Flags Data Set
Below are papers that cite this data set, with context shown.
Papers were automatically harvested and associated with this data set, in collaboration with Rexa.info.
Return to Flags data set page.
Ron Kohavi and George H. John and Richard Long and David Manley and Karl Pfleger. MLC++: A Machine Learning Library in C. ICTAI. 1994.
octag yes no yes no yes no yes no yes no yes no yes no yes no yes no yes no yes no yes no red yellow green blue red yellow green blue red yellow green blue sword balloon flag Figure 5: The Monk3 dataset (top), the concept predicted by ID3 (center), and the errors in black (bottom). 14 0 100 200 300 400 TS size Monk1 0.7 0.75 0.8 0.85 0.9 0.95 1 Accuracy HOODG ID3 Figure 6: Learning curves generated
George H. John and Ron Kohavi and Karl Pfleger. Irrelevant Features and the Subset Selection Problem. ICML. 1994.
The C4.5 program is the program that comes with Quinlan's book (Quinlan 1992); the ID3 results were obtained by running C4.5 and using the unpruned trees. On the artificial datasets, we used the ``-s -m1'' C4.5 flags which indicate that subset splits may be used and that splitting should continue until purity. To estimate the accuracy for feature subsets, we used 25-fold
Wl/odzisl/aw Duch and Karol Grudzi nski and Grzegorz Stawski. SYMBOLIC FEATURES IN NEURAL NETWORKS. Department of Computer Methods, Nicolaus Copernicus University.
Both FSM network and k-NN algorithm achieved similar results (within statistical accuracy) on the raw data as on the converted data (the number of attributes did not change here). On the Flags dataset, containing 194 samples, 3 continuous and 25 symbolic attributes, 8 classes (majority rate 30.8%) and no missing values, 10-fold cross-validation tests were performed. Large improvement is observed