Examine through plots §

---

Notes and Ideas: §

Diagnostics for the predictor X:

Want to check for outlying observations. To do so we want to examine the following plots:

1. Dot plot
2. Stem and Leaf plot
3. Box Plot
4. Sequence Plot (if the observations are time ordered)

The following are some of the possible violations of
assumptions:

1. Regression Function is non linear

2. Non-constant variance

3. Error terms not independent

4. Possibility of Outliers

5. Non Normal distribution of errors

6. Omission of some of the important predictors

The following Plots are considered:

1. Plot of explanatory and response
2. Plot of Residuals against the explanatory variables
   (Similar to plot of residuals to predicted)
3. Plot of Squared or absolute residuals and predicted
4. Plot of residuals against time sequence
5. Plot of residuals against omitted variables
6. Box Plot of Residuals
7. Normal probability Plot

---

Key words: §

---

TAGS §