Market Basket Analysis BySowjanya Alaparthi

Topics to be discussedIntroduction to Market basket analysisApriori AlgorithmDemo-1 ( Using self created table)Demo-2 ( Using Oracle sample schema)Demo-3 ( Using OLAP analytic workspace)

Introduction to Market Basket AnalysisDef: Market Basket Analysis (Association Analysis) is a mathematical modeling technique based upon the theory that if you buy a certain group of items, you are likely to buy another group of items. It is used to analyze the customer purchasing behavior and helps in increasing the sales and maintain inventory by focusing on the point of sale transaction data.Given a dataset, the Apriori Algorithm trains and identifies product baskets and product association rules

Definitions and TerminologyTransaction is a set of items (Itemset). Confidence : It is the measure of uncertainty or trust worthiness associated with each discovered pattern.Support : It is the measure of how often the collection of items in an association occur together as percentage of all transactionsFrequent itemset : If an itemset satisfies minimum support,then it is a frequent itemset. Strong Association rules: Rules that satisfy both a minimum support threshold and a minimum confidence threshold In Association rule mining, we first find all frequent itemsets and then generate strong association rules from the frequent itemsets

Definitions and Terminology- ContinuedApriori algorithm is the most established algorithm for finding frequent item sets mining.The basic principle of Apriori is Any subset of a frequent itemset must be frequent.We use these frequent itemsets to generate association rules.

Apriori AlgorithmCk: Candidate itemset of size kLk: Frequent itemset of size kL1={frequent items};For (k=1; Lk!=0; k++) do beginCk+1= Candidates generated from Lk;For each transaction t in the database doIncrement the count of all candidates in Ck+1 that are contained in tLk+1=candidates in Ck+1 with min_supportEndReturn UkLk;

Pictorial representation of Apriori algorithm

Step 1Scan the transaction database to get the support S of each 1-itemsets, compare S with min _sup, and get a set of frequent 1- itemsets,L1

Step 2Use Lk-1 join Lk-1 to generate a candidate k-itemsets. And use appiori propersty to prune the unfrequented k-itemssets from this set.

Step 3Scan the transaction database to get the support S of each candidate k-itemset in the final set, compare S with min_sup, and get a set of frequent k-itemsets,Lk

Step 4The candidate set = null

Step 5For each frequent itemset 1, generate all non empty subsets of 1

Step 6For every non empty subset s of 1, output the rule s =>(1-s) if confidence C of the rule s=> (1-s)(=support S of 1/support S of s )3

NO

YES

Demo-1Installations Oracle 10g enterprise edition SQL Plus Oracle Data Miner Client

Demo 1 - Data PreparationDownload the sample data, which is in excel sheet.write macro to convert data in excel sheet to insert queriesCreate a table and execute these insert queries in SQLplusAs we are connected to Oracle server, this table is then found in Oracle database

Demo-1 ConnectionsConnect Oracle Data Miner Client to Oracle Database Make sure the oracle listener is listeningDatabase instance ora478 is started. The port used is 1521 Give the hostname as oracle.itk.ilstu.edu

Demo-1Perform the activity, after installations and connections are made.

Demo-2 ( using oracle sample schema)Download Oracle 10g on your system and install itSelect the sample schema option during the custom installationLaunch Oracle Data Miner ClientIn order to use this sample scheme for our activity, we should have the system administrator privileges.The username is SH and password is password

Demo -2Administrator should perform some grants in sqlplusw to build this activity. They are alter user sh account unlock; alter user sh identified by password; grant create table to sh; grant create sequence to sh; grant create session to sh; grant create view to sh; grant create procedure to sh; grant create job to sh; grant create type to sh; grant create synonym to sh; grant execute on ctxsys.ctx_ddl to sh;

Demo-2The points to be noted before starting the activity are:Make sure the oracle listener is startedDatabase instance ORCL is started. The port used is 1521 Give the hostname as 127.0.0.1, which is a general hostname.

Demo-2Finally, the results from the model are published to a table, and this table forms the raw source for the new OLAP product dimension.At this point there is no information relating to revenue, costs or quantity. So, we need to extend the activity beyond association analysis to OLAP.

OLAPWe have to correctly format the results obtained from Association analysis for dimension mapping in OLAP. This can be done using OLAP DML or PL/SQL. In our activity we create a separate dimension that can hold the results from algorithm. For each dimension we can create Levels, hierarchies, attributes and mappings.

OLAP- Analytic workspaceLaunch Analytic workspace and give the login details as Username- sh Connection information- 127.0.0.1:1521:orcl This connects to Oracle sample schema SH on 1521 port and local host 127.0.0.1 and orcl database instance.

Demo 3- OLAP Analytic WorkspacePerform the activity and show the mappings

ConclusionWe have shown how Market basket analysis using association rules works in determining the customer buying patterns. This can be further extended using OLAP Analytic workspace as shown in demo-3, to add dimensions and cube to identify other measures like costs, revenue and quantity.

ReferencesBooks: [1] Michael J. A. Berry, Gordon Linoff.Data Mining Techniques: For Marketing, Sales, and Customer Support (Paperback).[2] J.Han,M. Kamber(2001) Data Mining, Morgan Kaufmann publishers, San Francisco,CA

Links:[3]. http://oraclebi.blogspot.com/2007/02/using-market-basket-analysis-to-add.html[4]. http://nymetro.chapter.informs.org/prac_cor_pubs/Ausleder-On-market-basket-analysis-May-04.pdf[5]. http://www2.sas.com/proceedings/sugi28/223-28.pdf[6] http://en.wikipedia.org/wiki/Market_basket_analysis[7] http://www.cs.ualberta.ca/~zaine/courses/cmput499/slides/lect10/sld053.htm[8] http://www.icaen.uiowa.edu/~comp/Public/Apriori.pdf

Questions??