1、實(shí)驗(yàn)三Pandas進(jìn)行數(shù)據(jù)預(yù)處理任務(wù) 1 插補(bǔ)用戶電量數(shù)據(jù)的缺失值。(1)讀取 misssing_data.csv 表中的數(shù)據(jù)。import pandas as pds二pd.read_excel(D:桌面實(shí)驗(yàn)三da tamissing_da ta.xls) print（數(shù)據(jù)輸出為：n,s）In 1: run file ( J C: /Us ers/Admin/ :spy 川巴m-py?/亡曰P!y J wdir=J 匚：/Us er s/ spyder-py3T) 數(shù)據(jù)輸出為:235,8333324.0343478-32310236.2708325.6379515,45641238.0521

2、328.0897517.09092235.9063NaN514.8900弓236.7604268.8324NaN4hlaN404.04804S6.09125237.4167391.2652516,23306238.6563380.8241NaM7237.6042388.0230435.35088238.031320S.43494S7.67509235 a.0729NaNNaM102弓乩.53134&9.787660.234711411.2069621.234612234.4688395.2343611,340813235.344.8221643.086314235.6354385.643264

3、2.348-215234.5521401.6234Nahl16236a：W&0409.6489602.934717235.2396416.8795弓呂9.345718235.4896NaN556-345219236.9688MaN弓予呂.3470查詢?nèi)笔е邓诘奈恢?。import pandas as pds=pd.read_excel(D:桌面實(shí)驗(yàn)三da tamissing_da ta.xls)print(缺失值的位置為：n,s.isnull().sum()In 2: runfile( JC:/Users/Admin/ spj/der-pyS/tewppy J wdir =JC;/Users/

4、Admin/ spyder-py3 缺尖價(jià)的f立置知 TOC o 1-5 h z 235.83332324.93434478.32314dtype: int&4import pandas as pd import numpy as nparr=np.array(0,1,2)missing_da ta二卩4.284_6乂。6丨(。:桌面實(shí)驗(yàn)三da tamissing_da ta.xls,names二arr) prin t(lagrange 插值前(False 為缺失值所在位置)，n,missing_da ta.no tnull()In 3: runfilef xC:/Users/Adniin/

5、spyder-py3/temp. pyJj wdir=rCr/Users/Admin/. spyd已廠-py弓”) lagrange插值前(Falsel缺失值所在f立畫)010TrueTrueTrue1TrueTrueTrue2TrueFalseTrue3TrueTrueFalse4FalseTrueTrue5TrueTrueTrue6TrueTrueFalse7TrueTrueTrue8TrueTrueTrue9TrueFalseFalse10TrueTrueTrue11FalseTrueTrue12TrueTrueTrue13TrueTrueTrue14TrueTrueTrue15True

6、TrueFalse16TrueTrueTrue17TrueTrueTrue18TrueFalseTrue19TrueFalseTrue(3)使用SciPy庫(kù)中interpolate模塊中的lagrange對(duì)數(shù)據(jù)進(jìn)行拉格朗日插值。( 4 )查看數(shù)據(jù)中是否存在缺失值，若不存在則說明插值成功。import pandas as pdimport numpy as nparr=np.array(0,1,2)missing_da ta二卩4.284_6乂。6丨(。:桌面實(shí)驗(yàn)三da tamissing_da ta.xls,names二arr)print(lagrange 插值前(False 為缺失值所在位置

7、)，n,missing_data.notnull()from scipy. interpolate import lagrangefor i in range(0,3):la=lagrange(missing_data.loc:,i.dropna().index,missing_data.loc:,i.dropna().values) list_d=list(set(np.arange(0,20).difference(set(missing_data.loc:,i.dropna().index) missing_data.loclist_d,i=la(list_d)print(第d 列缺失值

8、的個(gè)數(shù)為：d%(i,missing_data.loc:, i.isnull().sum()prin t(lagrange 插值后(False 為缺失值所在位置)n,missing_da ta.no tnull()VuidbldC1A14VuidbldC1A14TrueTiru eTrue15TrueTrueFal-se1&TirueTru&Tee17TrueIfueTrue邊True-FalseTrue19TrueFalseTruelagrsngelfi后Fohe為蔵矢 16斯?if立蓋B1TrueTiueTrue-1TrueTrueTrue2TrueTrueTrue3TrueTrueTrue

