Automatic extraction of reaction information from databases using classification and learning techniques

Work under development at RISC-Linz is outlined. By means of a set of tools for manipulating synthetic information from a reaction database, learn and classification algorithms are used to extract and organise synthetic reaction class descriptions (i.e. reaction patterns). This information is to be used either in an automatic synthetic planning system or for database organisation. The general principle and philosophy of the methods used are explained and two example procedures are suggested which use these principles.