KernelADASYN是一种过采样方法，用于处理类别不平衡问题。它结合了核密度估计和最近邻思想，通过Markov Chain Monte Carlo实现。参数包括比例、邻居数和核带宽。在采样过程中，若协方差矩阵不适定，会进行降维处理。该方法通过生成新的少数类样本来平衡数据集。
classKernelADASYN(OverSampling):"""Notes:
* The method of sampling was not specified, Markov Chain Monte Carlo has been implemented.
* Not prepared for improperly conditioned covariance matrix."""categories=[OverSampling.cat_density_estimation,
OverSampling.cat_extensive,
OverSampling.cat_borderline]def __init__(self, proportion= 1.0, k= 5, h= 1.0, n_jobs= 1):"""Constructor of the sampling object
Args:
proportion (float): proportion of the difference of n_maj and n_min to sample
e.g. 1.0 means that after sampling the number of minority
samples will be equal to the number of majority samples
k (int): number of neighbors in the nearest neighbors component
h (float): kernel bandwidth
n_jobs (int): number of parallel jobs"""super().__init__()
self.check_greater_or_equal(proportion,&
最低0.47元/天 解锁文章
扫一扫
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
