Disclaimer: Okay, I really do have the utmost respect for all law enforcement agents. I am sure Chief Jones must truly be a wonderful gentleman and I am grateful to him and all the other officers who put their lives on the line everyday to keep the rest of us safe. And I’m not just saying that in case I ever get pulled over in Trinity, TX. Well, okay, maybe that factors in a *little* bit in this disclaimer.
So Chief Jones is going to help me explain Bayesian inference to you folks today. Yes, the same Chief Jones from the stellar, suspense-filled, groundbreaking television masterpiece known as FAT COPS which I have highly recommended before. I am still on pins and needles to see if it will be renewed for a second season. I believe that it truly must be or else that would be proof right there that there is no more justice in the world. Furthermore, it should be expanded on to the silver screen. Yes, and forget about Order of Dimensions (you: Already have!), this … this right here must happen. Yes, John Krasinski, my future husband, and one of my two of my most favoritest people in the world (Not Dictionary Lady. The other one.) should really be calling their agents now to see if they could land the role of one Chief Steven Jones. I mean, by God, this man is an American, nay, a world, nay, a universal, nay, a multiversal hero. He risked facing the wrath of his wife to look for a poor, defenseless peacock, people! How many men would be brave enough to do that? Stupid enough and yet brave enough? And peacocks are not the only fowl he vowed to serve and protect either! Yes, and I smell
Razzie Oscar for whoever can pull off the role of such a brave, heroic, inspirational … okay, I’ll just move on to today’s lesson then.
Hey dere! So I’m gonna be helpin’ Irene talk about Bayesian inference with you folks and I’ll be a usin’ the example of that chained up peacock I found down dere in dat shack den. So like I said, I found it fishy chained up like dat in the first place, but I ain’t gonna be messin’ with mah wife more bein’ forty-five minutes late and den some so I just gotten there the next mornin’ for investigatin’ and seeing another bird, lets sayin’ a hoopoe, chained up again. Now, I agains free de bird and give it back to de rightful owner dere but again … somfin’s fishy. So I gets back to de same place a few times and each time see a funny lookin’ bird chained up down dere! Now, I reckon I have data to form a likelihood as is likely I be findin’ ‘nother funny lookin’ bird chained up if I go back down dere. But I know nuffin’ yet about dat’s shack’s owner! So if I wanna set up a prior, I could be usin’ a noninformative prior, like from them uniform distribution, so I be assumin’ equally likely probabilities that maybe the shack’s owner got somefin’s fishy goin’ on or maybe he ain’t got somefin’s fishy goin’ on but based on our likelihood, we git dis dere posterior distribution leadin’ us to believe that somefin’s indeed fishy down dere! And den mah Deputy Big Sexy calls me wid dis dere fax from the Huntsville office, sayin’ dis dere purple people-and-bird-eater from outah space dere been capturin’ dese birds, see, and chainin’ ’em up in dat poor shack owner’s shack like dat. Well, dat ain’t right! So our posterior changes with dis new prior information, see, and we know somefin’s fishy but it’s not wid the shack owner, it’s with the people-and-bird-eater from outah space, see? So dis dere is how we updated our posterior distribution. And …
Okay, maybe I better stop before I cause any more injustice to Chief Jones or Thomas Bayes. But join me next time when I embark on another attempt to make another statistical concept appealing to the masses. I promise it’ll be good next time — or at least less sucky. In the meanwhile, FAT COPS: The Movie — let’s make this happen, people!