Startseite

Newsletter-Management-System (NMS)

Intelligent information coverage with a “learning” text classification software

In-time information coverage has become very important in our society. Employees of many enterprises must be provided with current news and enterprise specific reports. Therefore, it is necessary to edit and to sort an extensive amount of data. Since searching for interesting and important news is effortful and documents must be checked on relevancy and sorted into categories, such information is usually brought to employees via newsletters or mailing lists.

The IntraFind newsletter management system makes it possible to collect current information or articles from different information sources (both charged and charge free) and check them automatically concerning relevancy. They are assigned to predefined categories and can easily be sent as a newsletter either fully automatic or with an additional step of manual quality assurance.

The newsletter management system is based on Intrafind’s automatic text classification software TopicFinder which "learns" how to allocate documents into individual classification categories via training-documents. This means, the longer this procedure is used, the more accurate the sorting becomes. The technology is based on linguistic and statistical pattern recognition methods (e.g support vector machines). In the newsletter management system this processing step is tied with a complete java-based web application, which newsletter editors can use to construct these newsletters very efficiently. The system automatically creates a suggestion which can still be edited and completed. The final steps are the creation of a PDF file with the corresponding layout and sending it out to the receiver circle.

Editing documents, creating a selection of newsletters, as well as sorting the articles, sighting and assigning them to predefined categories is done nearly fully automatically. Normally these processes require the use of highly qualified staff, which is a waste of resources. Performance and exclusive features

The product properties of the newsletter management system are outstanding:

  • Standard software, completely web based
  • Usage of the best available learning procedure for text classification (support vector Machine)
  • Flexible import of the relevant data (e.g. from Notes, databases, XML, FTP) 
  • Flexibly definable output format (e.g. PDF, Word or XML)
  • Automatic dispatch to defined receivers
  • Highly adaptable, without having to change the source code 
  • Little system prerequisites: Just servers with a servlet container like Tomcat and a current Java software are needed

Download: Product information NMS



Download: Case Study Merck