Abstract: A computer implemented method for training a neural network to capture a structural feature specific to a set of chemical compounds is disclosed. In the method, the computer system reads an expression describing a structure of the chemical compound for each chemical compound in the set and enumerates one or more combinations of a position and a type of a structural element appearing in the expression for each chemical compound in the set. The computer system also generates training data based on the one or more enumerated combinations for each chemical compound in the set. The training data includes one or more values with a length, each of which indicates whether or not a corresponding type of the structural element appears at a corresponding position for each combination. Furthermore, the computer system trains the neural network based on the training data for the set of the chemical compounds.

Abstract: Phenoxypropylamine derivatives of the general formula ##STR1## wherein R.sub.1 represents a hydrogen atom or a methyl or ethyl group, R.sub.2 and R.sub.3, independently from each other, represent a lower alkyl group, or together form a linear alkylene group having 4 to 7 carbon atoms which may be optionally substituted by a hydroxyl or hydroxymethyl group, and R.sub.4 represents a hydrogen atom or a group of the formula --R.sub.5 --Z in which R.sub.5 represents a lower alkylene group, and Z represents a hydrogen atom or an amino, mono- or di-(lower alkyl)-amino, hydrdoxy lower alkylamino, lower alkanoylamino, hydroxyl, lower alkoxy, lower alkanoyloxy, phenoxy, halophenoxy, benzoyloxy or halobenzoyloxy group, andthe salts thereof; to a process for production thereof; and to their medicinal use, particularly to antiulcer agents comprising these phenoxypropylamine derivatives or their salts.