Information Theory in the Context of Volcano Seismic Singals for Forecasting Purposes
Description:
Information Theory is a field that addresses how to send and reconstruct messages, through a noisy channel, with low probability of error, and how to quantitatively measure the information contained in this message. Studies performed by Nyquist and Hartley, found that, logarithmic function is the most appropriate to measure the amount of information contained in a message. Shannon complete these studies and propose the Entropy as the measure of the information of a random variable, were each of the possible outcomes of the variable has a defined probability of occurrence and the value associated with the outcome is aleatory.
Important word in Information theory are: Information is related with the knowledge obtained when a specific event is observed, for example, realizations of a random variable with low probability are less expected and give more information that one with high probability. Uncertainty is the amount of information we expect an experiment to reveal. For uniform probabilities distribution the result is unpredictable and the uncertainty is maximum, and if the distribution of probabilities is more concentrated then the result is more predictable and there is a lower uncertainty. Entropy indicate the amount of uncertainty associated with a variable when only its distribution is known. The entropy provided a mathematical way of determining the information contained in a message. In the context of Information Theory, the Information is independent of the previous knowledge and only depend of the source via the probability function of each of the symbols issued.
We propose the application of different definitions of entropy proposed in the literature to the time series of seismic signals of volcanic origin. Taking into account previous works developed, we consider that it is of interest to study the evolution throughout the different evolutionary periods of volcanoes. We propose the comparative study of differential entropy, Shannon entropy, approximation entropy and Rényi entropy.
Session: Multidisciplinary Approaches for Volcanic Eruption Forecasting [Poster Session]
Type: Poster
Date: 5/2/2024
Presentation Time: 08:00 AM (local time)
Presenting Author: M. Carmen
Student Presenter: No
Invited Presentation:
Authors
M. Carmen Benítez Presenting Author Corresponding Author carmen@ugr.es Universidad de Granada |
Pablo Rey-Devesa pablord@ugr.es Universidad de Granada |
Janire Prudencio janire@ugr.es Universidad de Granada |
Manuel Marcelino mmtitos@ugr.es Universidad de Granada |
Jesús Ibáñez jibanez@ugr.es Universidad de Granada |
|
|
|
|
Information Theory in the Context of Volcano Seismic Singals for Forecasting Purposes
Category
Multidisciplinary Aproaches for Volcanic Eruption Forecasting