Python clx.analytics.anomaly_detection.dbscan用法及代碼示例

用法: clx.analytics.anomaly_detection.dbscan(feature_dataframe, min_samples=3, eps=0.3)

將特征 DataFrame 傳遞給此函數以檢測特征 DataFrame 中的異常。該函數使用cuML DBSCAN 檢測異常並輸出相關標簽 0,1,-1。

參數：

:param feature_dataframe: Feature dataframe to be used for clustering：
:type feature_dataframe: cudf.DataFrame：
:param min_samples: Minimum samples to use for dbscan：
:type min_samples: int：
:param eps: Max distance to use for dbscan：
:type eps: float：

例子：

>>> import cudf
>>> import clx.features
>>> import clx.analytics.anomaly_detection
>>> df = cudf.DataFrame(
>>>         {
>>>             "time": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14],
>>>             "user": ["u1","u1","u1","u1","u1","u1","u1","u1","u1","u1","u5","u4","u2","u3"],
>>>             "computer": ["c1","c2","c3","c1","c2","c3","c1","c1","c2","c3","c1","c1","c5","c6"],
>>>         }
>>>     )
>>> feature_df = clx.features.frequency(df, entity_id="user", feature_id="computer")
>>> labels = clx.analytics.anomaly_detection.dbscan(feature_df, min_samples=2, eps=0.5)
>>> labels
    0   -1
    1   -1
    2   -1
    dtype: int32

相關用法

注：本文由純淨天空篩選整理自rapids.ai大神的英文原創作品 clx.analytics.anomaly_detection.dbscan。非經特殊聲明，原始代碼版權歸原作者所有，本譯文未經允許或授權，請勿轉載或複製。