# Perform regression analysis using linear mode, logarithmic model

The Flyair Company manufactures a sophisticated radar unit that is used in a fighter aircraft built by Seaways Aircraft. The first 85 units of the radar unit have been completed and Flyair is preparing to submit a proposal to Seaways Aircraft to manufacture the next 50 units. Flyair wants to submit a competitive bid, but at the same time, it wants to ensure that all the costs of manufacturing the radar unit are fully covered as part of this process. Flyair is attempting to develop a standard for the number of labor hours required to manufacture each radar unit. Developing a labor standard has been a continuing problem in the past but they believe that one way to do so is to predict the labor hours required from the number of units made. Columns K and L show the number of labor hours required for each of the first 85 units of production.

a. Create a scatter plot of the number of labor hours as a function of the units made. (In this context, such a plot is referred to as a learning curve.) Does the scatter plot suggest no relationship, a linear relationship or a curved relationship between labor hours and units made? Put your answer in cell adjacent and the details of your work in this sheet.

b. No matter what you found in the previous question, assume that Flyair wants to explore several kinds of fit between Labor Hours and Units Made (they want to explain or predict Labor Hours with Units Made). In particular, they want to compare a linear model with a logarithmic model, a power model and a polynomial model of order 2. See cells for model and variation, enter how much variation in Labor Hours is explained by Units Made for the 4 possible models. Then in cell R3, enter which model you prefer.

c. Using your preferred model from the previous question, how many labor hours do you predict will be required to make the next (the 86th) unit? Report your answer.