Simple AI Prediction Example

# Simple delivery time predictor
# This illustrates the full AI prediction cycle

# Step 1: Training data (distance_km, weight_kg) -> delivery_days
training_data = [
    ((10, 1), 1),    # 10km, 1kg -> 1 day
    ((50, 2), 2),    # 50km, 2kg -> 2 days
    ((100, 5), 3),   # 100km, 5kg -> 3 days
    ((200, 10), 5),  # 200km, 10kg -> 5 days
    ((500, 20), 8),  # 500km, 20kg -> 8 days
    ((30, 1), 1),
    ((80, 3), 2),
    ((150, 7), 4),
]

# Step 2: Find patterns (simplified)
# Pattern 1: days roughly = distance / 70
# Pattern 2: each 10kg adds about 1 day
# Real ML would find these automatically

def train_model(data):
    """Simplified training: find average relationships"""
    # In real ML, complex math finds the best formula
    # Here we use averages to illustrate
    distance_rates = []
    weight_rates = []
    
    for (distance, weight), days in data:
        distance_rates.append(days / distance)
        weight_rates.append(days / max(weight, 1))
    
    avg_distance_rate = sum(distance_rates) / len(distance_rates)
    avg_weight_rate = sum(weight_rates) / len(weight_rates)
    
    return avg_distance_rate, avg_weight_rate

dist_rate, weight_rate = train_model(training_data)

# Step 3: Make predictions
def predict_delivery_days(distance_km, weight_kg, dist_rate, weight_rate):
    days = (distance_km * dist_rate * 0.5) + (weight_kg * weight_rate * 0.5)
    return round(max(1, days))

# Step 4: Test on new deliveries
new_deliveries = [
    (75, 4),   # 75km, 4kg
    (300, 15), # 300km, 15kg
    (20, 0.5), # 20km, 0.5kg
]

print("Delivery Time Predictions:")
print("-" * 40)
for distance, weight in new_deliveries:
    days = predict_delivery_days(distance, weight, dist_rate, weight_rate)
    print(f"  {distance}km, {weight}kg -> Estimated {days} day(s)")