Title: | Continuous Development Models for Incremental Time-Series Analysis |
Version: | 0.1.3 |
Date: | 2018-05-01 |
Author: | Bijan Seyednasrollah, Jennifer J. Swenson, Jean-Christophe Domec, James S. Clark |
Maintainer: | Bijan Seyednasrollah <bijan.s.nasr@gmail.com> |
Description: | Using the Bayesian state-space approach, we developed a continuous development model to quantify dynamic incremental changes in the response variable. While the model was originally developed for daily changes in forest green-up, the model can be used to predict any similar process. The CDM can capture both timing and rate of nonlinear processes. Unlike statics methods, which aggregate variations into a single metric, our dynamic model tracks the changing impacts over time. The CDM accommodates nonlinear responses to variation in predictors, which changes throughout development. |
Depends: | R (≥ 3.3.0) |
Imports: | rjags |
License: | MIT + file LICENSE |
Encoding: | UTF-8 |
LazyData: | true |
RoxygenNote: | 6.0.1.9000 |
BugReports: | https://github.com/bnasr/phenoCDM/issues |
Suggests: | knitr, rmarkdown |
VignetteBuilder: | knitr |
NeedsCompilation: | no |
Packaged: | 2018-05-01 23:34:01 UTC; bijan |
Repository: | CRAN |
Date/Publication: | 2018-05-02 03:45:15 UTC |
Fit a CDM Model
Description
This function fits a CDM model on the input data as it is described by the phenoSim function.
Usage
fitCDM(x, z, connect = NULL, nGibbs = 1000, nBurnin = 1, n.adapt = 100,
n.chains = 4, quiet = FALSE, calcLatentGibbs = FALSE, trend = +1)
Arguments
x |
Matrix of predictors [N x p]. |
z |
Vector of response values [N x 1]. |
connect |
The connectivity matrix for the z vector [n x 2]. Each row contains the last and next elements of the time-series. NA values indicate not connected. |
nGibbs |
Number of MCMC itterations |
nBurnin |
Number of burn-in itterations. |
n.adapt |
Number of itterations for adaptive sampling |
n.chains |
Number of MCMC chains |
quiet |
logical value indicating whether to report the progress |
calcLatentGibbs |
logical value indicating whether to calculate the latent states |
trend |
time-series expected trend as -1:decreasing, +1:increasing, 0: not constrained |
Examples
#Summarize CDM Model Ouput
ssSim <- phenoSim(nSites = 2, #number of sites
nTSet = 30, #number of Time steps
beta = c(1, 2), #beta coefficients
sig = .01, #process error
tau = .1, #observation error
plotFlag = TRUE, #whether plot the data or not
miss = 0.05, #fraction of missing data
ymax = c(6, 3) #maximum of saturation trajectory
)
ssOut <- fitCDM(x = ssSim$x, #predictors
nGibbs = 200,
nBurnin = 100,
z = ssSim$z,#response
connect = ssSim$connect, #connectivity of time data
quiet=TRUE)
summ <- getGibbsSummary(ssOut, burnin = 100, sigmaPerSeason = FALSE)
colMeans(summ$ymax)
colMeans(summ$betas)
colMeans(summ$tau)
colMeans(summ$sigma)
Summarize Output of the CDM Model
Description
This function return a summary of the output from the Gibbs-Sampling of the CDM model.
Usage
getGibbsSummary(ssOut, burnin = NULL, colNames = NULL,
sigmaPerSeason = TRUE)
Arguments
ssOut |
CDM output list. |
burnin |
Number of burnin itterations . |
colNames |
vector of charachters includes names of each variable in the output. |
sigmaPerSeason |
logical value indicating whether each site/season has a separate process error |
Examples
#Summarize CDM Model Ouput
ssSim <- phenoSim(nSites = 2, #number of sites
nTSet = 30, #number of Time steps
beta = c(1, 2), #beta coefficients
sig = .01, #process error
tau = .1, #observation error
plotFlag = TRUE, #whether plot the data or not
miss = 0.05, #fraction of missing data
ymax = c(6, 3) #maximum of saturation trajectory
)
ssOut <- fitCDM(x = ssSim$x, #predictors
nGibbs = 200,
nBurnin = 100,
z = ssSim$z,#response
connect = ssSim$connect, #connectivity of time data
quiet=TRUE)
summ <- getGibbsSummary(ssOut, burnin = 100, sigmaPerSeason = FALSE)
colMeans(summ$ymax)
colMeans(summ$betas)
colMeans(summ$tau)
colMeans(summ$sigma)
Simulate Green-up Phenology Data
Description
This function return a set of simulated data for multiple green-up phenology time-series.
