Quasi-randomized nnetworks in Julia, Python and R

Posted on November 27, 2023 by T. Moudiki in Data science | 0 Comments

This article was first published on T. Moudiki's Webpage - Python , and kindly contributed to python-bloggers. (You can report issue about the content on this page here)
Want to share your content on python-bloggers? click here.

nnetsauce

nnetsauce, a package for quasi-randomized supervised learning (classification and regression), is currently available for R and Python. For more details on

nnetsauce

nnetsauce, you can read these posts.

I’ve always wanted to port

nnetsauce

nnetsauce to the Julia language. However, in the past few years, there was a little timing overhead (more precisely, a lag) when I tried to do that with Julia’s

PyCall

PyCall, based on my Python source code. This overhead seems to have ‘disappeared’ now.

Julia language’s

nnetsauce

nnetsauce is not a package yet, but you can already use

nnetsauce

nnetsauce in Julia.

Here’s how I did it on Ubuntu Linux:

Contents

1 – Install Julia

Run (terminal):

wget https://julialang-s3.julialang.org/bin/linux/x64/1.9/julia-1.9.4-linux-x86_64.tar.gz

wget https://julialang-s3.julialang.org/bin/linux/x64/1.9/julia-1.9.4-linux-x86_64.tar.gz

Run (terminal):

tar zxvf julia-1.9.4-linux-x86_64.tar.gz

tar zxvf julia-1.9.4-linux-x86_64.tar.gz

Run (terminal)(This is VSCode, but use your favorite editor here):

code ~/.bashrc

code ~/.bashrc

Add to

.bashrc

.bashrc (last line):

export PATH="$PATH:julia-1.9.4/bin"

export PATH="$PATH:julia-1.9.4/bin"

Run (terminal):

source ~/.bashrc

source ~/.bashrc

Run (terminal):

julia nnetsauce_example.jl

julia nnetsauce_example.jl

2 – Example using a nnetsauce classifier in Julia language

For Python user’s, notice that this is basically Python ^^

using Pkg

ENV["PYTHON"] = "" # replace with your Python path

Pkg.add("PyCall")

Pkg.build("PyCall")

Pkg.add("Conda")

Pkg.build("Conda")

using PyCall

using Conda

Conda.add("pip") # Ensure pip is installed

Conda.pip_interop(true) # Enable pip interop

Conda.pip("install", "scikit-learn") # Install scikit-learn

Conda.pip("install", "jax") # /!\ Only on Linux or macOS: Install jax

Conda.pip("install", "jaxlib") # /!\ Only on Linux or macOS: Install jaxlib

Conda.pip("install", "nnetsauce") # Install nnetsauce

Conda.add("numpy")

np = pyimport("numpy")

ns = pyimport("nnetsauce")

sklearn = pyimport("sklearn")

# 1 - breast cancer dataset

dataset = sklearn.datasets.load_breast_cancer()

X = dataset["data"]

y = dataset["target"]

X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X, y,

test_size=0.2, random_state=123)

clf = ns.Ridge2MultitaskClassifier(n_hidden_features=9, dropout=0.43, n_clusters=1,

lambda1=1.24023438e+01, lambda2=7.30263672e+03)

@time clf.fit(X=X_train, y=y_train) # timing?

print("\n\n Model parameters: \n\n")

print(clf.get_params())

print("\n\n Testing score: \n\n") # Classifier's accuracy

print(clf.score(X_test, y_test)) # Must be: 0.9824561403508771

print("\n\n")

using Pkg ENV["PYTHON"] = "" # replace with your Python path Pkg.add("PyCall") Pkg.build("PyCall") Pkg.add("Conda") Pkg.build("Conda") using PyCall using Conda Conda.add("pip") # Ensure pip is installed Conda.pip_interop(true) # Enable pip interop Conda.pip("install", "scikit-learn") # Install scikit-learn Conda.pip("install", "jax") # /!\ Only on Linux or macOS: Install jax Conda.pip("install", "jaxlib") # /!\ Only on Linux or macOS: Install jaxlib Conda.pip("install", "nnetsauce") # Install nnetsauce Conda.add("numpy") np = pyimport("numpy") ns = pyimport("nnetsauce") sklearn = pyimport("sklearn") # 1 - breast cancer dataset dataset = sklearn.datasets.load_breast_cancer() X = dataset["data"] y = dataset["target"] X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X, y, test_size=0.2, random_state=123) clf = ns.Ridge2MultitaskClassifier(n_hidden_features=9, dropout=0.43, n_clusters=1, lambda1=1.24023438e+01, lambda2=7.30263672e+03) @time clf.fit(X=X_train, y=y_train) # timing? print("\n\n Model parameters: \n\n") print(clf.get_params()) print("\n\n Testing score: \n\n") # Classifier's accuracy print(clf.score(X_test, y_test)) # Must be: 0.9824561403508771 print("\n\n")

