If float, the contamination should be in the range (0, 0.5]. Changed in version 0.22: The default value of ... Use dtype=np.float32 for maximum efficiency.
確定! 回上一頁