How to draw P-P plot and Q-Q plot manually ?
I have looked at different site and they explained in various way, such as one said for p-p plot in X-axis there is residual in ascending order and in Y-axis there is z-value.
And another said in p-p plot in X-axis there is cumulative probability$(\frac{i-.5}{N})$ and in Y-axis there is expected cumulative probability. But i don't know how to compute expected cumulative probability.
And for q-q plot in X-axis there is residual in ascending order and in Y-axis there is z-value, which is confusing with the first method of p-p plot.
Q-Q plot: There is often confusion on which axis is X and which is Y. One book says one way, another book the other way. In the Normal Q-Q plot, I prefer X to be Z-scores (theoretical quantiles) and Y to be the observed (sample) values. This way, the slope of the plot gives an estimate of standard deviation. ($Y = \mu + \sigma Z$).
P-P plot: can think of it in the same way as Q-Q plot, but the values will be converted to probabilities. X-axis will just have probabilities $\frac{i-.5}{N}$. On Y-axis, you'll have to calculate the probabilities based on estimated parameters. E.g for Normal, first estimate $\mu, \sigma$ and then get $p_i = \Phi [(y_i - \hat{\mu})/\hat{\sigma}]$, where $\Phi$ is the standard Normal CDF. If you start out with a Q-Q plot and then apply non-uniform stretching to both x- and y- axes, you'll get a P-P plot.
The definition of P-P plot you mentioned in 1st paragraph seems incorrect. That's a Q-Q plot. Q for quantile, P for probability.
For example, see here http://v8doc.sas.com/sashtml/qc/chap8/sect9.htm