如何使用sklearn EllipticEnvelope

时间:2019-05-21 15:13:59

标签: python-3.x machine-learning scikit-learn

从sklearn EllipticEnvelope获取异常值的索引

我正在尝试获取离群值的索引,但我无法做到

features, _= make_blobs(n_samples=100, centers=1, n_features=3)
    #update a few elements with large int:
    features[0,0] = 10000
    features[0,1] = 10000
    features[20,0] = 1000
    features[20,1] = 8000
    features[30,1] = 2000

    dataset = 


 pd.DataFrame({'Column1':features[:,0],'Column2':features[:,1],'Column3':features[:,2]})

    outlier_detector = EllipticEnvelope(contamination=.1)
    outlier_detector.fit(dataset)
    outlier_detector.predict((dataset))

我尝试了dataset.index,但是尽管使用了reshape(),它仍然告诉我“期望的2D数组,得到了1D数组”

有人有什么想法吗? 谢谢

0 个答案:

没有答案