Figure 12. Neural network learning fantasy of SAR/SPR modeling (adapted from ref.

Figure 12

Neural network learning fantasy of SAR/SPR modeling (adapted from ref. 15). Formal neurons are drawn as circles, and lines between neurons represent connection weights. Input neurons receive molecular patterns and represent molecular descriptors or attributes. The hidden layer neurons and their associated connection weights process the network stimuli, in a flexible system their connectivity adapts during the learning process, thereby extracting and storing the SAR/SPR-relevant features. The values calculated by the output neurons might be molecular activity, properties, etc.

From: A Conceptual Framework

Cover of Madame Curie Bioscience Database
Madame Curie Bioscience Database [Internet].
Austin (TX): Landes Bioscience; 2000-2013.
Copyright © 2000-2013, Landes Bioscience.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.