Modelos estadísticos para la detección de patrones en medio ambiente y economía :

Abstract

On average each year the information available on the network is duplicated and sources such as blogs, social networks, and web pages generate huge amounts of information that can be accessed through Web-Scraping techniques. In recent years, various methodologies have been proposed for measuring uncertainty based on news counts related to a set of keywords. The present work closely follows these methodologies. I use Topics modeling due to the variety of opinions and the amount of data obtained, which are probabilistic models. These models are based on two assumptions: 1) There are several different groups or text sources in an extensive collection of documents. 2) Texts from different sources tend to use different vocabulary. In particular, the Latent Dirichlet Allocation (LDA) algorithm addresses this work. From the problem of choosing the optimal number of topics to classifying the different documents related to uncertainty in different topics. Finally, the uncertainty to be broken down into its possible causes.