This subforum is intended to serve as a hub for discussions surrounding the IHSTAT_Bayes. Post here any comment / question / suggestion / bug reports.
IHSTAT Bayes – Peter Knott observations 30/3/22
Run on HP ProBook 430 G6
Intel(R) Core™ i7-8565U CPU @ 1.80GHz, RAM 8.00 GB (7.87 GB usable)
Windows 10 Pro, Version 10.0.19042 Build 19042
Selected “End”, opened up OK…
I’d keep the bar graph simple and 2 dimensional, just present the data, not the shading, also 95th perc, could either be 95thpercentile” or 95th%ile.
Percentile, Exceedance fraction, Arithmetic Mean are parameter estimates so in keeping with the IHSTATS naming convention I’d suggest the box is titled Lognormal parametric statistics” or Parameter estimates or something similar. I’d also be inclined to rename the arithmeric mean the “estimated”arithmetic mean so as not to generate controversy among prior IHSTAT users who may not have looked at Expostats before.
Exposure profile labels can overlap? With no ability to move them?
The drop down list ROS1…….are they examples?
Second time I ran the operation the QQ plot legend disappeared. Do you need to title the QQ plot?
What does Xaj do? (Upper right) Is it needed on the user input tab?
Auto scaling on the QQ plot and forcing axes through origin might be messy?
Is it possible to request new filename on save? i.e. force “Save as” to avoid writing over other files?
When I click on printer the dialog box is empty aside from Yes, No. also ran again with different data and QQ plot legend still missing? Something to do with the VBA error on start?
There would be real value in being able to just print the AIHA risk band plot, or output it as a graphics image (or pdf). Likewise the compliance results. I’m thinking about users who may have a few SEGS and want to put the data into a word table or excel. If you could output the results as a .csv or .xlsx which is easily cut / pasted/ imported into a table that would make the transition from IHSTATS Bayes to a report easier and with less opportunity for errors.
The IHDA-LE tool, presenting us with the possibility of Bayese inference in HI, allowed us to place prior information before the evaluation.
I don’t know if this will happen in a future revision of IHSTAT_Bayes. When I asked about it, I was told it was under study.
So, today where can we find the Bayese inference in its current version, to justify its name?