Clust-values < e-13 - Githubissues

holgerteichgraeber / TimeSeriesClustering.jl

Julia implementation of unsupervised learning methods for time series datasets. It provides functionality for clustering and aggregating, detecting motifs, and quantifying similarity between time series datasets.

MIT License

82 stars 23 forks source link

Clust-values < e-13 #34

Closed YoungFaithful closed 5 years ago

YoungFaithful commented 5 years ago

If values are too small and close to 0, the constraint is neglected of Gurobi. I fear that I'll need to revise my results in that regard. I think a simple round shall solve that.

holgerteichgraeber commented 5 years ago

Could you put a minimum working example, I'm not sure I understand the issue. Generally if you encounter a bug, could you include a minimum working example?

holgerteichgraeber commented 5 years ago

Ok, I see you unassigned me, so I guess you are working on it now.

YoungFaithful commented 5 years ago

Yes, I would have done so, if I wouldn't have fixed it on my own.

holgerteichgraeber commented 5 years ago

Nice, great!

YoungFaithful commented 5 years ago

Actually the problem is much less important than I thought. While checking possible cases I got some weird values. After an hour I figured out that the amount of iterations for kmeans is also for k=1 crucial as the solar ratios from 0-1 can scale to 7000 ;) Luckily JuMP is all right and the problem was setup the right way.

holgerteichgraeber commented 5 years ago

Why is the amount of iterations important for k=1?

Technically, it just calculates the mean over all values, which should not differ by iteration.
Why to the solar ratios scale to 7000?

YoungFaithful commented 5 years ago

I didn't understand that either. Let me have a look how I can recreate the issue and then put it into a branch

YoungFaithful commented 5 years ago

kBug-Branch