<> "The repository administrator has not yet configured an RDF license."^^ . <> . . "Theoretical analyses of cross-validation error and voting in instance-based learning"^^ . "This paper begins with a general theory of error in cross-validation testing of algorithms\nfor supervised learning from examples. It is assumed that the examples are described by\nattribute-value pairs, where the values are symbolic. Cross-validation requires a set of\ntraining examples and a set of testing examples. The value of the attribute that is to be\npredicted is known to the learner in the training set, but unknown in the testing set. The\ntheory demonstrates that cross-validation error has two components: error on the training\nset (inaccuracy) and sensitivity to noise (instability).\nThis general theory is then applied to voting in instance-based learning. Given an\nexample in the testing set, a typical instance-based learning algorithm predicts the designated\nattribute by voting among the k nearest neighbors (the k most similar examples) to\nthe testing example in the training set. Voting is intended to increase the stability (resistance\nto noise) of instance-based learning, but a theoretical analysis shows that there are\ncircumstances in which voting can be destabilizing. The theory suggests ways to minimize\ncross-validation error, by insuring that voting is stable and does not adversely affect\naccuracy."^^ . "1994" . . "6" . . "Journal of Experimental and Theoretical Artificial Intelligence"^^ . . . . . . . . "Peter D."^^ . "Turney"^^ . "Peter D. Turney"^^ . . . . . . "Theoretical analyses of cross-validation error and voting in instance-based learning (PDF)"^^ . . . . . . . . . "NRC-35073.pdf"^^ . . . "Theoretical analyses of cross-validation error and voting in instance-based learning (Indexer Terms)"^^ . . . . . . "indexcodes.txt"^^ . . "HTML Summary of #1821 \n\nTheoretical analyses of cross-validation error and voting in instance-based learning\n\n" . "text/html" . . . "Artificial Intelligence" . . . "Machine Learning" . . . "Statistical Models" . .