9、斗TrueTsUeTrue-5TrueTrueTrue6T rueTrueTrue7TrueTrueTrue呂TrueTmueTrue-9TrueTrueTrueIETrueTrueTrue11TrueTrueTrue-12TrueTrueTrue13TrueTrueTrue14TrueTrueTrue15TrueTmueTrue-1&TrueTrueTrue17TrueTrueTrueIBTrueTrueTrue19TrueTJUeTrue-irapcrt pandas az pd iraport nunpy as np amr-np-array（&a 2）ndiilnfi_data=pd-

10、read_ekcel（*02 矣質(zhì) 實(shí)軽三iddiai jossjrtgdotia.xtinaoei-arr） print（jJJffrfS-作機(jī)生亡幽輸后3菊譽(yù)注IP*a rjTJaFi55ing_data .notnull（）-frcun Aci.py4 interpcL-ate impurt la grange far 1 in range（03）:la-lagmangef ini55ing_data ,locs ,i. dropna （） indexj rd 55ing_dst a. locdnscKia （）List_d =lis t（iet （np: a range （0.,. 2

11、）difference （iet（rai5 sing_da 七日丄口匚；丘 dropn-a mis&ing_dBta.locli暑七_(dá)日i=lalis七jprdntf “華id 刑曲黃適的密缶念貂性gdata - lac :,!. isnullf ）. lbi（ ）printf l-irrrang書心 1,弋有缺矢應(yīng)瞬在垃蹩 uLssing ds ta nctnullf ）任務(wù) 2 合并線損、用電量趨勢(shì)與路線告警數(shù)據(jù)（1）讀取 ele_loss.csv 和 alarm.csv 表。（2）查看兩表的形狀。import pandas as pdele_loss二pd.read_csv（D:桌面實(shí)驗(yàn)

12、三da taele_loss.csv）alarm二pd.read_csv（D:桌面實(shí)驗(yàn)三da taalarm.csv, encoding二gbk）print（ele_loss 表的形狀為,ele_loss.shape）print（alarm 表的形狀為,alarm.shape）In 6: runfile （ J匚：/Mser：s/A曲1in/ spyder-py3/temp -py j wdir= rC:/Users/Admin/ spyder-py3r） elejoss的形狀為（49 4）alarm表的形狀芮（253）（3）以 ID 和 date 兩個(gè)鍵值作為主鍵進(jìn)行內(nèi)連接。（4）查看合并后

13、的數(shù)據(jù)。import pandas as pdele_loss=pd.read_csv（D:桌面實(shí)驗(yàn)三da taele_loss.csv）alarm二pd.read_csv（D:桌面實(shí)驗(yàn)三dataalarm.csv, encoding二gbk）print（ele_loss 表的形狀為,ele_loss.shape）prin t（alarm 表的形狀為,alarm.shape）merge=pd.merge（ele_loss,alarm,left_on=ID,date,right_on=ID,date,how=inner） print（合并后的表形狀為：,merge.shape）print（合并后

14、的表為：merge）In 7: runf ile( rC:/Users/Admin/ spyder-pyS/temppy J wdir= elE_loSS表的形狀為(49 4) alarms的形狀為(25. 3) 合并后的表形狀為：5)合并后的恙丸:IDdate eleloss alarm0212610012010/9/3858.00,151048電流不平衡1212610012010/9/6883.50,162778匚相電流過負(fù)荷2212610012010/9/16915.50.162028匸相電流過負(fù)荷3212610012010/9/16915.50,16202&C相電涼過負(fù)荷4212610

15、012910/9/17961.00-166650代相電涼過負(fù)荷5212610012010/9/17961.00,166650電涼不平衡任務(wù)3 標(biāo)準(zhǔn)化建模專家樣本數(shù)據(jù)(1)讀取 model.csv 數(shù)據(jù)。(2)定義標(biāo)準(zhǔn)差標(biāo)準(zhǔn)化函數(shù)。import pandas as pdmodel二pd.read_excel(D:桌面實(shí)驗(yàn)三da tamodel.xls,encoding二gbk) def Standard(data):data=(data-data.mean()/data.std()return dataS=Standard(model)print(標(biāo)準(zhǔn)化后的數(shù)據(jù)為：，n,S.head()In 8

