Fast Algorithms for Association Rule Mining

1
Fast Algorithms for Association Rule Mining

• Muhammad Aurangzeb Ahmad
• Nupur Bhatnagar

2
Background Motivation

• Basket Data
• Collection of records consisting of transaction
  identifier and the items bought in a
  transaction.
• Mining for associations among items in a large
  database of sales transaction to predict the
  occurrence of an item based on the occurrences of
  other items in the transaction.
• For Example

3
Problem Definition

• INPUT
• A set of transactions

Output Assosiation rules having support minsup
threshold confidence minconf threshold

• Objective
• Find rules that will predict the occurrence of an
  item based on the occurrences of other items in
  the transaction

books ? Stationary,books, Bags ?
grocery,Coke,utensils, coke ? books,
4
Major Contribution

• Proposed two new algorithms for fast association
  rule mining
• Apriori and AprioriTID, along with a hybrid
  of the two algorithms .
• Empirical evaluations of the performance of the
  proposed algorithms as compared with the
  contemporary algorithms.

5
Key Concepts

• Itemset
• A collection of one or more items in a market
  based transactions.
• k-Itemset
• An itemset with k number of items is referred to
  as k-itemset.
• Support
• Given a rule X-gtY
• Faction of transactions that contain both X and
  Y.
• Support (X union Y )/ N where N is the total
  number of transactions.
• Confidence
• Given the rule X-gtY
• It measures how often item in Y appear in
  transactions that contain X.
• Confidence X union Y / X

6
Key Concepts

• Large Itemset
• Itemsets having support greater than minimum
  support and minimum confidence are called
• as large itemsets.
• Small Itemset
• Itemsets having support less than minimum
  support and minimum confidence are called as
  small itemsets
• Association Rule Mining
• An asociation rule is an implication of the form
  X-gtY where X and Y are the itemsets .
• Candidate Itemsets
• A set of itemsets which are generated from a
  seed of itemsets which were found to be large in
  the previous pass having
• support minsup threshold
• confidence minconf threshold

7
Key Concepts -Apriori

• Input
• The market base transaction dataset.
• Process
• Determine large 1-itemsets.
• Repeat until no new large 1-itemsets are
  identified.
• Generate (k1) length candidate itemsets from
  length k large itemsets.
• Prune candidate itemsets that are not large.
• Count the support of each candidate itemset.
• Eliminate candidate itemsets that are small.
• Output
• Itemsets that are large and qualify the min
  support and min confidence thresholds.

8
Key Concepts

• AprioriTID
• Same candidate generation function as Apriori.
• Does not use database for counting support after
  the first pass.
• Encoding of the candidate itemsets used in the
  previous pass.
• Saves reading effort
• Apriori Hybrid
• Apriori Hybrid uses Apriori in the initial
  passes and switches to AprioriTid when it expects
  that the candidate itemsets at the end of the
  pass will be in memory

9
Validation Methodology-Synthetic data

• Generated synthetic data sets involving
  transactions to evaluate the performance of
  algorithms.
• Each itemset in a transaction has a weight
  associated with it, which corresponds to the
  probability of the itemset to be picked.

10
Validation Methodology-Weakness and Strength

• Strength
• Author use a substantial basket data for guiding
  the process of designing fast algorithms for
  association rule mining.
• Weakness
• Synthetic data set is used for validation. The
  data might be too synthetic as to not give any
  valuable information about real world datasets.

11
Assumptions

• Synthetic dataset is used.It is assumed that
  performance of the algorithm in the synthetic
  dataset is indicative of its performance on a
  real world dataset.
• All the items in the data are in a
  lexicographical order.
• It is assumed that all the data is present in
  the same site or table and there are no cases
  which there would be a requirement to make joins.

12
Possible Revision

• Some real world datasets should be used to
  perform the experiments .
• The number of large itemsets could exponentially
  increase with large databases. Modification in
  the representation structure is required that
  captures just a subset of the candidate large
  itemsets.

13
Questions?