การเขียนโปรแกรมแบบดั้งเดิมเทียบกับการเรียนรู้ของเครื่อง

What Is Traditional Programming?

Rules-Based Systems and Their Limits

# Traditional rules-based spam filter
def is_spam_traditional(email_text):
    spam_words = ['free', 'winner', 'click here', 'buy now']
    for word in spam_words:
        if word in email_text.lower():
            return True
    return False

print(is_spam_traditional('You are a WINNER! Click here for free stuff'))
# Output: True
print(is_spam_traditional('fr33 stuff for you!'))
# Output: False  -- fails on obfuscated spam

How Machine Learning Flips the Paradigm

# The ML paradigm shift
# Traditional: Input + Rules -> Output
# ML:          Input + Output -> Rules (learned automatically)

# Pseudocode conceptual example
from sklearn.naive_bayes import MultinomialNB
from sklearn.feature_extraction.text import CountVectorizer

emails = ['Free money now', 'Meeting at 3pm', 'Win a prize', 'Project update']
labels = [1, 0, 1, 0]  # 1=spam, 0=not spam

vectorizer = CountVectorizer()
X = vectorizer.fit_transform(emails)

model = MultinomialNB()
model.fit(X, labels)  # model learns the rules from data

Data as the New Source of Intelligence

When Traditional Programming Still Wins

Pattern Recognition: Where ML Shines

The Learning Process in Brief

# Conceptual training loop
import numpy as np

# Simulate a simple learning process
learning_rate = 0.1
weight = 0.0  # start with a random guess

for epoch in range(10):
    prediction = weight * 2.0
    target = 5.0
    error = target - prediction
    weight += learning_rate * error  # adjust based on error
    print(f'Epoch {epoch+1}: weight={weight:.3f}, error={error:.3f}')

Key Vocabulary: Model, Training, Inference

Generalisation: The True Goal

A Concrete Comparison Side by Side

# Traditional: hand-coded formula
def price_traditional(bedrooms, bathrooms, sqft):
    return 100000 + bedrooms * 50000 + bathrooms * 30000 + sqft * 150

# ML: coefficients learned from data
from sklearn.linear_model import LinearRegression
import numpy as np

# Training data (bedrooms, bathrooms, sqft)
X_train = np.array([[3, 2, 1500], [4, 3, 2000], [2, 1, 900]])
y_train = np.array([300000, 450000, 180000])

model = LinearRegression()
model.fit(X_train, y_train)  # learns coefficients from data
print('Learned coefficients:', model.coef_)

Why ML Is Transforming Every Industry

Quick Check

Lesson Recap