Wp. Walters et Dp. Dolata, MOUSE-III - LEARNING RULES OF CONFORMATIONAL-ANALYSIS FROM X-RAY DATA, Journal of molecular graphics, 12(2), 1994, pp. 130-138
MOUSE-III is learning program that finds rules of conformational analy
sis from raw crystallographic data. The program perceives molecular fe
atures, finds conformational classes in the data and then learns rules
that link features to classes. The rules that MOUSE learns are capabl
e of correctly assigning conformations to ring systems that were not u
sed for training with greater than 95% accuracy, when MOUSE was presen
ted with sufficient data. The rules also show a compression of as much
as 99% when compared to the raw data. This is accomplished through ab
straction and generalization. The algorithm is presented along with a
carefully worked example. An example of a learned rule is also present
ed and analyzed. Some conclusions about the scope and limitations of t
he learning process are presented.