What inspired this Wish List Request? A Pro opportunity with Thermofisher looking at using term selection to predict retention times for a protein chain based on the text form of the amino acids in the chain. We weren't able to use the technique because it doesn't treat the modeling type correctly.
Please describe the new or improved feature or solution you would like to see. Correct the DTM weightings in Text Explorer so that Log Freq and TF IDF settings result in Continuous columns, Binary and Ternary produce Nominal, and Frequency settings produce Ordinal. I provided examples for the issue in Case Number 00075358 to Tech Support around the issue when the incorrect modeling types are in the DTM for Term Selection
The short version is that Binary and Ternary are just categories - a value other than 0,1 (or 2 for ternary) are nonsensical. For frequencies, a value other than an integer implies it's possible to have 1.5 occurrences of a word in a sentence. In the case that triggered this request, it would imply it's possible to have 1.5 molecules. However, since all Term Weightings produce continuous columns in the DTM everywhere in Text Explorer, the predictions in the profilers do not reflect reality and will give strange results.
What would the value be to you and/or other users if this new feature is provided? The current settings make using the built-in modding capabilities in Text Explorer unusable for anything other TF IDF and Log Frequency situations.
Is this an external customer request? Thermofisher caught this will we were working on a Pro opportunity.
Is there a competitive software package that has this feature or solution that we should be aware of? N/A
Will this impact a renewal, expansion, or adoption at a customer, and does this put an account at risk if we do not address this? This is more of a fit-and-finish thing that got missed but will cause problems down the road.
