Difference between revisions of "Talk:Beta Tester Page"

 
Line 22: Line 22:
 
Functions
 
Functions
 
---------
 
---------
-I think the SampleWeighting function is extremely useful since in risk analysis we are interested inthe tails of the density distribution. It is particularly useful when that probability density is wide or skewed. I did not follow the instructions to change that weighting!  I would need to see an example of how to define wiethign to ehance the sampling of the tails. What that be achieve by specifing a weithing function!?
+
-I think the SampleWeighting function is extremely useful since in risk analysis we are interested inthe tails of the density distribution. It is particularly useful when that probability density is wide or skewed. I did not follow the instructions to change that weighting!  I would need to see an example of how to define weighting to enhance the sampling of the tails (!?). Would that be achieved by specifying a weighting function?

Revision as of 18:46, 13 April 2007

Pierre Richard 13 April 2007 comments on 4.0 build 4.0.0.39


Graphic


- Many excellent Graph setup changes: especially colour, line thickness, fill, and log scale are excellent improvements for presentations - There is a small square at bottom-right that does not fill with colour(!?) - 3D is of no apparent use - it does not improve readability of graphics - Can there be a B&W line style option (broken, dotted lines for eack key value (for use in publications)? - How do you select Number format from the x axis independently of the y axis?


Uncertainty Setup


- Probability bands - can you add 97.5% and 2.5% for an alpha=0.05. It is often required to present confidence intervals that have an alpha= 0.05.


Windows Cascading


BUG !!!!!!- Once, after cascading all open windows, the one with the basic model window partailly filled the window and did not expand correctly when I double-clicked on the top bar. I could not reproduce that problem again so I am not sure what I did differently. It seemed like a redraw malfunction, maybe prior operations required some large memory usage and that affected the redraw(!?). I had just done several runs with large uncertainty samples and I had aslo changes to the resampling method to Knuth!?

Functions


-I think the SampleWeighting function is extremely useful since in risk analysis we are interested inthe tails of the density distribution. It is particularly useful when that probability density is wide or skewed. I did not follow the instructions to change that weighting! I would need to see an example of how to define weighting to enhance the sampling of the tails (!?). Would that be achieved by specifying a weighting function?

Comments


You are not allowed to post comments.