In cboettig/nonparametric-bayes: Nonparametric Bayes inference for ecological models

opts_chunk$set(external = TRUE, cache = FALSE, cache.path = "bevholt-cache/")
read_chunk('gaussian-process-control.R')
library(knitcitations)

We use a Beverton-Holt model to drive the underlying dynamics, with parameters $A =$ r p[1] and $B =$ r p[2].

sigma_g <- 0.02
z_g <- function(sigma_g) rlnorm(1, 0, sigma_g) #1+(2*runif(1, 0,  1)-1)*sigma_g #
x_grid <- seq(0, 1.5 * K, length=101)
h_grid <- x_grid
profit = function(x,h) pmin(x, h)
delta <- 0.01
OptTime = 20
xT = 0
reward = profit(x_grid[length(x_grid)], x_grid[length(x_grid)]) + 1 / (1 - delta) ^ OptTime 
## x_0_observed is starting condition for simulation of the observed data.  
## It should be in preferred state for bistable model, 
## above Allee threshold for Allee model, 
## and near zero for BH or Ricker models
x_0_observed <- x_grid[2] 

We consider stochastic growth driven by a lognormal noise process, $X_{t+1} = z_g f(X_t)$, where $f$ is the stock recruitment curve and $z_g$ a lognormal shock with $\sigma_g$ = r sigma_g.

Simulate data

Estimates a Ricker curve with parameters $r =$ r p_alt[1] and $K =$ r p_alt[2]

We fit a Gaussian process with

The transition matrix of the inferred process

bibliography("html")

cboettig/nonparametric-bayes documentation built on May 13, 2019, 2:09 p.m.