Personal information filtering system - An application of associative memory

An adaptive personal information filtering system with associative memory is put to practical use. Associative memory is one of fundamental neural network models. We try to apply associative memory to the adaptive engine of our information filtering system in order to organize personal profiles which reflect user's interest. The trial suggests that associative memory is useful as the adaptive engine but not powerful to put to practical use in information filtering. Therefore the adaptive learning algorithm of the current system is modified through the statistical consideration - what keywords are useful for retrieving items in which the user wants to see. The current system gives the user items according to his or her interest estimated using the user's personal profile through user-system interaction. Our field tests show that the system can he put to practical use.