Content visualization of scientific corpora using an extensible relational database implementation
A method for supervised classification and visualization of collections of scientific publications is presented. By integrating a text classification module, which leads to class probability estimation, along with a dimensionality reduction technique, which represents each class in the 2-D space, any collection of unlabelled documents can be visualized. The classification and visualization modules have been trained on three different datasets and respective categorizations. We provide an example of our system’s functionality by visualizing the content of collections of publications which share a common funding scheme. In order to implement this, we have developed a funding mining submodule which identi- fies documents of particular funding schemes. All the individual modules have been implemented using the madIS system, which provides data analysis functionalities via an extended relational database.