Regression fit - The Noise factor for obscuring romantic relationships

Three hundred and fifty people at the ISB – and almost eighty women. So you’ve gotta have some sizzle goin, right?

Spot on. But its too early in the course for couples to be too brazen about it. The budding romance might not stand the glare of rabid attention. So couples stoop to impossible levels of discretion to camouflage their sex lives.

So how DOES one spot these couples? Heck –this is B-school! So what do you do when you are in doubt? Regress!!

(For the non-statisticians and non-B-schoolers, Regression is the legalized version of “stoned” behavior - seeing patterns where none exist - Editor)

So what you do is this – you compile the data for people arriving at the dining hall together, look for patterns of two people arriving at and leaving at the same time, assign highest weightage for early breakfast arrivals (go figure!), progressively lower weightage to arrivals toward lunchtime, and again progressively higher weightage to people going for coffee, early dinner, a dip in weightage for busy dinner time, and again higher weightage for later dinnertimes.

Regress – presto! – you have spanking new data on the hottest and newest couples on campus!!!! Coming from a block that has , on an average 2 IITians per quad, further analysis has inevitably shown the following R² interpretations -

(Again for you illiterates R² – being a measure of fit – in simpler terms – a measure of how much the pattern obscures the dots you used to make the pattern in the first place)

· 19% and above - A “new” couple – high on momentum and rosy-cheeked love. The gastronomic data equivalent (this is dining hall data, remember?) of the “can’t-keep-hands-off-each-other” phase.

· 14%-18.99% - Phase II – not yet naming their kids, but already into “you know your underarms stink” mode

· 11-13.99% - A “mature couple” - true sign of a relationship that’s ageing well. Probably indicates a five year plan that includes choice of school kids will go to.

This is all ridiculously “intuitive” – you might say. Now the IITian infested paparazzi team bring in game theory – Couples know they are couples. Couples also know people in all their morbid curiosity wanna know every gory detail of the relationship. Couples also know that Regression is the only tool the paparazzi can use (I DID warn ya– This IS B-school, mate!). So what do couples do ? Add Noise!!! (You were wondering about the low R² factors, weren’t ya?) (Noise and R² being opposing concepts – it is explained by the exact opposite of the R² explanation a few lines above. More dots than pattern. Thud.)

So Noise, huh? Yup! Couples add more noise by doing your best to be seen with as many different people as you can..or if they think they’re really smart, they add another pattern altogether by being seen with another chick/dude. But I am guessing this one is like all smart-aleck approaches, it probably brings in a set of problems that Mr./Ms. Smart-Aleck hadn’t bargained for!

As our non-resident evangelist economist says “go forth and shift those graphs!”

PS: IPR for the concept - in collaboration with Reverend Shankar, Smoker Sandy and Regressor Raj,