jeremycruzz/msds301-wk7

msds301-wk7

Setup (MAC i9, zshrc)

• NOTE: Executable files are included but I built them on an i9 Macbook
• Clone repo with git clone --recursive git@github.com:jeremycruzz/msds301-wk7.git
• Install GoLearn System Dependencies instructions here
  • Make sure you have g++ and MYSS2instructions here
  • Install OpenBLAS in any directory
    • git clone https://github.com/xianyi/OpenBLAS
    • cd OpenBLAS
    • make
    • alternatively brew install openblas
  • Link to library
    • CGO_LDFLAGS="-L/path/to/OpenBLAS -lopenblas" go install github.com/gonum/blas/cgo

Building executable

• Run go build ./cmd/isotree

Running Go executable

• Run ./isotree

Results

I decided to go with the GoLearn Package as I had experience with the package from the previous weeks assignment. I thought it was pretty straightforward when doing linear regression but it was not quite as straight forward since I did not load the data from a csv. Creating the dataframe from scratch proved to be a bit difficult. There was a weird way to add attributes and I was not able to add a class attribute as a categorical attribute but was forced to create a float attribute for the label. Unit tests were not done but I would regularly print out the data and rebuild the project as I was working on it. This is wear scripting languages have an advantage over go, as you can run a few lines and explore the data without needing to rebuild a project. After running the tests we can see that the results that Go produced were similar to those that python produced.

For the firm, I would reccomend that we try to explore a different go package that is easier to work with. The Go program was much faster than the R and python scripts and the speed would be an advantage in our data processing pipelines.

Language	Average	Median	Min	Max
R	0.265768	0.265265	0.187304	0.388845
Python	0.443680	0.438875	0.371611	0.631608
Go	0.445456	0.444328	0.304279	0.661401