We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

eess.SP

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Electrical Engineering and Systems Science > Signal Processing

Title: Streaming Generalized Cross Entropy

Abstract: We propose a new method to combine adaptive processes with a class of entropy estimators for the case of streams of data. Starting from a first estimation obtained from a batch of initial data, model parameters are estimated at each step by combining the prior knowledge with the new observation (or a block of observations). This allows to extend the maximum entropy technique to a dynamical setting, also distinguishing between entropic contributions of the signal and the error. Furthermore, it provides a suitable approximation of standard GME problems when the exacted solutions are hard to evaluate. We test this method by performing numerical simulations at various sample sizes and batch dimensions. Moreover, we extend this analysis exploring intermediate cases between streaming GCE and standard GCE, i.e. considering blocks of observations of different sizes to update the estimates, and incorporating collinearity effects as well. The role of time in the balance between entropic contributions of signal and errors is further explored considering a variation of the Streaming GCE algorithm, namely, Weighted Streaming GCE. Finally, we discuss the results: in particular, we highlight the main characteristics of this method, the range of application, and future perspectives.
Comments: 22 pages, 11 figures
Subjects: Signal Processing (eess.SP); Information Theory (cs.IT)
Journal reference: Soft Computing (2019)
DOI: 10.1007/s00500-019-04632-w
Cite as: arXiv:1811.09710 [eess.SP]
  (or arXiv:1811.09710v2 [eess.SP] for this version)

Submission history

From: Mario Angelelli [view email]
[v1] Fri, 9 Nov 2018 13:17:49 GMT (165kb,D)
[v2] Tue, 14 Jan 2020 13:14:47 GMT (200kb,D)

Link back to: arXiv, form interface, contact.