The study of point processes using topological data analysis and applications
We live in an age where technology has been increasingly growing over
over the years. Such development is so fast that one can regard this
century as the age of information. Given the countless amount of data
being produced continuously, many researchers have been showing concern
to come up with new techniques to tackle all this data and to extract
as much information as possible - hidden through its complexity.
Among some of these new techniques being developed, Persistent Homology
shows to be a promising way to handle the geometric patterns hidden all
over the data. This is so, given that Persistent Homology makes use of
the field of Simplicial Homology to spotlight the patterns of data. With
that in mind, the core of our analysis will be based on making use of
this field of Topological Data Analysis.
So far, we have worked basically on two problems. Firstly, we have
studied a specific class of non stationary point processes and proposed
some convergences on some of its persistent features. Finally, we
started studying the behavior of Fake news on Twitter, applying
Persistent Homology on the graph made of all tweets
collected to detect its patterns.