Technical Note : Selecting a Classification Method by Cross-Validation

摘要: If we lack relevant problem-specific knowledge, cross-validation methods may be used to select a classification method empirically. We examine this idea here show in what senses does and not solve the selection problem. As illustrated empirically, lead higher average performance than application of any single strategy, it also cuts risk poor performance. On other hand, is no more or less form bias simpler strategies, applying appropriately ultimately depends same way on prior knowledge. In fact, seen as partial information about applicability alternative strategies.