Usage
phenoSim(nSites = 1000, nTSet = c(3:6), p = 2, beta = NULL, sig = 0.1,
tau = 0.01, miss = 0, plotFlag = FALSE, ymax = 1, trend = +1)
Arguments
nSites |
Number of sites/seasons |
nTSet |
A vector of integer values. Length of each time-series will be randomly sampled from this vector. |
p |
Number of predictors in the model. |
beta |
Beta coefficients |
sig |
Process error. |
tau |
Observation error. |
miss |
Fraction of missing data. |
plotFlag |
logical value indicating whether to plot the resulted time-series. |
ymax |
Asymptotic maximum values. |
trend |
time-series expected trend as -1:decreasing, +1:increasing, 0: not constrained |
Examples
#Simulate Phenology Data
ssSim <- phenoSim(nSites = 2, #number of sites
nTSet = 30, #number of time steps
beta = c(1, 2), #beta coefficients
sig = .01, #process error
tau = .1, #observation error
plotFlag = TRUE, #whether plot the data or not
miss = 0.05, #fraction of missing data
ymax = c(6, 3) #maximum of saturation trajectory
)
Plot Simulated Phenology Data
Description
This function plots the time-series data described with a connectivity matrix.
Usage
phenoSimPlot(z, connect, add = FALSE, col = "blue", ylim = range(z, na.rm
= TRUE), pch = 1, lwd = 1)
Arguments
z |
A vector of time-series data [n x 1] |
connect |
The connectivity matrix for the z vector [n x 2]. Each row contains the last and next elements of the time-series. NA values means not connected. |
add |
logical value indicating whether the plot should be overlaid on the current panel. |
col |
The color variable as charachter |
ylim |
Range of the y axis |
pch |
pch value for the symbols |
lwd |
lwd value for line tickness |
Examples
#Simulate Phenology Data
ssSim <- phenoSim(nSites = 2, #number of sites
nTSet = 30, #number of time steps
beta = c(1, 2), #beta coefficients
sig = .01, #process error
tau = .1, #observation error
plotFlag = TRUE, #whether plot the data or not
miss = 0.05, #fraction of missing data
ymax = c(6, 3) #maximum of saturation trajectory
)
#Plot Simulated Data
phenoSimPlot(ssSim$z, ssSim$connect)
Plot Observed vs Predicted
Description
This function plot posterior distributions of the parameters.
