如何使用Numba运行此Dataframe循环？

Question

#Stastical Features

timeSeries = [375, 750, 1125, 1875, 3750, 7500, 15000]

for i in range(15000, masterDF.shape[0]):

    for tms in timeSeries:
    
        ROCseries = masterDF['minROC'][i-tms:i]
        masterDF.loc[i, 'rocKurtosis{}'.format(tms)] = ROCseries.kurtosis()
        masterDF.loc[i, 'rocSkew{}'.format(tms)] = ROCseries.skew()
        masterDF.loc[i, 'rocAbsSum{}'.format(tms)] = ROCseries.abs().sum()

masterDF大约有5300万行，我可以做的是将索引、时间序列和偏移值存储在一个数组中，但是我如何为它分配相应的列呢