mgcharm-150527055232-lva1-app6891

24
MG-CHARM ANURAG SAXENA by Minimal generator close hierarchical association rule mining

Upload: anurag-saxena

Post on 12-Apr-2017

245 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: mgcharm-150527055232-lva1-app6891

MG-CHARM

ANURAG SAXENA

by

Minimal generator close hierarchical association rule mining

Page 2: mgcharm-150527055232-lva1-app6891

Contents Abstract

Introduction

Scope

Existing System

Proposed System

Algorithm

Example

Design

Functional Diagram

Structural Diagram

Screens

Conclusion

References

Page 3: mgcharm-150527055232-lva1-app6891

Abstract In Data Mining, Frequent Closed Itemsets (FCIs) are usually fewer than frequent

itemsets.

However, it is necessary to find Minimal Generators (mGs) for association rule from

them.

The finding mGs approaches based on generating candidate loose timelines when the

number of FCIs are large.

To overcome the loose of timelines we implemented an algorithm called MG-CHARM.

Page 4: mgcharm-150527055232-lva1-app6891

Introduction An efficient and fast algorithm for finding mGs of FCIs.

Time of MG-CHARM is fewer than the time of finding mGs after finding all closed

itemsets (CHARM), especially in case of the length of each FCI is long.

At present, almost all algorithms for mining mGs of FCIs are based on Apriori

algorithm.

Applying for Market Basket Analysis.

Page 5: mgcharm-150527055232-lva1-app6891

ScopeAlthough Market Basket Analysis conjures up pictures of shopping carts, and

supermarket shoppers, it is important to realize that there are many other areas in which

it can be applied.

These include:

Analysis of credit card purchases.

Analysis of telephone calling patterns.

Identification of fraudulent medical insurance claims.

Analysis of telecom service purchases.

Page 6: mgcharm-150527055232-lva1-app6891

Existing System In first method, found candidates that are mGs first, then defined their closures to

find out FCIs.

In second method, found all FIs using CHARM algorithm.

Then used level-wise method to find out all mGs that correspond to each closed

itemset.

Page 7: mgcharm-150527055232-lva1-app6891

Proposed System Both of the previous methods have the disadvantage in large size of frequent

itemsets since the number of considered candidates is large.

The proposed method is a fast algorithm for mining mGs based on CHARM.

It overcomes the disadvantage of above two methods by using CHARM to generate

FCI and also find mGs of them.

This proposed algorithm is named as MG-CHARM.

Page 8: mgcharm-150527055232-lva1-app6891

AlgorithmInput: The database D and

support threshold minS

Output: all FCI satisfy minSup and their

Method:MG-CHARM(D, minSup)[0]={IjXl(li) ,(li): t,E 1/\ cr(/ i) :2:

minSup}MG-CIIARM-EXTEND([0], C= 0 )Return CMG-CHARM-EXTEND([P], C)for each lixI(lj),mG(li) in [PJ doPi= Pj U Ii and [PiJ = 0for each IjxI(lj),mG(Ij) in [PJ,with j

> i doX = Ij and Y=I(li) n 1(1j)MG-CHARM-PROPERTY(XxY,/

i,lj ,Pi, [Pi],[P])SUBSUMPTION-CHECK(C, Pi)MG-CIIARM-EXTEND([Pi],C)MG-CHARM-

PROPERTY(XxY,li,lj,Pi, [Pi],[P])if cr(X):2: minSup thenif I(Ii) = I@ then II property IRemove Ij from PPi=Pi UIjmG(Pi) = mG(Pi) +mG(lj)else if I(li)C I@ then II property 2Pi =Pj uljelse if l(lj) :::J I(lj) then II property

3Remove Ij from [P]Add Xx Y,mG(Ij) to [Pi]else if I(li) *-1(1j) then II property 4AddXxY, u [mG(li), mG@]to [Pi]

Page 9: mgcharm-150527055232-lva1-app6891

Example1. Transaction Database

Consider the database

Transaction ID

Content

1 A, C, T, W2 C, D, W3 A, C, T, W4 A, C, D, W5 A, C, D, T,

W6 C, D, TExample database

Item TransactionsA 1, 3, 4, 5C 1, 2, 3, 4, 5, 6D 2, 4, 5, 6T 1, 3, 5, 6W 1, 2, 3, 4, 5

Vertical format

Page 10: mgcharm-150527055232-lva1-app6891

2. Support The number of transactions that has the given item set.

Support of ACW=4

3. Frequent Itemset An item set is said to be frequent item set if it is greater than or equal to the

minSup.

minSup is the value given by the user.

If minSup = 4 then ACW is a frequent item set.

Page 11: mgcharm-150527055232-lva1-app6891

4. Galois Connection

t(ACW) = t(A) t(C) t(W) i(245) = i(2) i(4) i(5)

= 1345 123456 12345 = CDW ACDW ACDTW

= 1345 = CDW

X t(X)

i(Y) Yi

t

Page 12: mgcharm-150527055232-lva1-app6891

5. Closure OperatorIf c(X) = i(t(X)) then c(X) is called closure operator.

c(AW) = i(t(AW))

= i(1345)

= ACW

6. Frequent Closed Itemset An item set X is said to be frrequent closed itemset if X is

- Frequent Itemset and

- Closure

Page 13: mgcharm-150527055232-lva1-app6891

Illustrations

Consider minSup = 50%

Since number of transactions = 6

minSup = 3

{}123456

A1345T1356 C123456D2456 W123

45

W CATD

Page 14: mgcharm-150527055232-lva1-app6891

Illustration of updating mG {}123

456

A1345T1356 C123456

D2456

W12345

W CATDC

DT56DT

DWC245

DWDA45DA

Page 15: mgcharm-150527055232-lva1-app6891

Functional DiagramUse Case Diagram

Page 16: mgcharm-150527055232-lva1-app6891

Sequence Diagram

Page 17: mgcharm-150527055232-lva1-app6891

Structural Diagram

Page 18: mgcharm-150527055232-lva1-app6891

ScreensLogon Screen

Page 19: mgcharm-150527055232-lva1-app6891

Recruit Employee Screen for Admin

Page 20: mgcharm-150527055232-lva1-app6891

Frequent Items

Page 21: mgcharm-150527055232-lva1-app6891

Frequent Items

Page 22: mgcharm-150527055232-lva1-app6891

Conclusion This is a new method for mining mGs of FCIs need not generate candidates.

Experiments showed that the time of updating mGs of frequent closed itemsets is insignifant. Especially in case of the large case of frequent itemsets, the time of updating is very fewer than the methods implementd before.

In future, we can apply this achievement to the problems of mining non-redundant association rules, non-redundant rules query.

Page 23: mgcharm-150527055232-lva1-app6891

References

[1]. IEEE supporting paper on the title “Fast Algorithm for Mining Minimal Generators of Frequent Closed Itemsets and Their Applications”. Link: http://ieeexplore.ieee.org/

[2]. Previous knowledge about mining frequent patterns, associations, and correlations from the text book ‘Data Mining Concepts and techniques by Micheline Kamber 2nd Edition’.

[3]. Information about Market basket Analysis.Link: http://www.information-drivers.com/market_basket_analysis.php

Page 24: mgcharm-150527055232-lva1-app6891

THANK YOU