In Fall 2011, UTD opened a new buffet where there are many food selections for f

Question

In Fall 2011, UTD opened a new buffet where there are many food selections for faculty and students. For simplicity, suppose five types of foods are offered daily: salad, hamburger, taco, soup and pasta. Suppose you are the manager and you decide to use associate rules (manually) to figure out what foods customers tend to purchase together. You recorded selections by five customers as shown in the Table below. You also decide to use the following cutoffs: minimum support 40% and minimum confidence 80%. What valid rules will you generate? Provide detailed steps with your relevant calculations. Also report support, confidence and lift for the final rules you generate.

Explanation / Answer

We will be using Apriori algorithm to generate the rules.

Sacnning the above table, we get the below table.

Minimum support = 40% * 5 = 2

As, the support of all food items is equal or greater than 2, we select all food items for next steps.

All combinations of 2 food items with their support is given below.

Selecting only those combinations of food items with the support of all food items is equal or greater than 2.

All combinations of 3 or 4 food items with their support is given below.

So, only Soup, Hamburger, Pasta has the support of 2.

So, Frequent food items with their support are

For rule Salad Hamburger :
support = support({Salad, Hamburger }) = 40%
confidence = support({Salad, Hamburger }) / support({Salad }) = 100%

For rule Hamburger Salad :
support = support({Salad, Hamburger }) = 40%
confidence = support({Salad, Hamburger }) / support({Hamburger }) = 66.67%

For rule Hamburger Soup :
support = support({Soup, Hamburger }) = 40%
confidence = support({Soup, Hamburger }) / support({Hamburger }) = 66.67%

For rule Soup Hamburger :
support = support({Soup, Hamburger }) = 40%
confidence = support({Soup, Hamburger }) / support({Soup }) = 66.67%

For rule Hamburger Pasta :
support = support({Pasta, Hamburger }) = 40%
confidence = support({Pasta, Hamburger }) / support({Hamburger }) = 66.67%

For rule Pasta Hamburger :
support = support({Pasta, Hamburger }) = 40%
confidence = support({Pasta, Hamburger }) / support({Pasta }) = 50%

For rule Pasta Soup :
support = support({Soup, Pasta }) = 60%
confidence = support({Soup, Pasta }) / support({Pasta }) = 75%

For rule Soup Pasta :
support = support({Soup, Pasta }) = 60%
confidence = support({Soup, Pasta }) / support({Soup }) = 100%

For rule Soup, Pasta Hamburger :
support = support({Soup, Hamburger, Pasta}) = 40%
confidence = support({Soup, Hamburger, Pasta }) / support({Soup, Pasta }) = 66.67%

For rule Soup, Hamburger Pasta:
support = support({Soup, Hamburger, Pasta}) = 40%
confidence = support({Soup, Hamburger, Pasta }) / support({Soup, Hamburger }) = 100%

For rule Pasta, Hamburger Soup :
support = support({Soup, Hamburger, Pasta}) = 40%
confidence = support({Soup, Hamburger, Pasta }) / support({Pasta, Hamburger }) = 100%

So, the following are rules with more than 80% confidence.

Salad Hamburger

Soup Pasta

Soup, Hamburger Pasta

Pasta, Hamburger Soup

Food Support Salad 2 Hamburger 3 Taco 2 Soup 3 Pasta 4

Related Questions

Navigate

• Browse (All)
• Browse I
• Subjects

---