Machine Learning

This is a repository for some of the Machine Learning projects I have worked on.

My Beloved, the Iris Dataset

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay
from sklearn.model_selection import train_test_split, GridSearchCV

df = sns.load_dataset("iris")
y = df.iloc[:, -1].copy().to_numpy()
X = df.iloc[:, :-1].copy().to_numpy()
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3)
params = {'max_depth': list(range(3,12))}
model = RandomForestClassifier(oob_score=True)
gs = GridSearchCV(model, param_grid=params, scoring='accuracy', n_jobs=-1)
gs.fit(X_train, y_train)
clf = gs.best_estimator_

print(f"Score (Train): {clf.score(X_train, y_train):.4f}")
print(f"Score (Test): {clf.score(X_test, y_test)}")
print(f"OOB Score: {clf.oob_score_:.4f}")

cm = confusion_matrix(y_test, clf.predict(X_test), normalize="true")
disp_cm = ConfusionMatrixDisplay(cm, display_labels=clf.classes_)
disp_cm.plot()
plt.show()

importances = pd.DataFrame(clf.feature_importances_, index=df.columns[:-1])
importances.plot.bar()
plt.tight_layout()
plt.show()