Re: Calling a stored model within the predict() function

After a short discussion I decided to post my compete code and data set. I hope this makes sense.

Thanks,

Jim



# READ DATA
# =========
SampleCluster.txt
setwd("C:/Program Files/R/library/mboost/data")
SampleClusterData <- read.table("SampleCluster.txt",header=TRUE,sep=",")
indep <- names(SampleClusterData) [-98]# INPUTS

# GENERATE FORMULA
# ===============
modelformula <- as.formula(paste("CR ~",paste(indep,collapse="+")))

# CREATE MODEL
# ===========
setwd("C:/Program Files/R/library/mboost")
library(mboost)

# Additive Model for same Variables
# =========================
model <- gamboost(modelformula,data=SampleClusterData,baselearner="bbs")

# STORE AND CALL MODEL FILE
# ======================
save(model,file="model.RData") # STORE MODEL
# ...
# New R-Session
setwd("C:/Program Files/R/library/mboost")
model <- load("model.rdata")


# PREDICT WITH A STORED MODEL OBJECT
# ==============================
Prediction <- predict(model, newdata= SampleClusterData)

# WRITE PREDICTION TO FILE
# =====================
write.table(Prediction,file="C:/Prediction.txt")