In boydpe/nphawkes: Nonparametric Hawkes Processes

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(nphawkes)

# Data Setup
# quake = read.csv("~/OSU/MS/eq_test1.csv")
# refd = "1/1/2010"
# full_time = lubridate::mdy(refd) + lubridate::duration(quake$t, "days")
# quake$date = lubridate::date(full_time)
# quake$time = paste(lubridate::hour(full_time),lubridate::minute(full_time),round(lubridate::second(full_time)), sep=":")

In this document, we will demonstrate the use of the nphawkes package. We will use the eq.RData file, found in this package.

This file contains 471 simulated earthquakes and contains the location in latitude and longitude (in [0,1] $x$ [0,1] unit square), date, time of day, and magnitude ($m$) of each event.

We will first employ the Model Independent Stochastic Declustering algorithm via the nph() function. This function must be executed at least in terms of a temporal Hawkes process, but it may also be analyzed as a spataio-temporal or marked spatio-temporal Hawkes process. Provided the given information (date, spatial location, and marks), we will use all three triggering components.

boydpe/nphawkes documentation built on March 31, 2022, 11:03 a.m.