1 / 11

Power Models

Power Models. Presentation 2-10. Power Models. Consider the following dataset about the cost of pizzas from Joanies Pizzaria. Of course, we start with a scatterplot. Power Models. From the scatterplot, we probably suspect it is not linear. It may be exponential or something else.

sorena
Télécharger la présentation

Power Models

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Power Models Presentation 2-10

  2. Power Models • Consider the following dataset about the cost of pizzas from Joanies Pizzaria. • Of course, we start with a scatterplot.

  3. Power Models • From the scatterplot, we probably suspect it is not linear. • It may be exponential or something else.

  4. Power Scatterplots • Power models appear when each x is raised to a constant power. • For powers between 0 and 1, the model is logarithmic.

  5. Power Situations • Area and Volume • Comparing length to area (2nd power) or volume (3rd power) • Such as the pizza problem or relating foot length to weight (as in most real regressions, there are outliers). • Predicting the height of a tree from its cross sectional area of its trunk.

  6. Power Transformation • The trick is to change the data such that it is linear. • This can be done by taking the logarithm of both the explanatory and response variables. • Remember properties of logarithms (when you multiply powers, you add (linear) exponents • So, essentially, you change the x values and y values to exponents • Now look at the new scatterplot of (log x, log y)

  7. Power Transformation • Now look at the new scatterplot of (log x, log y) • This scatterplot looks more centered about a line as opposed to a curve. • Now that it is linear, we do a least-squares regression line on the transformed (log x, log y)

  8. Power Transformation • The r-squared value looks good (0.9797) • The regression line (LSRL) is given as: y = 1.3389x – 0.1148 • This is not quite right as it should be: log y = 1.3389*log x – 0.1148

  9. Power Transformation • Now to solve the equation for y, so we don’t have to deal with the logarithm. • Start by exponentiating, then using properties of logarithms to solve for y-hat.

  10. Power Transformation • This could also be done using ln or natural logarithms. • The only difference would be base e instead of base 10.

  11. Power Transformation • This concludes the presentation.

More Related