Design Patterns MCP Server

{ "id": "association-rule-mining", "name": "Association Rule Mining", "category": "Big Data Analysis", "description": "Discovers frequent patterns and relationships between items in large datasets using market basket analysis techniques", "when_to_use": "Market basket analysis\nRecommendation systems\nCross-selling opportunities\nFraud detection\nBiological pattern discovery", "benefits": "Uncovers hidden relationships\nActionable business insights\nScalable to large datasets\nUnsupervised learning approach", "drawbacks": "Computationally expensive\nMay generate too many rules\nRequires careful parameter tuning\nInterpretability challenges", "use_cases": "E-commerce recommendations\nHealthcare pattern analysis\nWeb usage mining\nSupply chain optimization\nGenomic research", "complexity": "High", "tags": [ "big-data", "unsupervised-learning", "pattern-mining", "market-basket-analysis", "frequent-itemsets" ], "examples": { "python": { "language": "python", "code": "from mlxtend.frequent_patterns import apriori, association_rules\nimport pandas as pd\n\ndef association_rule_mining(transactions, min_support=0.01, min_confidence=0.2):\n \"\"\"\n Mine association rules from transaction data\n \n Parameters:\n - transactions: DataFrame with binary columns (items) and rows (transactions)\n - min_support: Minimum support threshold\n - min_confidence: Minimum confidence threshold\n \"\"\"\n \n # Find frequent itemsets\n frequent_itemsets = apriori(transactions, min_support=min_support, use_colnames=True)\n \n # Generate association rules\n rules = association_rules(frequent_itemsets, metric='confidence', min_threshold=min_confidence)\n \n # Sort by lift (strength of association)\n rules = rules.sort_values('lift', ascending=False)\n \n return {\n 'frequent_itemsets': frequent_itemsets,\n 'rules': rules,\n 'summary': {\n 'total_rules': len(rules),\n 'avg_confidence': rules['confidence'].mean(),\n 'avg_lift': rules['lift'].mean()\n }\n }\n\n# Example usage\n# Sample transaction data\ndata = {\n 'bread': [1, 0, 1, 1, 0],\n 'milk': [1, 1, 0, 1, 1],\n 'butter': [0, 1, 1, 1, 0],\n 'cheese': [0, 0, 1, 1, 1],\n 'eggs': [1, 1, 0, 0, 1]\n}\n\ntransactions_df = pd.DataFrame(data)\n\n# Mine association rules\nresult = association_rule_mining(transactions_df, min_support=0.2, min_confidence=0.5)\n\nprint(f\"Found {result['summary']['total_rules']} association rules\")\nprint(\"Top rules by lift:\")\nprint(result['rules'][['antecedents', 'consequents', 'support', 'confidence', 'lift']].head())" } } }