Interactive Frequent Itemset Mining

✨ NEW: Clean, organized structure with interactive terminal interface!
All results displayed in terminal - no file clutter!

A comprehensive data mining project implementing three algorithms for frequent itemset mining and association rule discovery.

🎯 Features

• Three Mining Algorithms:

  • Brute Force (custom implementation)
  • Apriori (mlxtend library)
  • FP-Growth (mlxtend library)

• Five Transactional Databases:

  • Amazon (Technology)
  • BestBuy (Electronics)
  • Walmart (Groceries)
  • Target (Clothing)
  • Costco (Household)

• Interactive Interface:

  • User-friendly menu system
  • Real-time results display
  • Configurable parameters
  • No file clutter (results shown in terminal)

📁 Project Structure

Data mining/
├── main.py              # Interactive interface (START HERE)
├── data/                # Transaction databases (CSV files)
│   ├── Amazon_transactions.csv
│   ├── BestBuy_transactions.csv
│   ├── Walmart_transactions.csv
│   ├── Target_transactions.csv
│   └── Costco_transactions.csv
├── src/                 # Source code
│   ├── generate_databases.py
│   ├── brute_force_mining.py
│   ├── library_based_mining.py
│   └── run_brute_force.py
└── docs/                # Documentation
    ├── README.md
    ├── DATA_CREATION_REPORT.md
    ├── BRUTE_FORCE_ALGORITHM_REPORT.md
    ├── ALGORITHM_COMPARISON_REPORT.md
    └── ... (other reports)

📖 Getting Started

👉 See GETTING_STARTED.md for detailed walkthrough with examples!

🚀 Quick Start

Prerequisites

pip install mlxtend pandas numpy

Run the Interactive Tool

python main.py

What You'll See

1. Welcome Screen - Introduction to the tool
2. Database Selection - Choose from 5 databases
3. Algorithm Selection - Pick Brute Force, Apriori, or FP-Growth
4. Configuration - Set minimum support and confidence
5. Results Display - View frequent itemsets and association rules in terminal
6. Repeat or Exit - Run another analysis or quit

💡 Example Usage

SELECT DATABASE
1. Amazon      - Technology & Electronics
2. BestBuy     - Consumer Electronics
3. Walmart     - Groceries
4. Target      - Clothing & Fashion
5. Costco      - Household Items

Enter database number: 1

SELECT ALGORITHM
1. Brute Force
2. Apriori
3. FP-Growth

Enter algorithm number: 2

CONFIGURATION
Minimum support (0-1, default 0.2): 0.2
Minimum confidence (0-1, default 0.6): 0.6

[Results displayed in terminal...]

📊 Sample Output

FREQUENT ITEMSETS
  Total frequent itemsets found: 18

  📦 1-Itemsets (12 found):
     {HDMI_Cable} - Count: 10, Support: 0.4000
     {Router} - Count: 10, Support: 0.4000
     {USB_Cable} - Count: 10, Support: 0.4000
     ...

ASSOCIATION RULES
  Total rules generated: 15

  Top 15 rules:
  Rule                                                Supp     Conf     Lift
  {Laptop} → {Keyboard}                             0.2000   1.0000   5.0000
  {Mouse} → {Keyboard}                              0.2000   1.0000   5.0000
  ...

🔧 Advanced Usage

Regenerate Databases

cd src
python generate_databases.py

This creates new deterministic transaction data in the data/ folder.

Run Individual Algorithms (Non-Interactive)

cd src

# Brute Force only
python run_brute_force.py

# Apriori + FP-Growth
python library_based_mining.py

📈 Parameters

Minimum Support

• Range: 0.0 to 1.0 (percentage) or integer (absolute count)
• Default: 0.2 (20%)
• Effect: Lower = more itemsets found, slower execution

Minimum Confidence

• Range: 0.0 to 1.0
• Default: 0.6 (60%)
• Effect: Lower = more rules generated

🎓 Educational Value

Brute Force

• Shows exhaustive search approach
• Demonstrates exponential complexity
• Educational baseline for comparison

Apriori

• Illustrates pruning optimization
• Uses downward closure property
• Standard industry algorithm

FP-Growth

• Tree-based pattern growth
• Most efficient for large datasets
• Modern mining approach

📚 Documentation

Detailed reports are available in the docs/ folder:

• DATA_CREATION_REPORT.md - How databases were created
• BRUTE_FORCE_ALGORITHM_REPORT.md - Brute force implementation details
• ALGORITHM_COMPARISON_REPORT.md - Performance comparison of all algorithms
• QUICK_START_GUIDE.md - Detailed usage instructions

🏆 Key Results

All three algorithms produce identical results (verified):

• ✅ Same frequent itemsets
• ✅ Same association rules
• ✅ Same support/confidence/lift values

Performance (all 5 databases)

• Brute Force: 0.050s
• Apriori: 0.013s (3.85x faster)
• FP-Growth: 0.017s (2.94x faster)

🔍 Common Use Cases

Product Bundling

Find items frequently bought together to create bundles.

Store Layout

Place related products near each other based on co-occurrence.

Recommendation Systems

Suggest items based on shopping cart contents.

Inventory Management

Predict demand for complementary products.

🐛 Troubleshooting

"mlxtend not found"

pip install mlxtend

"File not found" error

Make sure you're running main.py from the project root directory.

Database files missing

Run python src/generate_databases.py to recreate them.

📝 License

Educational project for data mining coursework.

👤 Author

Data Mining Project - Frequent Itemset Mining Implementation

---

Start mining: python main.py 🚀