- julia 1+, download binaries from https://julialang.org/downloads/ or build from source https://github.com/JuliaLang/julia/tree/v1.1.0
- jupyter installed through miniconda3. Use the appropriate 64 bit installer for Python 3.7 from https://docs.conda.io/en/latest/miniconda.html. Make sure conda is added to system path and once it is install jupyter with
conda install jupyter
. jupyter should be in the system path as well which can be confirmed withwhich jupyter
in the terminal. - to set up proper environment for notebook, clone the repository to your computer and run
julia envsetup.jl
within that directory. Doing so will install all the package dependencies. This assumes julia1.x is in your PATH variable, otherwise just type the path to the julia binary to run the setup script.
- Clone the repository to a local directory and run the envsetup.jl script as explained above
- Navigate to directory in terminal and execute
jupyter notebook
. A web browser should open displaying the FluxRegressionComparison directory. Click on the notebookJulia 1.5 Regression Framework Comparison.ipynb
- Alternatively click the binder badge to open the repository in the browser and click on
Julia 1.5 Regression Framework Comparison.ipynb
once the repository loads.
Presentation on Machine Learning Frameworks for Regression
[Document on Measuring Estimation Uncertainty Using Log Likelihood Cost Function](https://docs.google.com/document/d/1odQX81f04KFCcqdplzjVbyVvRZqu-IW8WpHk4U5sgb0/edit?usp=sharing