Excel Answers

Design a worksheet in MS Excel to implement Naive Bayes Classifier. The worksheet shall

include the following:

• It should allow users to create and store training data tuples. Let the attributes in the

tuples be nominal and output labels belong to one among ‘K’ classes.

• It should compute “K’ class aprioi probabilities.

• It should compute class conditional probabilities of each attribute taking values from

the set of values

• it should utilize these computed class conditional probabilities to classify

future/unknown data tuples based on MAP criterion.Please refer to the table given below. It has four nominal attributes viz. Weather, temperature,

Humidity and Wind. Let the output variable be Play which would take two values either ‘Yes’

or ‘No’ i.e. K = 2.

Design a worksheet in MS Excel to implement Decision Tree Classifier. The worksheet shall

include the following:

• It should allow users to create and store training data tuples. Let the attributes in the

tuples be nominal and output labels belong to one among ‘K’ classes.

• It should compute one of the following measures to decide an attribute to be used for

splitting the data tuples.

◦ Information gain

◦ Gain ratio

◦ Gini index

• It should split data tuples under consideration into multiple subsets based on the

selected attribute that maximizes the measure used and depending on number of

values the selected attribute would take.

• Repetition of the above steps of computing the measure to split further and continue

until a suitable stopping criteria is met.

Question 1

The Unilever Company plans to allocate some or all of its monthly advertising budget of GH¢82,000 in the Mankato area. It can purchase local radio spots at GH¢120 per spot, local TV spots at GH¢600 per spot, and local newspaper advertising at GH¢220 per insertion.

The company's policy requirements specify that the company must spend at least GH¢40,000 on TV and allow monthly newspaper expenditures up to GH¢60,000.

The payoff from each advertising medium is a function of the size of its audience. The general experience of the firm is that the values of insertions and spots in terms of "audience points" (arbitrary unit), are as given below:

Radio 40 audience points per spot

TV 180 audience points per spot

Newspapers 320 audience points per insertion

(a) Formulate a linear programming model for this problem.

(b) Use solver to find optimal solution and sensitivity report.