Bioinformatics 2002 Bergen , Norway April 4 −
نویسندگان
چکیده
The paper contains the comparison between several class prediction methods (the K-Nearest Neighbour (KNN) algorithms and some variations of it) for classification of tumours using gene expression data. The KNN is a traditional classifier that uses a set of attributes for class prediction. Also are considered, the cases when these attributes (for KNN algorithm) are un-weighted (i.e. they all have the same relevance for classification) and weighted (they have different relevance). Also has been tested a Hybridised Genetic Algorithm with KNN classifier which has been demonstrated to be a good classifier in terms of accuracy in others situations. The experimental results demonstrated that the Simple W-KNN algorithm with stairs weights (introduced in this paper) is competitive against the other variations of KNN algorithm, for larger values of k, in terms of classification-accuracy using gene expression data.
منابع مشابه
Long Waiting Times for Elective Hospital Care – Breaking the Vicious Circle by Abandoning Prioritisation
Background Policies assigning low-priority patients treatment delays for care, in order to make room for patients of higher priority arriving later, are common in secondary healthcare services today. Alternatively, each new patient could be granted the first available appointment. We aimed to investigate whether prioritisation can be part of the reason why waiting times for care are often...
متن کاملHealth Promotion at Local Level in Norway – Who, What, When, and How: A Response to Recent Commentaries
متن کامل
Op-brai130105 2393..2404
1 Department of Neurology, Haukeland University Hospital, Bergen, Norway 2 Department of Clinical Medicine, University of Bergen, Norway 3 Centre for Nuclear Medicine/PET, Department of Radiology, Haukeland University Hospital, Bergen, Norway 4 Institute for Surgical Sciences, University of Bergen, Norway 5 Department of Biological and Medical Psychology, University of Bergen, Norway 6 Departme...
متن کاملNo evidence for rare TRAP1 mutations influencing the risk of idiopathic Parkinson’s disease
1 Department of Neurology, Haukeland University Hospital, 5021, Bergen, Norway 2 Department of Clinical Medicine, University of Bergen, 5020, Norway 3 Center for Medical Genetics and Molecular Medicine, Haukeland University Hospital, 5021, Bergen, Norway 4 Department of Clinical Science, University of Bergen, 5020, Norway 5 Computational Biology Unit, Department of Informatics, University of Be...
متن کاملErratum to: Study protocol for a multicenter investigation of reablement in Norway
Author details Centre for Care Research Western Norway, Bergen University College, P.O Box 7030, 5020 Bergen, Norway. Department of Nursing, Bergen University College, P.O Box 7030, 5020 Bergen, Norway. Departement of Occupational Therapy, Physiotherapy and Radiography, Bergen University College, P.O Box 7030, 5020 Bergen, Norway. Haraldsplass Deaconess University College, Ulriksdal 10, 5009 Be...
متن کامل