Alternative methods for novelty detection and correlations

Hey mates I have the following project:

Imagine having two datasets A and B. Each dataset consits of 101 time series with the same lenght and identical time stamps. The two datasets where taken from the same experiment, therefore the data structure is identical. From the 101 time series there is one particulary signal that is of interest in both datasets. Lets call that signal X(t)_101. Now we have the following case that the signal X(t)_101 from dataset A (good dataset) is significantly different than from dataset B (bad dataset). The goal now is to analyse the remaining 100 signals and find the signals that have the biggest effect on making our signal of interest X(t)_101 go bad. So in the end I would get some sort of ranking, on which signal is responsible for the change of our signal X(t)_101.

So far I am looking into novelty detection and correlation, are there any other keywords I should consider for this topic?

Thanks

Topic anomaly-detection correlation time-series

Category Data Science

1
answered at

I'm doing something similar and found that the best approach has to be comprehensive. My process obtains a ranking of signals based on degree and measure of anomaly, both as a function of 'bads' (if available). If information on 'bads' is unavailable then I perform the same study but as a function of the degree of anomaly. In both cases, I perform hierarchical clusters. These are helpful in grouping signals that have similar behavior.

About

Geeks Mental is a community that publishes articles and tutorials about Web, Android, Data Science, new techniques and Linux security.