Document classification using connectionist models

As part of research into document analysis, we implemented a method for classification of binary document images into textual or non-textual data blocks using connectionist models. The four connectionist models considered were backpropagation, radial basis function, probabilistic connectionist, and Kohonen's self-organizing feature map. The performance and behavior of these connectionist models are analyzed and compared in terms of training times, memory requirements, and classification accuracy. The experiments carried out on a variety of medical journals show the feasibility of using the connectionist approach for textual block classification and indicate that in terms of both accuracy and training time the radial basis function connectionist should be preferred.<<ETX>>