I'm an electrical engineer and I'm not so familiar with data mining concepts (I studied clustering concepts though). I have a data set that includes the measured electricity consumption of households over a year (So, the time series is too long).
I want to calculate the entropy for each household separately based on its time series to see which ones show more regular patterns.
Could you please introduce me some proper entropy measures for this?
(Note: the time series is long hence, I need a suitable measure for that)
Also, if you could introduce me a reference that explains the difference between regularity, complexity, and variability I would be grateful.
Dear Md. Milon and Ratul,
Thanks for your answers. I'll read them and if I could find anything else I'll share it here.
Dear Yi,
Thanks a lot for your response. I had read your other papers previously and also these two papers that include entropy measures in their studies. I have cited them in my paper which has been submitted recently for IEEE Trans. on Smart Grids.
[1] -"Clustering of electricity consumption behavior dynamics toward big data applications"
[2] -"Smart meter driven segmentation: What your consumption says about you"
[3] -"Household energy consumption segmentation using hourly data"
However, in all of these papers, the entropy, for example, Shannon entropy is calculated using the processed data. For instance, in your paper you calculated the Shannon entropy of the state transition matrix which is resulted from HMM or in [3] they compare the customer' consumption with a load shape encoding dictionary and then calculate the entropy.
My main question, however, is how to apply entropy measures on the raw data for example a time series? I wonder that is this helpful at all and can give meaningful results?
- Badges
- Science topic
- Similar topics
- Computer Science
- Data Mining