rpy glm覚え書き - fun-sci @ ウィキ

テストデータでscipy、rpyのglmの使い方を勉強してみる。

出展はハーバード大学講義テキスト
生物統計学入門のChapter 19
データは公開されているものなのでOKかと

コードはこんな感じ

from scipy import *
from rpy import *
import csv

# low_infants.txtは下の「データ」のテキストファイル
rd1 = csv.reader(file("low_infants.txt")) 

low_inf = []

for line in rd1:
  low_inf.append(line)

# この時点では文字列のarray
low_infa = array(low_inf)

# 文字列を64精度の数値に変換
low_infa = low_infa.astype(float64)

x1=low_infa[:,4].tolist()
x2=low_infa[:,2].tolist()
y=low_inf[:,1].tolist()

ret = r.glm(r("y~x1+x2"),data=r.data_frame(x1=x1,x2=x2,y=y,family='gaussian'))

ret['coefficients']

係数は本に載っていたように出てきた。
一応、うまく行ってる？

データ

27,41,29,1360,37,0
29,40,31,1490,34,0
30,38,33,1490,32,0
28,38,31,1180,37,0
29,38,30,1200,29,1
23,32,25,680,19,0
22,33,27,620,20,1
26,38,29,1060,25,0
27,30,28,1320,27,0
25,34,29,830,32,1
23,32,26,880,26,0
26,39,30,1130,29,0
27,38,29,1140,24,0
27,39,29,1350,26,0
26,37,29,950,25,0
27,39,29,1220,25,0
26,38,29,980,28,0
29,42,33,1480,30,0
28,39,33,1250,31,1
27,38,29,1250,37,0
25,36,28,920,31,0
25,38,30,1020,21,0
24,34,27,750,21,0
31,42,33,1480,30,0
28,37,32,1140,23,0
23,32,28,670,33,1
26,36,29,1150,18,0
25,34,28,1030,20,0
23,33,29,560,29,0
28,37,30,1260,26,0
35,36,31,900,23,0
24,31,30,620,33,0
33,39,31,1440,26,0
28,39,29,1350,38,0
26,37,27,1170,23,0
26,36,27,1170,23,0
26,37,27,1170,26,0
28,40,32,1420,27,0
29,43,31,1475,27,0
28,37,28,1200,19,0
27,36,30,970,30,1
27,38,29,1190,27,0
28,39,28,1350,31,0
28,41,31,1460,29,0
22,32,27,570,23,1
22,32,25,620,35,0
28,40,30,1200,27,0
24,32,28,660,35,0
28,40,28,1280,28,0
23,33,25,830,21,0
22,31,23,680,26,0
23,32,27,640,21,0
28,39,28,1440,29,0
25,38,27,1150,32,0
27,35,27,850,32,0
24,29,26,760,24,0
23,20,25,620,19,0
21,32,23,660,41,0
25,35,26,985,33,0
22,33,24,690,17,0
27,38,29,1200,26,1
28,40,29,1370,20,0
26,35,27,1170,25,0
28,38,30,1200,30,0
27,39,30,1100,32,0
31,43,32,1480,33,0
30,38,33,1350,34,1
26,38,27,1160,23,0
27,38,31,1330,30,0
24,36,26,960,22,0
25,39,27,1080,35,0
25,36,27,1000,26,0
29,41,35,1490,34,1
25,35,28,880,18,1
29,41,30,1370,23,0
31,40,31,1325,18,1
29,39,30,1400,20,0
26,37,27,1240,29,0
23,31,25,660,18,0
23,34,25,780,25,0
25,37,26,950,14,0
25,36,29,850,39,0
28,40,29,1200,32,0
30,42,34,1440,35,1
26,38,30,1100,23,1
26,38,29,1250,18,0
29,42,33,1420,28,1
28,38,30,1400,19,0
27,36,29,1420,36,0
24,34,24,900,29,0
29,38,33,1160,33,1
23,34,25,820,39,0
28,41,32,1410,29,1
27,39,31,1300,33,1
26,38,31,1110,30,0
26,37,31,820,30,1
27,40,29,1150,28,0
28,35,32,880,35,1
28,41,33,1320,36,1
26,38,28,1080,36,0