CHEM312 Problem Set: Predicting Maximum Daily
Ozone Concentrations Using Multivariable Linear Regression
The ozone data set contains 8 hour ozone readings from the Mecklenburg
County Line Site from 1990-1997. This monitoring station is located
within one mile of Charlotte (Lowe's) Motor Speedway and has the highest
average ozone readings among all ozone monitors in both Carolinas. The
data include, in addition to ozone readings, daily max temperatures, humidity,
and average wind speeds.
Requirement:
- Conduct three
single factor regression models to predict ozone as a function of only
temperature, wind speed, and humidity respectively. For each model:
- Write the
applicable equation (specifying units)
- State the fraction
of ozone variance accounted for by the regression
- Comment on whether
the variable is a factor that significantly affects ozone concentration;
fully support your answer.
- Conduct three
two-variable regression models to predict ozone levels using temperature,
wind speed and humidity factors
- Write the
applicable equation (specifying units)
- Comment on the
fraction of variance accounted for by each
- Identify the best two-variable model and compare
this to the best single-factor model
- Develop a three-variable model to predict ozone
from temperature, wind speed and humidity data.
- Write the
applicable equation (specifying units)
- Comment on the
fraction of ozone variance accounted for by the model
- Compare this model
to others and to its effectiveness in predicting ozone concentrations