R: Bradley-Terryモデル その1

先日の生態学会自由集会で勉強してきたBradley-Terry モデルをためしてみた。

まずはhelpにある例をためしてみる。

> library(BradleyTerry2)
> 
> data(baseball, package = "BradleyTerry2")
> 
> head(baseball)
  home.team away.team home.wins away.wins
1 Milwaukee   Detroit         4         3
2 Milwaukee   Toronto         4         2
3 Milwaukee  New York         4         3
4 Milwaukee    Boston         6         1
5 Milwaukee Cleveland         4         2
6 Milwaukee Baltimore         6         0
> 
> tapply(baseball$home.wins, baseball$home.team, sum)
Baltimore    Boston Cleveland   Detroit Milwaukee  New York   Toronto 
       10        25        18        27        28        24        22 
> 
> baseballModel1 <- BTm(cbind(home.wins, away.wins), home.team, away.team,
+                       data = baseball, id = "team")
> summary(baseballModel1)

Call:
BTm(outcome = cbind(home.wins, away.wins), player1 = home.team, 
    player2 = away.team, id = "team", data = baseball)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.6539  -0.0508   0.4133   0.9736   2.5509  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)    
teamBoston      1.1077     0.3339   3.318 0.000908 ***
teamCleveland   0.6839     0.3319   2.061 0.039345 *  
teamDetroit     1.4364     0.3396   4.230 2.34e-05 ***
teamMilwaukee   1.5814     0.3433   4.607 4.09e-06 ***
teamNew York    1.2476     0.3359   3.715 0.000203 ***
teamToronto     1.2945     0.3367   3.845 0.000121 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 78.015  on 42  degrees of freedom
Residual deviance: 44.053  on 36  degrees of freedom
AIC: 140.52

Number of Fisher Scoring iterations: 4

タグ：R