logo

Association Rules (Relevant Output)

11 Pages978 Words303 Views
   

Data Mining and Visualisation for Business Intelligence (ITC516)

   

Added on  2020-03-16

About This Document

Rule 1 – For any customer who makes a purchase of brush, a conditional probability of 1 exists for making a subsequent purchase of nail polish. Rule 2 - For any customer who makes a purchase of nail polish, a conditional probability of 06322 exists for making a subsequent purchase of brush. Rule 3 - For any customer who makes a purchase of nail polish, a conditional probability of 0.5919 exists for making a subsequent purchase of bronzer.

Association Rules (Relevant Output)

   

Data Mining and Visualisation for Business Intelligence (ITC516)

   Added on 2020-03-16

ShareRelated Documents
DATA MINING ANDVISUALIZATIONStudent Id and Name[Pick the date]
Association Rules (Relevant Output)_1
Question 1Association Rules (Relevant Output)(i)The question highlights that the interpretation of the first three rules needs to beperformed (Rouse, 2004).Rule 1 – For any customer who makes a purchase of brush, a conditional probability of 1 existsfor making a subsequent purchase of nail polish.1
Association Rules (Relevant Output)_2
Rule 2 - For any customer who makes a purchase of nail polish, a conditional probability of06322 exists for making a subsequent purchase of brush.Rule 3 - For any customer who makes a purchase of nail polish, a conditional probability of0.5919 exists for making a subsequent purchase of bronzer.(i)A given rule is declared to be redundant if the underlying support closely resembles theexpected value on the basis of the ancestor rule. The redundancy for the given data isapparent in rule 2 where the support level is same as that of the ancestor which was onexpected lines. Thus, rule 2 can be considered as redundant for the cosmetic data as noincremental value is added (Prithiviraj & Porkodi, 2011). The above rules utility can be truly determined by paying attention to the support (captured bylift ratio) and confidence (captured by confidence level). Rules that have a higher lift ratio aregiven a higher significance as is represented in the above output as well. Moreover, those whichhave a higher confidence level are indicative of higher conditional probability. This wouldtypically enable finding of hidden and unexpected relations (Ragsdale, 2014). (ii) XL miner (75% minimum confidence)2
Association Rules (Relevant Output)_3
Unlike, the association rule output with 50% minimum confidence, in case of 75% minimumconfidence, the rules that are represented in the output have been trimmed to one. This is quitenatural since only the rules which have higher than 75% confidence level would be part of theoutput. However, a disadvantage of increasing the minimum confidence interval is that, somerules which are otherwise important on account of high support might not be considered(Liebowitz, 2015). 3
Association Rules (Relevant Output)_4

End of preview

Want to access all the pages? Upload your documents or become a member.

Related Documents
Data Mining and Redundant Rule Assignment
|9
|993
|143

Data Mining Assignment PDF
|9
|840
|46

Data Mining Assignment
|9
|972
|38

Data Mining Assignment | Business Analytics Assignment
|8
|876
|157

Data Mining Assignment -Dendrogram
|9
|973
|68

Data Mining Association Rules & Clustering
|9
|951
|62