Just brainstorming - Reliability prediction, Weibull analysis and the airline industry


#1

I’m one of those graduates that studied for a masters in ‘operational research’ after my engineering degree and went on to help my clients understand in service failures of aircraft engines. The other option was a career in finance but it never rubbed off. Today I’m applying stochastic processes, probabilistic analysis, that sort of thing to historical events, usually failures of one kind or another. The other element of my work is deterministic. I now spend much of time predicting when events shall likely occur using both deterministic (if new product/service) and probabilistic analysis for in-service services of products.

I’m always amazed at the accuracy achieved. On some occasions, albeit a year ago or so, my model instructed the withdrawal of an aircraft from service two weeks before an engine failure. If your an airline operator it is always far less expensive to know when failures are going to occur rather than reacting to them. Airlines will pay for this information.

Time is the problem that is always against me. Lots of data, no analysis, is an event imminent? No one knows.

The question, can something like the Numerai network facilitate such analysis? The models are not necessary complicated, 2-3 weeks to set up, but are generally specific to some life distribution.


#2

I’m not really sure what you’re asking at the end? The competition at Numerai specifically uses data provided by Numerai for one purpose. At least at present, there isn’t any announced plan to include other datasets/problems in the competition, and even if there were Numerai is a hedge fund not a general machine learning platform like Kaggle so the other data would certainly be financially related to inform how they invest their assets under management.


#3

I see. Might look into Kaggle… I’m looking for opportunities to monetise problem solving of this sort, i.e. recognise the problem and then determine the life distribution. The rewards have a strong financial incentive to operators.