AnalyticBridge

Data Intelligence, Business Analytics

What are the most difficult things to predict, among these 19 challenges?

Please share your opinion and let us know how to make better predictions for the following problems. Also how do you define better?

Earthquake time and intensity forecasts
Solar flares time and intensity forecasts
Stock market prices for individual stocks or indexes, buy/sell signals
Efficiency of new pharma drugs (in terms of sales, or in terms of health improvement)
Weather forecasts in UK, the NorthWest or other challenging locales
Bankruptcy prediction
Econometric: how to predict the success of a new immigration policy, or a new tax regulation, or a new law to stop drug abuse or crime
Asteroid collisions with earth
Early detection of a pandemic
Chance for a borrower to repay her loan
Terrorist attacks
Chance of a 5 or 10 or 20-year survival for a particular person
Spam detection
Bogus reviews detection
Chance for an ad to result in a sale
Risk of car accident within the next 12 months for a particular individual
Chances of succeeding at University based on high school metrics
Chances of succeeding in a job based on job interview metrics
Chances of succeeding for a start-up, if you invested $10MM in the start-up in question

Comment

Comment by Shaizan on August 31, 2012 at 5:39am: It would be interesting to predict #16 Risk of car accident within the next 12 months for a particular individual

Comment by Benhenni on August 29, 2012 at 7:06pm: Regarding #11, I wasn't able to find the references, but I read about models doing correct predictions in few cases, giving probabilities of occurring events in a given area considering what happened in the neighboring. I know such models are now used somewhere in the United States by police officers to to "optimize" their patrol around predicted hot points.

Comment by Alfred on August 23, 2012 at 10:30am: I think economics and medical problems are the hardest, so I would say 7 then 6 12 and 19, since they are affected by non-measurable parameters.

I would add a last one thing to reach a round figure :

20. Long Term Demographic projections...?

Comment by Matthew A. Riebel on August 20, 2012 at 5:25pm: I hope it's 18 because I was continually rejected on interviews for a long time based on my social struggles even though I knew I would perform the job very well if given the opportunity.

I also think 11 is hard to predict. Consider how unprecedented September 11th was.

Comment by Nissim Matatov on August 19, 2012 at 5:10am: I think it depends on organization . Each time the challenge lie in different part of analytics solution : data or algorithm issue. I mean , for example , paragraph 10 can be the easiest and hardest in two different organizations.

You can talk about aggregated measure based on existing practices . May be it's worth to run an appropriate survey.

Welcome to
AnalyticBridge

Sign Up
or Sign In

Or sign in with:

1	Why is it so hard to count this way?
2	Yelp won't hire unemployed data scientists
3	Normalized odds-ratio
4	Reverse clustering for tactical warfare
5	Time Period for Analytical Positions Recruitment.
6	Seven Questions on Adopting Analytics Culture
7	Outliers : Capping rule
8	Correlation and R-Squared for Big Data
9	Cross-validation in R: a do-it-yourself and a black box approach
10	Do You Have What It Takes To Be a Data Scientist?