Create or machine-learn fuzzy logic rules for use with an on-line inference engine
New DocAndys SaaS service supports user-created embeddable Fuzzy Logic Expert Systems. Use rule language Darl to hand-create or machine-learn rule sets from data and use them via REST interfaces.
Making use of our new rule language, DARL, users can hand-create or machine-learn rule sets from data on-line that can be used via REST interfaces.
Darl is unique in that well-tried Supervised, Reinforcement and Unsupervised learning algorithms exist that create Darl code directly as a result of the training process. The rule-sets created are readily readable and annotated with learning statistics as comments.
The Darl language itself is extremely simple, but modular and composable. It is possible to mix hand-coded and machine-learned code in the same solution and maintain each independently.
Rule sets can be automatically converted to questionnaires using our own incremental inference technology which reverse-engineers rule sets to find the most salient input variables, and presents those first to the user.
A full suite of on-line editors are provided for the Darl source, questionnaire text and formatting.
Our REST interfaces are media-agnostic, meaning you can create conversations, eliciting user input and inferring from the same, over web sites, apps, tweets, instant messages, texts or emails.
A wide variety of technologies and platforms are supported by code examples and down-loadable apps, especially Microsoft API Apps and Logic Apps.
On-line documentation and tutorials are provided, and for deeper insights we have a new book available to download or read in print called DARL - AI Online: Build a Fuzzy Logic Expert System
Usage of the site and creation of rule sets is free for low volume traffic and charged per use for higher volumes.
Contact sales@docandys.com and follow us on Twitter @doctorandys and our FaceBook page.
Darl® is a trademark of Dr Andy's IP.