library(conjoint)
#Load the three csv files first then you should be able to display them
cprof
cmatrix
clevels
# Alter row names
rownames(cprof)<-cprof[,1] # make the row names the first column imported
cprof<-cprof[,-1] # now remove the column that was imported with the row names
rownames(cmatrix)<-cmatrix[,1] # make the row names the first column imported
cmatrix<-cmatrix[,-1] # remove the first column which contained the row name
# Show Matrix for each profile - you can see the different attribues for each car
# (style Sport(1)/Luxury(2), color Red(1)/Black(2), gastype Eco(1)/NonEco(2))
cprof
#Now examine the first row of the matrix created
cmatrix[1,] # Corresponds to the first persons preference of each of the profiles.
#Lets look at the model for the first respondent
caModel(y=cmatrix[1,], x=cprof) # pass in the first row of preference matrix, then the profile matrix as x
#Notice the results are showing negative and positive numbers for the estimate. These are synonmous with
# This shows the vector of the utility for the first respondent. Notice that some of the levels are missing becuase this
# is a regression equation, howver if you look at the vector at the bottom all the values are there.
# To intrepet this look for the higher values as having more utility than the other attributes of the same factor
caUtilities(y=cmatrix[1,], x=cprof, z=clevels)
# Further we can use this to identify the relative importance of each factor for a given subject
caImportance(y=cmatrix[1,], x=cprof)
caImportance(y=cmatrix[2,], x=cprof)
caImportance(y=cmatrix[3,], x=cprof)
# this will provive all the utilities for every single observation, use this command and interpret the same way
caPartUtilities(y=cmatrix, x=cprof, z=clevels)
# Model Paramters for the whole sample
Conjoint(y=cmatrix, x=cprof, z=clevels)
#Notice how at the bottom each of the factors has a given "Average importance" Associated with them.
#38.76 / 30.23/ 31.01 <- meaning that style of the car has the higher average importance
# We can also use this to segment out the customers in 3 to 5 clusters, to see how they might group together.
caSegmentation(y=cmatrix, x=cprof, c=3)