Holt-Winters seasonal method

[rchanda]

No description provided.

[sryza]

This looks awesome. Thanks a ton.

Curious to hear any details of the R comparison? The error tends to be higher than R? How much higher? Can still accept the PR if so, just want to be able to document this properly.

A few notes:

• Line endings should be under 100 characters
• Indentation is inconsistent in some files
• The branch currently has merge conflicts
• Classes and methods need documentation to describe what they do and which methods are used
• Files need copyright headers at the top

Are you interested in fixing these things? If not, let me know, and I can work on them when I get the chance.

[dircsem]

Why does the forecast method receive two vectors? Isn't better to create the forecast vector inside this method?

[dircsem]

Ok, but don't you think that would be better to maintain the same structure than is in ARIMA?
I can do this change, and some others like
for (i <- 0 to (initSeason.size - 1)
to
for (i <-0 until (initSeason.size )
if I have write access to this branch.

[sryza]

you are a champion. This looks great. The main nits I have are cosmetic:

• Indentation needs to be kept at a consistent 2 spaces.
• There needs to be a space between "if" and "for" keywords and the parentheses that follow them.
• The patch needs to be rebased so that there are no merge conflicts.

Let me know if you're interested in making these changes. If not, I'll happily do so.