What value should the class label be in regression?

[BramVanroy]

In the README it says:

: : ... . . . Each line contains an instance and is ended by a '\n' character. For classification, is an integer indicating the class label (multi-class is supported). For regression, is the target value which can be any real number.

However, we found that the label doesn't need to be an integer on Linux, as it also works if you use a string. For instance, using UNK (from unknown) works - but not on Windows.

To ensure a similar experience across operating systems, which default value is encouraged? Documentation says 'any integer', so can I just use 0?

[cjlin1]

To train regression you must put a target value there.

Bram Vanroy writes:

In the README it says:

: : ...
.
.
.

Each line contains an instance and is ended by a '\n'
character. For
classification, is an integer indicating the class label
(multi-class is supported). For regression, is the target
value which can be any real number.

However, we found that the label doesn't need to be an integer on Linux, as it also works if you use a string. For instance, using UNK (from unknown) works - but not on Windows.

To ensure a similar experience across operating systems, which default value is encouraged? Documentation says 'any integer', so can I just use 0?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.*

[BramVanroy]

But the target value is unknown, right? It's the one you are trying to predict.

[cjlin1]

for prediction any value is ok Bram Vanroy writes:

But the target value is unknown, right? It's the one you are trying to predict.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.*

[BramVanroy]

So just using something like the following, where the label is 0 is okay?

0 1:4.458333333333333 2:24.0 3:0.20833333333333334 4:8.333333333333334 5:29.166666666666668 6:87.5 8:1.0

[cjlin1]

yes Bram Vanroy writes:

So just using something like the following, where the label is 0 is okay?

0 1:4.458333333333333 2:24.0 3:0.20833333333333334 4:8.333333333333334 5:29.166666666666668 6:87.5 8:1.0

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.*