, 34(1): 28–36. February 2001.\n
\n
@article{Dreiseitl2001,\n title = {A Comparison of Machine Learning Methods for the Diagnosis of Pigmented Skin Lesions.},\n author = {Dreiseitl, S. and {Ohno-Machado}, L. and Kittler, H. and Vinterbo, S. and Billhardt, H. and Binder, M.},\n year = {2001},\n month = feb,\n journal = {J Biomed Inform},\n volume = {34},\n number = {1},\n pages = {28--36},\n doi = {10.1006/jbin.2001.1004},\n url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.13.2630},\n abstract = {We analyze the discriminatory power of k-nearest neighbors, logistic regression, artificial neural networks (ANNs), decision tress, and support vector machines (SVMs) on the task of classifying pigmented skin lesions as common nevi, dysplastic nevi, or melanoma. Three different classification tasks were used as benchmarks: the dichotomous problem of distinguishing common nevi from dysplastic nevi and melanoma, the dichotomous problem of distinguishing melanoma from common and dysplastic nevi, and the trichotomous problem of correctly distinguishing all three classes. Using ROC analysis to measure the discriminatory power of the methods shows that excellent results for specific classification problems in the domain of pigmented skin lesions can be achieved with machine-learning methods. On both dichotomous and trichotomous tasks, logistic regression, ANNs, and SVMs performed on about the same level, with k-nearest neighbors and decision trees performing worse.},\n copyright = {All rights reserved},\n pubmedid = {11376540},\n keywords = {11376540,Adult,Algorithms,Anonymous Testing,Artificial Intelligence,Bias (Epidemiology),Carcinoma,Child,Comparative Study,Computational Biology,Computer-Assisted,Computerized,Confidentiality,Data Interpretation,Databases,Decision Trees,Demograph,Diagnosis,Differential,Disclosure,DNA,Female,Gene Expression,Gene Expression Profiling,Gene Expression Regulation,Genetic Markers,Humans,Internet,Logistic Models,Lung Neoplasms,Male,Medical Records Systems,Melanoma,Messenger,Middle Aged,Multivariate Analysis,Neoplasm,Neoplasms,Neoplastic,Neural Networks (Computer),Nevus,Non-U.S. Gov't,Oligonucleotide Array Sequence Analysis,P.H.S.,Pigmented,Privacy,Research Support,Rhabdomyosarcoma,RNA,ROC Curve,Sarcoma,Skin Diseases,Skin Neoplasms,Skin Pigmentation,Small Cell,Software,Statistical,Statistics,U.S. Gov't,y},\n file = {/Users/staal/Documents/Zotero/storage/2JCESE87/dreisetl2001-JBI.pdf;/Users/staal/Documents/Zotero/storage/6MG8J3W4/dreisetl2001-JBI.pdf;/Users/staal/Documents/Zotero/storage/FW4VR7RH/dreisetl2001-JBI.pdf}\n}\n\n
\n
\n We analyze the discriminatory power of k-nearest neighbors, logistic regression, artificial neural networks (ANNs), decision tress, and support vector machines (SVMs) on the task of classifying pigmented skin lesions as common nevi, dysplastic nevi, or melanoma. Three different classification tasks were used as benchmarks: the dichotomous problem of distinguishing common nevi from dysplastic nevi and melanoma, the dichotomous problem of distinguishing melanoma from common and dysplastic nevi, and the trichotomous problem of correctly distinguishing all three classes. Using ROC analysis to measure the discriminatory power of the methods shows that excellent results for specific classification problems in the domain of pigmented skin lesions can be achieved with machine-learning methods. On both dichotomous and trichotomous tasks, logistic regression, ANNs, and SVMs performed on about the same level, with k-nearest neighbors and decision trees performing worse.\n