If a linear model fits the data well then variable X can be used to predict variable Y; meaning variable X causes variable Y.
A linear regression can be used when the residuals are normally distributed and the data exhibits a linear relationship (visualized through plotting and computing the correlation).
Fill in the missing words. Hint: The two words are 'within' and 'outside'.
Interpolation occurs the dataset's observational range. Extrapolation occurs the dataset's observational range.
Why is extrapolation potentially problematic? (multiple answers may be correct)
- It could be that the variables outside of this range follow a completely different shape requiring distinctly different equations depending on the variable's value..
- It could be that the variables outside this range follow a similar shape, such as linear, but the parameters of the equation could be entirely different..
- Extrapolation can be computationally expensive..
- Assuming that the pattern holds for data outside the observed range is a very large assumption..
What's the first step in Quality Controlling your data after plotting?
- Remove values larger than the mean of the data +- standard devation of the data..
- Read the 'readme' file..
- Manually remove values that look extreme from the plot..
Match the image with the best description about Quality Control.
Fill in the missing words 'offset' and 'slope' appropriately.
We compute the first then use the answer to compute the .
The plot above shows some data (dots) with 3 linear fits (a,b, and c). Visually, which fit is the best?
- B.
- C.
- A.
I have an RMSE of 2.5 Degrees F for a linear model. Using the empirical rule, what would the 95% confidence interval be?
- +- 2.5 Degrees F.
- +-5 Degrees F.
- +- 7.5 Degrees F.
If I have an R-squared value of 0.93, then variable X can explain 93% of the variability in variable Y.