记录| 解决statsmodels-OLS报错“Pandas data cast to numpy dtype of object.Check input data with np.asarray “

import statsmodels.api as sm

对初始dataframe ' data ' 新添列 ' R '，初始操作

N, M = data.shape
data['R'] = ''
data.iloc[0, -1] = 0
for i in range(1,N):
    data.iloc[i, -1] = (data.iloc[i, 1] - data.iloc[i-1, -1])/data.iloc[i-1, -1]

使用OLS，

df = data.dropna()
Y = df['R'].copy()
X = df.iloc[:, 0:2]

mod = sm.OLS(Y, X)
res = mod.fit()
coef = list(res.params)

print(coef)

报错："Pandas data cast to numpy dtype of object.Check input data with np.asarray "

首先，使用网上查到的方法仍报错

mod = sm.OLS(Y, X.astype(float))

可能是新添列出现了问题， print(data['R'])后显示dtype为object，更改

data['R'] = 0

print(data['R'])后显示dt