Abstract
In this paper, advice-giving software which uses various strategy-evaluation criteria is described. Ultimately the software will make use of an artificial neural network to connect the scores of strategies on the evaluation criteria with their overall desirability scores, and such ‘learning’ will, in theory, enable the software to give better and better advice the more it is used. To test this, a preliminary experiment was conducted to see whether simulated neural networks can actually mimic relationships between strategy-evaluation criterion scores and overall scores. The results suggest that artificial neural networks might be more accurate than conventional statistical methods at predicting overall strategy scores.
Get full access to this article
View all access options for this article.