using Pkg
ENV["PYTHON"] = ""  # replace with your Python path
Pkg.add("PyCall")
Pkg.build("PyCall")
Pkg.add("Conda")
Pkg.build("Conda")

using PyCall
using Conda

Conda.add("pip")  # Ensure pip is installed
Conda.pip_interop(true)  # Enable pip interop
Conda.pip("install", "scikit-learn")  # Install scikit-learn
Conda.pip("install", "jax")  # /!\ Only on Linux or macOS: Install jax
Conda.pip("install", "jaxlib")  # /!\ Only on Linux or macOS: Install jaxlib
Conda.pip("install", "nnetsauce")  # Install nnetsauce
Conda.add("numpy")

np = pyimport("numpy")
ns = pyimport("nnetsauce")
sklearn = pyimport("sklearn")


# 1 - breast cancer dataset

dataset = sklearn.datasets.load_breast_cancer()

X = dataset["data"]
y = dataset["target"]

X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X, y, 
test_size=0.2, random_state=123)

clf = ns.Ridge2MultitaskClassifier(n_hidden_features=9, dropout=0.43, n_clusters=1, 
lambda1=1.24023438e+01, lambda2=7.30263672e+03)

@time clf.fit(X=X_train, y=y_train) # timing?

print("\n\n Model parameters: \n\n")
print(clf.get_params())

print("\n\n Testing score: \n\n") # Classifier's accuracy
print(clf.score(X_test, y_test)) # Must be: 0.9824561403508771
print("\n\n")

# 2 - wine dataset

dataset = sklearn.datasets.load_wine()

X = dataset["data"]

y = dataset["target"]

X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X, y,

test_size=0.2, random_state=123)

clf = ns.Ridge2MultitaskClassifier(n_hidden_features=15,

dropout=0.1, n_clusters=3,

type_clust="gmm")

@time clf.fit(X=X_train, y=y_train) # timing?

print("\n\n Model parameters: \n\n")

print(clf.get_params())

print("\n\n Testing score: \n\n") # Classifier's accuracy

print(clf.score(X_test, y_test)) # Must be 1.0

print("\n\n")

# 2 - wine dataset dataset = sklearn.datasets.load_wine() X = dataset["data"] y = dataset["target"] X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X, y, test_size=0.2, random_state=123) clf = ns.Ridge2MultitaskClassifier(n_hidden_features=15, dropout=0.1, n_clusters=3, type_clust="gmm") @time clf.fit(X=X_train, y=y_train) # timing? print("\n\n Model parameters: \n\n") print(clf.get_params()) print("\n\n Testing score: \n\n") # Classifier's accuracy print(clf.score(X_test, y_test)) # Must be 1.0 print("\n\n")

# 2 - wine dataset

dataset = sklearn.datasets.load_wine()

X = dataset["data"]
y = dataset["target"]

X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X, y, 
test_size=0.2, random_state=123)

clf = ns.Ridge2MultitaskClassifier(n_hidden_features=15,
dropout=0.1, n_clusters=3, 
type_clust="gmm")

@time clf.fit(X=X_train, y=y_train) # timing?

print("\n\n Model parameters: \n\n")
print(clf.get_params())

print("\n\n Testing score: \n\n") # Classifier's accuracy
print(clf.score(X_test, y_test)) # Must be 1.0
print("\n\n")

To leave a comment for the author, please follow the link and comment on their blog: T. Moudiki's Webpage - Python .

Want to share your content on python-bloggers? click here.

Python-bloggers

Data science news and tutorials - contributed by Python bloggers

Quasi-randomized nnetworks in Julia, Python and R

1 – Install Julia

2 – Example using a nnetsauce classifier in Julia language

Related