Day 5 — ML Fundamentals | AI Engineering Internship

0–5 min
Real Hook

5–25 min
Core Concept

25–45 min
Live Coding

45–55 min
You Code It

55–60 min
Cheat Sheet

🔥 Real-World Hook (0–5 min)

MagicBricks and 99acres don't have humans pricing every listing. A regression model looks at area, location, bedrooms, age — and outputs a price estimate in milliseconds.

The same math Gauss used in 1809 to predict asteroid orbits. Today it prices 10 million homes per day. That's ML in production.

📈 The ML Workflow (5–25 min)

💡 Every ML Project Has 5 Steps

1. Load data — CSV, database, API.
2. Split — 80% train, 20% test. Never evaluate on training data.
3. Train — model.fit(X_train, y_train).
4. Evaluate — RMSE, R² on the test set.
5. Predict — model.predict(new_data).

⚠ Overfitting

A model that memorises training data but fails on new data. Like a student who copies answers for last year's exam but can't solve new problems. Always check test score vs train score.

💻 Live Demo — House Price Predictor (25–45 min)

>import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error, r2_score

# Synthetic house price dataset
np.random.seed(42)
n = 200
area    = np.random.randint(600, 3000, n)           # sq ft
rooms   = np.random.randint(1, 6, n)
age     = np.random.randint(0, 30, n)              # years
price   = 2000*area + 50000*rooms - 3000*age + np.random.normal(0, 80000, n)

df = pd.DataFrame({"area": area, "rooms": rooms,
                      "age": age,  "price": price})

# 1. Split
X = df[["area", "rooms", "age"]]
y = df["price"]
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# 2. Train
model = LinearRegression()
model.fit(X_train, y_train)

# 3. Evaluate
y_pred = model.predict(X_test)
rmse   = mean_squared_error(y_test, y_pred) ** 0.5
r2     = r2_score(y_test, y_pred)
print(f"RMSE: ₹{rmse:,.0f}  R²: {r2:.3f}" )  # R² close to 1.0 is good

# 4. Predict a new house
new_house = [[1800, 3, 5]]           # 1800 sqft, 3 rooms, 5 yr old
est_price = model.predict(new_house)[0]
print(f"Estimated price: ₹{est_price:,.0f}" )

📋 You Code It (45–55 min)

✍ Task

Add a location_score feature (1–5) to the dataset. Re-train. Does R² improve? Print the model's coefficients with model.coef_ to see which feature matters most.

⚡ Day 5 Cheat Sheet

train_test_split(X, y, test_size=0.2)

80/20 split

model.fit(X_train, y_train)

Train the model

model.predict(X_test)

Generate predictions

r2_score(y_test, y_pred)

R²: 1.0 = perfect

mean_squared_error(...)**0.5

RMSE: error in same units

model.coef_

Feature weights

model.intercept_

Bias term

StandardScaler().fit_transform(X)

Normalize features

← Day 04 Day 06 · Classification →