16、: run file ( J 匚；/Us ers/Admin/ spyder -py3/t emp pyS wdir=J 匚；/Us ers/Admin/ spyder-py3J) 標(biāo)準(zhǔn)化后的數(shù)據(jù)為：電里趨勢(shì)下隱指標(biāo) 線損指標(biāo)告警類指標(biāo)是否朝屆電0&.6127770.974549-0.1612162.500687&.612777-1.&2259&2.6536162.5006873433520.974549-0.1612162.5006873,&03&99-1.&2259&-1.099494398516&.1347120.974549-1.099494398516使用函數(shù)分別對(duì)3 列數(shù)據(jù)進(jìn)行標(biāo)準(zhǔn)

17、化。查看標(biāo)準(zhǔn)化后的數(shù)據(jù)import pandas as pdmodel二pd.read_excel(D:桌面實(shí)驗(yàn)三da tamodel.xls,encoding=gbk)def Standard(data):data=(data-data.mean()/data.std()return dataS=Standard(model)print(標(biāo)準(zhǔn)化后的數(shù)據(jù)為：，n,S.head()def MinMaxScale(data):data=(data-data.min()/(data.max()-data.min()return dataM=MinMaxScale(model)print(離差標(biāo)準(zhǔn)化后

18、的數(shù)據(jù)為：，n,S.head()import numpy as npdef DecimalScaler(data):data=data/10*np.ceil(np.log10(data.abs().max()return dataD=DecimalScaler(model) prin t(小數(shù)疋標(biāo)差標(biāo)準(zhǔn)化的數(shù)據(jù)為:,n,D.head()In 9: runfile( JC:/Users/Admin/ spyder-py3/temppy J wdir=JC:/Users/Admin/:已廠_py弓 標(biāo)準(zhǔn)化后的數(shù)據(jù)芮：電里趨勢(shì)下眸指標(biāo)線損指標(biāo)告警類指標(biāo)是否竊掃電&0-6127770.974549-0

19、.1612162.50068710-612777 -1.02259ft2.6536162.5006872-0.3433520.974549-0.1612162.50068733.003099 -1.02259ft-1.099494 -0.39851640-1347120.974549-1.099494 -0.39&516離差標(biāo)準(zhǔn)化后的數(shù)據(jù)芮:電重趨勢(shì)下眸指標(biāo)線損指標(biāo)告誓類指標(biāo)是否竊掃電00.6127770.974549-0.1612162500S8710-612777 -1.02259ft2.65361625006872-0.3433520.974549-0.161216250068733.&

20、03099 -1.02259ft-1.099494 -039851640-1347120.974549-1.099494 -039&516小數(shù)定標(biāo)差標(biāo):隹化的埶據(jù)芮：電重翁勢(shì)下陳指標(biāo) 線損指標(biāo) 告誓類指標(biāo) 是否密掃電&0.41.00.11.010.40.00.41.020.21.00.11.030.90.00.00.040.31.00.00.0任務(wù) 4 完成以下操作1、使用如下方法規(guī)范化數(shù)組：200,300,400,600,1000 (1)離差標(biāo)準(zhǔn)化(2)標(biāo)準(zhǔn)差標(biāo)準(zhǔn)化(3)小數(shù)定標(biāo)標(biāo)準(zhǔn)化import pandas as pdnum=200,300,400,600,1000df=pd.DataF

21、rame(num)print(df.describe()df0=(df-df.min()/(df.max()-df.min() #離差標(biāo)準(zhǔn)化 print(df0)print(df.mean() #標(biāo)準(zhǔn)差標(biāo)準(zhǔn)化df1=(df-df.mean()/df.std()print(df1)df2=(df/(10*4) #小數(shù)疋標(biāo)標(biāo)準(zhǔn)化print(df2)In 10 : runfile (C:/Users/Admin/, spyder-py3/temp.py13 wdir=JC:/Users/Ad/sin/. spyder-py3r 0count5. &mean500.0&0std316.227766mi

22、n200.0&00&02.5%300.O&00&05&%&00&075%600.O&00&0ma x1600.&0& 0.0& TOC o 1-5 h z &.1259.250&.5&S1.0&5&S.&dtype: float640& -9.948683-0.632456-0.31622B&.3162281.5S1139& 0.02 TOC o 1-5 h z 0.030.040.060.102、12 個(gè)銷售價(jià)格記錄如下：5，10，11，13，15，35，50，55，72，92，204，215。使用等寬法對(duì)其進(jìn)行離散化 處理(分 3 類)import pandas as pdnum=5,10,11,13,15,35