R语言在hadoop的线性回归的可用代码

X=matrix(rnorm(20000),ncol=10)

X.index=to.dfs(cbind(1:nrow(X),X))

y=as.matrix(rnorm(2000))

Sum=function(.,YY)

  keyval(1,list(Reduce('+',YY)))

XtX=values(

  from.dfs(

    mapreduce(

      input=X.index,

      map=function(.,Xi){

        yi=y[Xi[,1],]

        Xi=Xi[,-1]

        keyval(1,list(t(Xi) %*% Xi))

      },

      reduce=Sum,

      combine=TRUE)))[[1]]

Xty=values(

  from.dfs(

    mapreduce(

      input=X.index,

      map=function(.,Xi){

        yi=y[Xi[,1],]

        Xi=Xi[,-1]

        keyval(1,list(t(Xi) %*% yi))

      },

      reduce=Sum,

      combine=TRUE)))[[1]]

 solve(XtX,Xty)