序列数据平滑（SG平滑、滑动平均平滑） |
您所在的位置：网站首页 › excel中插值函数 › 序列数据平滑（SG平滑、滑动平均平滑） |
序列数据平滑（SG平滑、滑动平均平滑）
|
经过 序列数据缺失值插补（线性插值） 处理，时间序列数据已经完整。如果需要，序列数据时序平滑对时序分析有重要帮助。其中以滑动平均平滑和SavitzkyGolay（SG）平滑较为常用。 本文以缺失值插补后的时序 MODIS 数据为输入数据，介绍基于 gma 库 的这两种数据平滑方法的实现方式。 目前，gma 提供了 SavitzkyGolay （SG平滑）和 MovingAverage（滑动平均）两种数据平滑方式。 安装 gma: pip install gma 1 读取数据 import pandas as pd # 读取原始数据 InFile = r'D:\xxxx\NDVI.xlsx' Data = pd.read_excel(InFile)[['NDVI']] print(Data.values) [0.223985 0.223985 0.223985 0.233396 0.227466 0.229259 0.230363 0.231399 0.22672 0.225723 0.224726 0.212029 0.236226 0.229548 0.236484 0.248894 0.266551 0.275269 0.274135 0.270156 0.3743545 0.478553 0.590506 0.721067 0.7600655 0.799064 0.742108 0.804224 0.83514 0.777888 0.734315 0.700915 0.602857 0.553114 0.397294 0.318083 0.289676 0.253421 0.25891 0.24601 0.23812 0.238319 0.238759 0.232261 0.253731 0.236231 ] 2 时序NDVI数据平滑 import gma # 初始化平滑参数------平滑窗口为5，平滑次数为2 SMD = gma.math.Smooth(Data['NDVI'], 5, 2) # SG 平滑，平滑多项式阶数为2，过滤器的样本间距为 1，边缘值处理方法为 'interp'。并将平滑结果添加到 ‘SG’ 列。 Data['SG'] = SMD.SavitzkyGolay(Polyorder=2, Delta=1, Mode='interp') # 滑动平均平滑。并将平滑结果添加到 ‘ME’ 列。 Data['ME'] = SMD.MovingAverage()Data的数据结果如下： NDVISGME00.2239850.2228650.22487710.2239850.2251620.22560420.2239850.2271710.22658430.2333960.2292080.22786140.2274660.2295050.22849550.2292590.229670.22891960.2303290.230110.22895170.2313990.230050.22799880.226720.2285750.22694190.2257230.2246140.226264100.2247260.2214310.226087110.2120290.2213310.227059120.2362260.2273980.230943130.2295480.2317610.236196140.2364840.2383550.243119150.2488940.2497210.250959160.2665510.2658610.26285170.2752690.2691010.281041180.2741350.2680470.310279190.2701560.2916880.355611200.3743550.3658170.419193210.4785530.478180.494744220.5905060.5993730.572358230.7210670.7067520.645911240.7600660.7653050.70615250.7990640.7778750.747505260.7421080.7765360.769282270.8042240.7999210.778868280.835140.8120620.771852290.7778880.7907990.748991300.7343150.7401770.710194310.7009150.6886550.657338320.6028570.6194240.589679330.5531140.5253110.516098340.3972940.4137350.44203350.3180830.3268820.377134360.2896760.2791630.325689370.2534210.2613680.28864380.258910.2528090.264981390.246010.2456630.252024400.238120.2410120.245428410.2383190.2364210.241954420.2387590.2373410.240452430.2322610.2394080.239434440.2537310.2405440.239483450.2362310.2416330.239223 3 制图查看 Data.plot()
|
【本文地址】
今日新闻 |
推荐新闻 |