拟合混合效应随机森林模型时出现内存错误消息

时间:2019-04-16 00:18:20

标签: python

我正在尝试在python中使用一个新包,即MERF(混合效果随机森林)。当我想将200000+行的数据与少量的簇(<100)拟合时,该模型始终会输出一条内存错误消息。我认为问题在于群集的数量。当我使用非常大的群集号(> 10000)时,它给出了有效的输出。

from merf import MERF

merf = MERF()

clusters_train = np.array([])
for i in np.arange(len(df_train)):
    if 1570<=df_train['fss'][i]<=1875:
        clusters_train = np.append(clusters_train,1)
    elif 1510<=df_train['fss'][i]<=1569:
        clusters_train = np.append(clusters_train,2)
    elif 1450<=df_train['fss'][i]<=1509:
        clusters_train = np.append(clusters_train,3)
    elif 1340<=df_train['fss'][i]<=1449:
        clusters_train = np.append(clusters_train,4)
    elif 1001<=df_train['fss'][i]<=1339:
        clusters_train = np.append(clusters_train,5)
    else:
        clusters_train = np.append(clusters_train,0)
clusters_train = pd.Series(clusters_train)

X_train = df_train[['ccs','pydx','gbr']]

Z_train = df_train[['pct30p','pct90p']]

y_train = y_train['bad']

merf.fit(X_train, Z_train, clusters_train, y_train)

Error Message

0 个答案:

没有答案