You Can Be Naughty, But You Can’t Fake Out Santa

It was Friday the 13^th; Friday, December 13^th. The RIViR Program is always a high intensity affair and our team ended the week with a Christmas celebration luncheon in Annapolis. We had a merry and jolly time, and just as lunch was served my teammates looked at me with anticipation and wonderment.

I was raising a fork full of food when I noticed and asked, “What?” I wondered if I dropped food on my shirt.

“Do you think you’ll get the call?” someone asked.

I was thoughtful and very demure, “The Clauses are in a good place. They’ve been running things up North using RIViR for a while now. Fully patched and ready to go.”

I went back to my food, and so we all continued our merriment.

The afternoon ended, as Friday afternoons end, and my trip home was delightfully uneventful. It was when I got out of my car and shut the door when my phone buzzed. I checked the screen.

‘Mrs. Claus, North Pole’ splashed across the screen. I smelled cinnamon (or is that nutmeg?). I took the call.

“Hello Will. I’m so happy you had a merry afternoon with the RIViR team,” the Clauses are always watching. “I need a few minutes of your time.”

It was freezing outside, but I felt pleasantly comfortable. Like it was 73 degrees and sunny. I wonder if that’s quantum effects working or maybe just Christmas magic.

“Sure, what’s up Mrs. C?” I replied.

“You know, most of our children the world over, are good ones. They consistently make the nice list. However, we’ve detected some strange outliers this year, and our usual 1 or 2 percent variance on the naughty list is even lower,” she said. “That seems odd.”

“Oh. Well maybe tweaking a parameter or tw—“

“Now Will. I think we need more than tweaking. You’re fresh off of NHCAA and I’m sure you can deploy some new behavioral analysis techniques you learned at that conference,” she chided.

“Uhhh, sure-“, I replied.

“Will, I know you’ve been a good boy this year, for the most part. But, you went to NAMPI and NHCAA. That’s two times you’ve been to New Orleans this year… Oh my.”

“Understood, Mrs. Claus. We’re on it.”

You Need More Than Data to Fight Fraud

You’ve heard, ‘He’s making a list, he’s checking it twice.’ That makes good caroling material. The naughty and nice lists are constantly updated. The North Pole’s MFCU is constantly in the mix. For those of you not hip to the lingo MFCU stands for Merriment Fraud Control Unit. The N.P. MFCU samples both lists periodically looking for bad little children faking the funk and pretending to be good. They also verify the system isn’t producing false positives and making sure the good kids aren’t being placed on the naughty list. It would be a devastating Christmas for well behaved, kind, and respectful children to wake up and find coal in their stockings.

You see, the vast majority of kids are good. Then, there are a small but significant population who are downright naughty and subtly cheat the system. They don’t commit obvious offenses that are easy to detect. They try to cover up their dirty deeds by performing the occasional good one and are seemingly nice. These naughty kids require more than straightforward data analysis.

At Qlarant, our data scientists do a lot of upfront work identifying outlying youth on both sides of the good and bad spectrum. Beyond the obvious outliers, things get trickier. Are outliers truly outliers? Are they doing something different? Are they the goodest of the good and baddest of the bad, really?

That’s where behavioral analysis comes in, and extrapolating good and bad behaviors across the population. This involves more than processing claims of good and bad deeds.

Systematizing Behavioral Analysis

False positives end up on the naughty list, and fraudsters on the good list because data can only take you so far. We collaborate with the best and brightest minds who understand the behaviors that good and bad actors make. Behavioral patterns require a deeper dive than straightforward analytical processing. More features, metadata like social media, and publicly available information are needed for behavioral processing. For instance some small humans make public shows of doing good deeds and do dirty deeds to others that result in complaints on social media.

These behaviors are codified into our analytic techniques. That way the North Pole’s MFCU isn’t chasing shadows and fully utilize their resources. Throughout the year things change, and behaviors can shift. To account for this, we add a feedback loop to our systems so analysis and identification are constantly refined.

AI + HI = Merry Christmas!

We systematize all of this in RIViR using our Nerdy Algorithm technology. The Clauses already had a nimble setup with RIViR. Their data processing is configurable, providing the ability to update algorithms to account for behavioral patterns. We’ve activated smart workflows for them providing human feedback when outlying good or bad children are identified. This human feedback helps reduce false positives and at the same time refine analytical processing.

This is a win-win-win for the Clauses, good little children, and Christmas.

While you’re reading this, you can rest a little more assured, knowing that the Clauses and Qlarant are using the very best techniques and technologies making it a Merry Christmas for all of us.