Last I checked I was "tense". Is it theoretically possible to explain why your models came to that conclusion? Is there a theoretical data view that could say "here's the comments that contributed to you being considered "tense" (or even better: specific parts of the comments).
From my perspective, without a way to concretely back up any conclusion made, I simply cannot consider the complex-sounding systems to be any better than RNG.
I alluded to this in my first comment, but I think it's imperative that if you're going to explicitly call out what mood you think someone was/is/will be, you need to back it up with "why" so that, at the very least, someone can say, "that's clearly an unfair representation of my mood and here's where the algorithms messed up."
I'm going to go on a bit of a limb and assume that you're just having some fun and exploring an interesting feature. And I think that's really quite cool. But I think we're getting into a software engineering ethics realm that I'm ill-equipped to dig in to. You've got a site that applies "moods" to individuals in a way that doesn't appear to be verifiable. And even if they are, and are accurate, is that even the right thing to do?
I don't want to ruin your fun or twist your arm into changing this. I'm just responding to my red-alert alarm that lives in my stomach. It's set off by the discovery that one of my few online accounts where I don't hide my identity has been assigned a mood label.