Hi! What column types are valid to be used as output for predictive modeling? I’m using a simple table with text and long text columns. No options are visible in the “Column to Predict” dropdown.
@tony Mind helping out?
@Dave The valid predicted output fields are Yes/No, Enum, Ref, Number, Price, and Decimal.
Care to share what you’re trying to predict? I’d love to help out.
Hi @tony Thank you for the info. I figured this would be a bit of a long shot. To give you an idea of what we do and what I’m looking to accomplish long term; our agency provides behavioral therapy to treat autism and other developmental disabilities. Our therapists are currently using AppSheet apps to collect data on structured treatments, which is used (among other things) to create periodic progress reports that are submitted to funding sources. Generating these reports is one of the most time consuming aspects of our practice, and the time spent is non-billable. I am ultimately looking to automate the generation and delivery of data insights as standardized natural-language narratives/reports using machine learning. This is a very new (maybe non-existent?) approach in an industry that is already slow to adopt technology solutions. Still, I believe this approach lends itself well, given that our data sets are fairly structured and non-ambiguous.
Short term, I would like to be able to generate brief (one-paragraph) progress summaries based on data collected on targeted skill areas. I realize this might be beyond what the AppSheet platform is designed to do currently, but I would be very interested in any insights you might have.
Thanks for the info. Right now it sounds like the predictive models are not what you’re looking for. AppSheet’s predictive models are useful for predicting a value, like a category or a number or a yes/no. For example, estimating the cost of a project based on its attributes, or categorizing a comment based on the words that appear in it.
That makes sense. I appreciate the additional information. You guys have a great product, I’ve been very happy with it so far.