Graded Lab Part II

• Examine additional variables:
  1. Protein
  2. Fat
  3. Sodium
  4. Carbo
  5. Sugars
  6. Potass
  7. Vitamins
• Open training file, cerealTrain_YourInitials

• Add new sheet and build Combo Chart with Name as Dimension and the following expressions:
  • Only(Rating), label it ActualRating (make it Bar Chart)
  • Formula for Sugars: $(=LINEST_M(Rating, Sugars))*Only(Sugars)+$(=LINEST_B(Rating,Sugars)), label it PredictedbySugars. (Making it Line Chart)
  • Formula for additional 2 most significant variables (replace Sugars with variable name in the formula above)(Make these Line Charts)

• Compare Actual and Predicted Rating values for each Model. Analyze the results. Is it possible to identfy the best model by looking on the chart.

• Find Test_R^2 for additional models. Compare the performance.

• Create Combo Chart to compare Actual Rating with Predicted Rating for several models. Analyze the chart

Documentation and Submission

• Document R^2 and Linear Regression Formula for each Model in the Word file cerealResultComplete.docx
• Submit: THREE FILES: Word File with the Summary, 2 QlikView files (training and testing)