and compare it to new dynamic content changes. If, say, a small change occurs, then no additional human input is required, and the previous decisions are recalled and re-executed, Quiver explains.
The Quiver categorization engine employs two algorithms, the first of which identifies which topic or topics a document appears to fall under. The second ranks the relevance of the content for a topic.
QKS Classifier is built in Java, J2E compliant and platform independent and features APIs based on industry standards such as XML and SOAP. The company adds that the Output and Display Interface produces XML output and flexible display alternatives.
Pricing starts at $125,000 for the full client server platform and varies depending on the size of the taxonomy and the number of information manager seat licenses required.
In June, Quiver announced a agreement with Inktomi to integrate QKS Classifier with the Inktomi search products to jointly develop and market enterprise search and categorization solutions.