pandas.DataFrame遍历

pandas.DataFrame遍历速度会很慢，为了方便没有用其他语言写

计算当前行数据前N行的平均值，方差，标准差等

#当前行前300条数据协方差
data_speed_cv = pd.DataFrame([0] * 300)

for i in range(300, len(data)):
    data_tmp = data[(data.index >= i - 300) & (data.index < i)]
    speed_cv_std = std(data_tmp.speed)#标准差
    speed_cv_mean = mean(data_tmp.speed)#均值
    speed_cv = speed_cv_std / speed_cv_mean#协方差
    data_speed_cv = data_speed_cv.append([speed_cv])

data_speed_cv = data_speed_cv.as_matrix()
data['speed_cv'] = data_speed_cv