Some Julia specific coding advice and a fix.

[Nosferican]

• .gitignore: including .DS_Store which is a Mac OS specific binary file which should be ignored.
• Your Array where eltype <:Union{Float64,Missing} when is better to have them <:Float64 if no missing values.
• For sum is better to use the broadcasted optional function rather than allocate the vector as before.
• Fixed predictions.

[hpoit]

Thanks Jose! What I really need is to minimize J with LLS and get it down to zero https://github.com/hpoit/ML-Coursera/blob/master/week%202/ang_assignment1ab-v02.jl#L51, so my regression line can be the same as Andrew Ng's, which is just a little bit higher in y-intercept and slope. Still need your help!