We use a variety of classication tasks to test the dierent weighting algorithms.
Because we hypothesize that dierent data sets will require diering degrees of
locality for greatest accuracy, we include a range of articial and real-world domains.
The data sets used are shown in the leftmost column of Table 1. The rst
six of these (LED-7, LED-24, Monks-2, Lymph, Promoters, and Soybean)2 are
from the UCI machine learning repository The tasks selected are a subset of those proposed as a benchmark by Zheng (1993). Tasks
We use a variety of classi cation tasks to test the di erent weighting algorithms.Because we hypothesize that di erent data sets will require di ering degrees oflocality for greatest accuracy, we include a range of arti cial and real-world domains.The data sets used are shown in the leftmost column of Table 1. The rstsix of these (LED-7, LED-24, Monks-2, Lymph, Promoters, and Soybean)2 arefrom the UCI machine learning repository The tasks selected are a subset of those proposed as a benchmark by Zheng (1993). Tasks
การแปล กรุณารอสักครู่..