Algorithms for clustering data
Algorithms for clustering data
C4.5: programs for machine learning
C4.5: programs for machine learning
A conceptual version of the K-means algorithm
Pattern Recognition Letters
| Hi-index | 0.00 |
This paper presents a st udy on mixed similarity mea su res (MSM) th at allows doing classification and cluste ring in many sit uations with out discreti eat.ion. For supervised classification we do experimental com parative studies of cl as sifiers bu ilt by dec isio n t ree induction system C4.5 and k ne a rest n eighbor ru le usin g MS:-'L Fo r unsu per vised clustering we first intro duce an extension of k-means algorithm for m ixed numeric and symbolic da ta, t hen evaluate clusters obtained by th is algorithm with natural classes. Exp erimental studies allow us to draw conclusions (meta-knowledge) that are significant in pract ice about t he mutu al use of discreti zation techniques and MSM.