utterances-bot commented 3 years ago

Hotelling's T^2 in Julia, Python, and R - Data & The World

czylabsonasa commented 3 years ago

for julia (at least in 1.6) the convert does not work anymore, so instead of:

versicolor = convert(Matrix, iris[iris.Species .== "versicolor", 1:2])
virginica = convert(Matrix, iris[iris.Species .== "virginica", 1:2])

use

versicolor = Matrix(iris[iris.Species .== "versicolor", 1:2])
virginica = Matrix(iris[iris.Species .== "virginica", 1:2])

jose-matos commented 9 months ago

Matrix operations aren’t quite as native to Python, so we lack an operator for performing it and have to use np.matmul() twice as a result.

That is not true since 2015 with python 3.5: What is new in Python 5.5

Also the transpose can be used as .T instead of .transpose(). With this the above code becomes:

t_squared = (nx*ny)/(nx+ny) * delta.T @, np.linalg.inv(S_pooled)) @ delta

On the other hand both Python and R support unicode for variable names, just like Julia. So in both cases you could have used δ instead of delta.

gjanesch / data-and-the-world