This article is a follow-up to an earlier post titled "Related Posts - A New Approach". If you've not read that, I suggest you go over and take a look now.
So, why a follow up to my earlier post? There are 2 possible factors. When you start to pay attention to a specific thing, your news feeds and the Internet gods start to show you more information on those things you are interested in. But that is not all. You start to notice headlines, as they strike a chord in your brain & relate to what you have been spending time on.
In my case, I've found 2 things that have turned my thinking on it's head regarding related posts.
The first was "Omnity" (https://www.omnity.io/ though their site appears to be down currently). This is an interesting approach to creating relationships amongst large amounts of unstructured data, such as text or web pages. The way this is funded is probably as interesting as their technology implementation, so I highly recommend you go take a look.
The second was Bibblio (https://www.bibblio.org/). This is a Machine Learning implementation much, much closer to what I was investigating with Microsoft Azure Decision Services. This service effectively ticks all the boxes for the kind of "related content" service I had in mind at this time. I will be signing up to this and giving their free plan a run through it's paces to check the results.
This doesn't mean that my interest in Machine Learning has come to an end. Just that my immediate use for it may be satisfied via a COTS service offering, rather than "rolling my own".