0
$\begingroup$

I haven't studied maths for over 30 years, but have been asked to solve a problem.

We have a very large data set, each data point contains 11 numbers. We would like to identify a recurring pattern within the data set. The pattern will consist of multiple data points.

Any advice on how to write an algorithm to carry out this task?

Thank you...

$\endgroup$
  • $\begingroup$ It depends on the data. There's no general algorithm for any data set out there. That being said, you can start by looking for an arithmetic progression, and then move on to a geometric progression. $\endgroup$ – barak manos Nov 5 '15 at 9:40
  • $\begingroup$ Surely these would only work where the pattern being detected was sequential? $\endgroup$ – user5308141 Nov 5 '15 at 10:39
0
$\begingroup$

What you ask belongs to a discipline called data mining. Of course, there is no single answer to your question, and the few useful algorithms that may exist to attack the problem, are considered highly non-trivial intellectual property of private firms that invested quite a lot in developing them.

$\endgroup$
  • $\begingroup$ Hi, thanks. So one way to solve this problem would be to use a Semi-supervised anomaly detection technique, construct a model representing normal behavior from a given normal training data set, and then testing the likelihood of a test instance to be generated by the learnt mode? $\endgroup$ – user5308141 Nov 5 '15 at 10:37
  • $\begingroup$ Yes, precisely so. You got it right. $\endgroup$ – uniquesolution Nov 5 '15 at 11:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.