rule-mining-algs 0.1.1

The package contains a few selected association rule mining algorithms.

Repository for Assocation Rule Mining Algorithms

Several algorithms for mining association rules have been implemented in this repository.

NOTE: The algorithms are implemented in pure Python which makes them rather slow.

Installation

pip install rule-mining-algs

Different Algorithms thus far

• AIS
• Apriori
• FP-Growth
• h-Clique (all-confidence pushed into Apriori-algorithm)
• Quantitative Association Rules
• AClose
• Minimal Non-redundant rules
• Rule Generation for apriori-like Algorithms
• Brute-Force Mining of Classification Rules
• Clustering to find intervals for numeric attributes
• Evolutionary Algorithm to Discover Itemsets (GAR)
• Evolutionary Algorithm to Discover Rules with fixed consequent (GAR-PLUS)

Datasets

• store_data.csv: Store data
• agaricus-lepiota.data: Mushroom dataset

Build and Run models

• Transformers: static_discretization(equi-depth/width partitioning of numeric attributes), cluster_interval_data(birch clustering to find intervals for numeric attributes)
• Itemset_Miners: see Algorithms section
• Rule_Miners: generate_rules(Standard algorithm to generate rules from itemsets), min_redundant_rules(only usable with a_close itemset miner)

---

Example for a dataset with a single attribute

from algs.models import StandardMiner
from algs.data import load_store_data

data_df = load_store_data() # Load store dataset
# Choose model
m = StandardMiner()
# Set parameters for the algorithms
m.set_args(m.itemset_miner, {"min_support": 0.005})
m.set_args(m.rule_miner, {"min_conf": 0.5})
# Run the algorithm on the dataset
output = m.run(data_df)

Example for a DB containing several categorical attributes

from algs.data import load_shroom_data
from algs.quantitative import static_discretization
from algs.rule_gen import get_classification_rules

shrooms = load_shroom_data()
mine_quant = StandardMiner(static_discretization)
names = {name: 0 for name in shrooms.columns}
# Set arguments for transformer, itemset and rule miner
mine_quant.set_args(mine_quant.transformer, {"discretization": names})
mine_quant.set_args(mine_quant.itemset_miner, {"min_support": 0.15})
mine_quant.set_args(mine_quant.rule_miner, {"min_conf": 0.65})
rules = mine_quant.run(shrooms)
# Post processing step to obtain rules having only the label in the consequent
classification_rules = get_classification_rules(rules, "label")