Selasa, 12 Agustus 2014

Somebody Needs to Start Calling “Foul” On Proprietary Predictive Algorithms When They Cannot Be Replicated For Accuracy As This Accelerates Inequality and Promotes Even More Data Selling For Profit

Predictive analytics have been oversold in many areas.  I’m  not saying there’s not good use for looking for patterns and determining outcomes as there is but it is getting to the point to where anyone can create a model and claim “it works”.  When the models are proprietary, there’s no way to check for accurate outcomes and decisions are made on such models. There’s a lot of false positives and nobody questions that element and thus so people are getting hurt when that occurs.  In healthcare we have seen it for years with patterns that look for fraud with false positives when the parameters are nailed down too tight.  That has been going on for years.

There is a direct correlation here with inequality as every time a score is made, and it can’t be verified, folks just assume well it must be right when in fact it is not.  It certainly allows for more data to be sold whenever a score is created as now you have some new data to sell to someone else to assist in their decision making.  I call the Attack of the Killer Algorithms, which it is when flawed data denies consumers access. 

Data Selling and the Direct Correlation To Accelerated Inequality - Epidemic Spreading Like A Virus Moving Money Keeping Corporations Cash Rich and Consumers Cash Poor

It even gets worse as sometimes there’s no science at all with models and data scientists or quants just guess..and consumer is denied access, on a “guess” of a number component of a formula.  This FICO medication adherence is one of those as nobody can replicate the proprietary model.  So you just believe any numbers and stats tossed at you?  We sure have a lot of that today. 

FICO Medication Adherence Scoring Should Be Banned As It’s Quantitated Justifications for Profit That Hurts US Consumers Using Proprietary Algorithms That Cannot Be Replicated For Accuracy or Audited

Almost a couple years ago I agreed with a banker too that due to this fact that half of the analytics purchased will be a waste of investment.  We have too many folks looking for “Algo Fairies” when they don’t exist.   A couple years ago I predicted this too and it’s here today with big data being used out of context with discriminatory practices against consumers as everyone just believes anything tossed at them. 

Big Data/Analytics If Used Out of Context and Without True Values Stand To Be A Huge Discriminatory Practice Against Consumers–More Honest Data Scientists Needed to Formulate Accuracy/Value To Keep Algo Duping For Profit Out of the Game

It’s an interesting world indeed that is ending up to be the undoing of what we have developed as our culture over the years, one side hiding risk and other parts exploiting it.  Is there any balance here when you bring in the error factors that are always there and fictional models?  Of course not and it is making everybody crazy and this give a lot people a big area of control and moves a lot of money to the direction of very few in the US so the land of opportunity tends to shrink a bit.  You have to deal with and try to make sense of the fiction that’s mixed in out here.

Again I have written my share of SQL queries and have been a query monster in search of value in querying everything I could get my hands on and did I find value in everything I queried..NO.  There’s a lot of deception out there on where predictive analytics works and what context the numbers are used.  After the fact you can stir up up any old story just looking at data and really miss what occurred.  I just sat there with disbelief with the recent Facebook and OKCupid data scientist stories and scratched my head on how folks could think they could have any impact there at all.  Code runs hog ass wild and you have no control as it’s there to make money and knows how to grind you day in and day out.

Here’s the cause to create a law to where every data seller would need to be licensed and disclose what kind of data they sell and to who.  Scroll down and watch the 4 videos in the footer and the message in this post will become crystal clear as to what’s really going on behind closed server doors that you can’t see, touch, feel or talk to.  Be a skeptic when you need to be and that’s becoming more frequent today as if you don’t, well you’re “Algo Duped” at your own risk.  BD 

Tidak ada komentar:

Poskan Komentar