Day 42: KNN & Decision Trees

import numpy as np
import pandas as pd
st = pd.read_csv("students_performance.csv")
st["risk"] = (st["average"] < 60).astype(int)
new = np.array([1.8, 72])                               # study hours, attendance
d = np.sqrt(((st[["study_hours", "attendance"]].to_numpy() - new) ** 2).sum(axis=1))
nearest = st.assign(dist=d).nsmallest(5, "dist")[["student_id", "study_hours", "attendance", "average", "risk", "dist"]]
print(nearest.round(2).to_string(index=False))
print("Vote → at risk" if nearest["risk"].mean() > 0.5 else "Vote → not at risk")

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import make_pipeline
st = pd.read_csv("students_performance.csv")
X = st[["study_hours", "attendance"]]; y = (st["average"] < 60).astype(int)
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=0.3, random_state=0, stratify=y)
print(" k   raw   scaled")
for k in [1, 3, 7, 15, 31, 61]:
    raw = KNeighborsClassifier(n_neighbors=k).fit(X_tr, y_tr).score(X_te, y_te)
    sc = make_pipeline(StandardScaler(), KNeighborsClassifier(n_neighbors=k)).fit(X_tr, y_tr).score(X_te, y_te)
    print(f"{k:2d}  {raw:.3f}  {sc:.3f}")

import pandas as pd
import matplotlib.pyplot as plt
from sklearn.tree import DecisionTreeClassifier, plot_tree, export_text
st = pd.read_csv("students_performance.csv")
X = st[["study_hours", "attendance"]].assign(internet=(st["internet"] == "Yes").astype(int))
y = (st["average"] < 60).astype(int)
tree = DecisionTreeClassifier(max_depth=3, min_samples_leaf=15, random_state=0).fit(X, y)
print(export_text(tree, feature_names=list(X.columns), class_names=["ok", "at_risk"]))
plt.figure(figsize=(12, 5))
plot_tree(tree, feature_names=X.columns, class_names=["ok", "at risk"], filled=True, rounded=True, fontsize=8, impurity=False, proportion=True)
plt.title("Each box is a question; leaves give the share of at-risk students"); plt.show()

import numpy as np
import pandas as pd
st = pd.read_csv("students_performance.csv"); y = (st["average"] < 60).astype(int)
gini = lambda s: 1 - (s.mean() ** 2 + (1 - s.mean()) ** 2)
print(f"Gini of all students: {gini(y):.3f}")
for cut in [1.5, 2.0, 2.5, 3.0]:
    left, right = y[st["study_hours"] <= cut], y[st["study_hours"] > cut]
    w = (len(left) * gini(left) + len(right) * gini(right)) / len(y)
    print(f"split study_hours ≤ {cut}: left risk {left.mean():.2f}, right risk {right.mean():.2f}, weighted Gini {w:.3f}")

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
l = pd.read_csv("loan_applications.csv"); l["lti"] = l["loan_amount"] / (l["monthly_income"] * 12)
X = l[["credit_score", "lti", "years_employed", "monthly_income"]]; y = l["approved"]
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=0.25, random_state=42, stratify=y)
for leaf in [1, 5, 20, 50]:
    t = DecisionTreeClassifier(min_samples_leaf=leaf, random_state=0).fit(X_tr, y_tr)
    print(f"min_samples_leaf={leaf:2d}  leaves={t.get_n_leaves():3d}  train={t.score(X_tr, y_tr):.3f}  test={t.score(X_te, y_te):.3f}")

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import make_pipeline
l = pd.read_csv("loan_applications.csv"); l["lti"] = l["loan_amount"] / (l["monthly_income"] * 12)
X = l[["credit_score", "lti"]]; y = l["approved"]
xx, yy = np.meshgrid(np.linspace(480, 900, 220), np.linspace(0, 3, 220))
grid = pd.DataFrame({"credit_score": xx.ravel(), "lti": yy.ravel()})
models = [("Logistic", make_pipeline(StandardScaler(), LogisticRegression())),
          ("KNN k=15", make_pipeline(StandardScaler(), KNeighborsClassifier(15))),
          ("Tree depth 4", DecisionTreeClassifier(max_depth=4, random_state=0))]
fig, ax = plt.subplots(1, 3, figsize=(13, 3.8), sharey=True)
for a, (name, m) in zip(ax, models):
    z = m.fit(X, y).predict(grid).reshape(xx.shape)
    a.contourf(xx, yy, z, alpha=0.35, cmap="RdYlGn"); a.scatter(X["credit_score"], X["lti"], c=y, cmap="RdYlGn", s=5)
    a.set_title(f"{name} (train acc {m.score(X, y):.2f})"); a.set_xlabel("credit score"); a.set_ylim(0, 3)
ax[0].set_ylabel("loan ÷ yearly income"); plt.tight_layout(); plt.show()

import pandas as pd
from sklearn.tree import DecisionTreeClassifier
l = pd.read_csv("loan_applications.csv"); l["lti"] = l["loan_amount"] / (l["monthly_income"] * 12)
l["salaried"] = (l["employment"] == "Salaried").astype(int)
feats = ["credit_score", "lti", "years_employed", "monthly_income", "loan_amount", "salaried"]
t = DecisionTreeClassifier(max_depth=5, min_samples_leaf=20, random_state=0).fit(l[feats], l["approved"])
print(pd.Series(t.feature_importances_, index=feats).round(3).sort_values(ascending=False).to_string())