Little Bits of Artificial Intelligence

Pure Python Code: Unsupervised Learning (Simple K-Means)

import random
# Step 1: Create fake 2D data points
points = [(random.randint(0, 20), random.randint(0, 20)) for _ in range(10)]

# Step 2: Pick 2 random points as initial cluster centers
centroids = random.sample(points, 2)

def distance(p1, p2):
    # Euclidean distance
    return ((p1[0] - p2[0]) ** 2 + (p1[1] - p2[1]) ** 2) ** 0.5

def assign_clusters(points, centroids):
    clusters = {i: [] for i in range(len(centroids))}
    for point in points:
        distances = [distance(point, c) for c in centroids]
        closest = distances.index(min(distances))
        clusters[closest].append(point)
    return clusters

def update_centroids(clusters):
    new_centroids = []
    for points in clusters.values():
        x_coords = [p[0] for p in points]
        y_coords = [p[1] for p in points]
        avg_x = sum(x_coords) / len(points)
        avg_y = sum(y_coords) / len(points)
        new_centroids.append((avg_x, avg_y))
    return new_centroids

# Step 3: Run K-Means algorithm manually
for i in range(5):  # Repeat 5 times
    clusters = assign_clusters(points, centroids)
    centroids = update_centroids(clusters)

# Step 4: Print the final groups
for i, group in clusters.items():
    print(f"\nGroup {i+1}:")
    for point in group:
        print(point)

What’s Happening Here:

1. Random points are created.
2. We choose 2 random centers (you can change to 3 or more).
3. The algorithm:

• Assigns each point to the closest center
• Updates the center to be the “average” of points near it
• Repeats this 5 times to “learn” better groupings

Output Example:

Group 1:
(3, 2)
(2, 1)
(4, 3)

Group 2:
(15, 18)
(14, 17)
(16, 19)

Even though it’s very basic, this shows how unsupervised learning works without any help — the computer figures out groups on its own.

We’ll draw a simple grid, place the points on it, and show their cluster group using different symbols (like X, O, *, etc.).

Unsupervised Learning with ASCII Visualization

import random

# Grid size (like a board of 20x20)
GRID_SIZE = 20

# Step 1: Create random 2D points
points = [(random.randint(0, GRID_SIZE-1), random.randint(0, GRID_SIZE-1)) for _ in range(10)]

# Step 2: Choose 2 random points as cluster centers
centroids = random.sample(points, 2)

def distance(p1, p2):
    return ((p1[0]-p2[0])**2 + (p1[1]-p2[1])**2) ** 0.5

def assign_clusters(points, centroids):
    clusters = {i: [] for i in range(len(centroids))}
    for point in points:
        dists = [distance(point, c) for c in centroids]
        closest = dists.index(min(dists))
        clusters[closest].append(point)
    return clusters
def update_centroids(clusters):
    new_centroids = []
    for pts in clusters.values():
        avg_x = sum(p[0] for p in pts) / len(pts)
        avg_y = sum(p[1] for p in pts) / len(pts)
        new_centroids.append((avg_x, avg_y))
    return new_centroids

# Run clustering steps
for _ in range(5):
    clusters = assign_clusters(points, centroids)
    centroids = update_centroids(clusters)

# Cluster symbols for visualization
symbols = ['X', 'O', '*', '#']

# Draw the grid with clusters
print("\n Clustering Grid View:\n")
for y in range(GRID_SIZE):
    row = ""
    for x in range(GRID_SIZE):
        point = (x, y)
        printed = False
        for idx, group in clusters.items():
            if point in group:
                row += symbols[idx]  # Use symbol for the cluster
                printed = True
                break
        if not printed:
            row += '.'  # Empty space
    print(row)

# Optional: print group members
print("\n Cluster Groups:")
for i, group in clusters.items():
    print(f"Group {i+1} ({symbols[i]}): {group}")

What we’ll See:

A 20×20 grid, with:

• Points marked as X, O, etc. (based on which group they belong to)
• Empty cells as .

A list showing which points are in which group

Sample Output (snipped):

………………..
………………..
……….X………
………………..
…..O…………..
…..O…………..
………………..
………………..
………………..
……X………….
………………..
………………..
…….X…………
………………..
………………..
………………..
………………..
………………..
………………..
………………..

Cluster Groups:
Group 1 (X): [(10, 2), (6, 9), (7, 12)]
Group 2 (O): [(5, 4), (5, 5)]

What It Teaches:

• The computer groups points by closeness, without knowing the groups in advance.
• The ASCII grid gives a visual feeling of clustering without using graphics.
• Kids (or beginners) can see learning in action — like coloring similar things together on paper.

Unsupervised Learning – Brainstorming Session