News Article
Groundbreaking solution in automated high volume document classification
NIJMEGEN– 22nd of November 2011 – Edmond Document Solutions, a specialist and a leading vendor in strategic output management solutions, has added a groundbreaking solution to the Eddoc Document Solutions. This new solution for high volume classification of documents, Linguistic Classification System (LCS), was created in cooperation with the Radboud University of Nijmegen. Emeritus professor Kees Koster was one of the driving forces in the development of the software.
Professor Kees Koster explains: ‘Classification is important to manage document collections. Without the classification system you actually have nothing but a pile of documents. Automatic classification adds value for archiving and workflow management, by categorizing of and search for documents. Classifying documents manually is very time consuming, expensive and in addition to this, inconsistent with the possibility of human error. Nevertheless, there are many questions left even if one decides to use IT solutions. Multiple classification systems already exist and all of them are more or less equally effective because the underlying algorithms are quite similar.”
Koster continues: “Our Linguistic Classification System in Eddoc document management Suite offers specific content and context sensitivity. So not only words but also the sequence and context of words determine the relation between a document and the content of a document. This way any collection of documents can be classified easier, faster and more efficiently. This efficiency is what makes the big difference between LCS and other solutions in the market, especially when it comes to high volume of documents. That’s the USP of the LCS.”
Paul Bogers, CEO of Edmond, is enthusiastic and convinced of the possibilities of the LCS solution: ‘We have been able to develop the LCS software ready and applicable for business applications, like our own Eddoc Weaver output management software. Now it’s possible to automatically classify up to 95% of the documents correctly in a high volume environment. Besides that, the software is also able to separately queue the documents that are not classifiable at once. An expert can review and classify those documents manually after which the self learning ability of the system will enhance future results. With this feedback the classifier will be adjusted automatically.”
Kees Koster foresees a bright future for the LCS: ‘Classification serves a higher goal: it supports the most important business processes and eliminates manual process of classification and human error. This system is especially meant for so called 'high recall' processes. In other words, you need to find ALL the relevant documents, like in eDiscovery projects. This is where the added value of the LCS comes to life: all classifications are done in the same consistent and high speed manner, based on the exact parameters set by the user. Only then you are able to unlock random document collections the right way and for a company that's what it’s all about. Isn’t it?’
The complete interview of Kees Koster as well as further information about the LCS system is available here.