Usage
plotPOGibbs(o, p, nburnin = NULL, xlim = range(o, na.rm = TRUE),
ylim = range(p, na.rm = TRUE), xlab = "Observed", ylab = "Predicted",
colSet = c("#fb8072", "#80b1d3", "black"), cex = 1, lwd = 2, pch = 19)
Arguments
o |
Observed vector |
p |
Predicted Gibbs samples |
nburnin |
numbe of burn-in itterations |
xlim |
x-axis range |
ylim |
y-axis range |
xlab |
x-axis label |
ylab |
y-axis label |
colSet |
vector of colors for points, bars and the 1:1 line |
cex |
cex value for size |
lwd |
line width |
pch |
pch value for symbols |
Examples
ssSim <- phenoSim(nSites = 2, #number of sites
nTSet = 30, #number of Time steps
beta = c(1, 2), #beta coefficients
sig = .01, #process error
tau = .1, #observation error
plotFlag = TRUE, #whether plot the data or not
miss = 0.05, #fraction of missing data
ymax = c(6, 3) #maximum of saturation trajectory
)
ssOut <- fitCDM(x = ssSim$x, #predictors
nGibbs = 200,
nBurnin = 100,
z = ssSim$z,#response
connect = ssSim$connect, #connectivity of time data
quiet=TRUE)
summ <- getGibbsSummary(ssOut, burnin = 100, sigmaPerSeason = FALSE)
colMeans(summ$ymax)
colMeans(summ$betas)
colMeans(summ$tau)
colMeans(summ$sigma)
par(mfrow = c(1,3), oma = c(1,1,3,1), mar=c(2,2,0,1), font.axis=2)
plotPost(chains = ssOut$chains[,c("beta.1", "beta.2")], trueValues = ssSim$beta)
plotPost(chains = ssOut$chains[,c("ymax.1", "ymax.2")], trueValues = ssSim$ymax)
plotPost(chains = ssOut$chains[,c("sigma", "tau")], trueValues = c(ssSim$sig, ssSim$tau))
mtext('Posterior distributions of the parameters', side = 3, outer = TRUE, line = 1, font = 2)
legend('topleft', legend = c('posterior', 'true value'),
col = c('black', 'red'), lty = 1, bty = 'n', cex=1.5, lwd =2)
yGibbs <- ssOut$latentGibbs
zGibbs <- ssOut$zpred
o <- ssOut$data$z
p <- apply(ssOut$rawsamples$y, 1, mean)
R2 <- cor(na.omit(cbind(o, p)))[1,2]^2
#Plot Observed vs Predicted
par( mar=c(4,4,1,1), font.axis=2)
plotPOGibbs(o = o , p = zGibbs,
xlim = c(0,10), ylim=c(0,10),
cex = .7, nburnin = 1000)
points(o, p, pch = 3)
mtext(paste0('R² = ', signif(R2, 3)), line = -1, cex = 2, font = 2, side = 1, adj = .9)
legend('topleft', legend = c('mean', '95th percentile', '1:1 line', 'latent states'),
col = c('#fb8072','#80b1d3','black', 'black'),
bty = 'n', cex=1.5,
lty = c(NA, 1, 2, NA), lwd =c(NA, 2, 2, 2), pch = c(16, NA, NA, 3))
Plot Posterior Distributions
Description
This function plot posterior distributions of the parameters.
Usage
plotPost(chains, trueValues = NULL, outline = FALSE)
Arguments
chains |
Gibbs sampling chains |
trueValues |
numeric vector of true values |
outline |
logical value whether showing outliers |
Examples
ssSim <- phenoSim(nSites = 2, #number of sites
nTSet = 30, #number of Time steps
beta = c(1, 2), #beta coefficients
sig = .01, #process error
tau = .1, #observation error
plotFlag = TRUE, #whether plot the data or not
miss = 0.05, #fraction of missing data
ymax = c(6, 3) #maximum of saturation trajectory
)
ssOut <- fitCDM(x = ssSim$x, #predictors
nGibbs = 200,
nBurnin = 100,
z = ssSim$z,#response
connect = ssSim$connect, #connectivity of time data
quiet=TRUE)
summ <- getGibbsSummary(ssOut, burnin = 100, sigmaPerSeason = FALSE)
colMeans(summ$ymax)
colMeans(summ$betas)
colMeans(summ$tau)
colMeans(summ$sigma)
par(mfrow = c(1,3), oma = c(1,1,3,1), mar=c(2,2,0,1), font.axis=2)
plotPost(chains = ssOut$chains[,c("beta.1", "beta.2")], trueValues = ssSim$beta)
plotPost(chains = ssOut$chains[,c("ymax.1", "ymax.2")], trueValues = ssSim$ymax)
plotPost(chains = ssOut$chains[,c("sigma", "tau")], trueValues = c(ssSim$sig, ssSim$tau))
mtext('Posterior distributions of the parameters', side = 3, outer = TRUE, line = 1, font = 2)
legend('topleft', legend = c('posterior', 'true value'), col = c('black', 'red'),
lty = 1, bty = 'n', cex=1.5, lwd =2)