After a discussion with others, I guessed the latter may be more possible,
according to the .info file
※ 引述《brooky (未夠班)》之銘言：
: Hi,
: I am a little confused with the requirement of hw2.
: The requirement says,
: "We are interested in whether a single nominal attribute
: with 40 possible values or 40 binary attributes is better"
: Is this means, we should build classification tree according to
: these 40 binary attributes separately, and compare the result to
: "attribute with 40 values"?
: ex: according to "binary attribute 1"
: v.s.
: according to "40-value attributes"
: Or should we build a classification tree according to 40 binary
: attributes all together?
: ex: according to "binary attribute 1~40"
: v.s.
: according to "40-value attributes"
: If so, how to do this?
since you will have two distinct trees, so their result needs to be
considered as usual