PPM-LASSO: Point process models with LASSO-type penalties

Description

This package contains tools to fit point process models with sequences of LASSO penalties ("regularisation paths"). Regularisation paths of Poisson point process models or area-interaction models can be fitted with LASSO, adaptive LASSO or elastic net penalties. A number of criteria are available to judge the bias-variance tradeoff.

Details

The key functions in ppmlasso are as follows:

Useful pre-analysis functions:

findRes

Determine the optimal spatial resolution at which to perform analysis

getEnvVar

Interpolate environmental data to species presence locations

griddify

Ensure a matrix of environmental data is on a rectangular grid

ppmdat

Calculate observation weights and set up design matrix for fitting

pointInteractions

Calculate interpoint interactions for fitting area-interaction models

sampleQuad

Set up a regular grid of quadrature points

Creating regularisation paths of point process models:

ppmlasso

Fit a regularisation path of point process models

plotFit

Plot the fitted intensity of a ppmlasso object

plotPath

Plot the regularisation path of a ppmlasso object

print.ppmlasso

Print output from a ppmlasso object

predict.ppmlasso

Make predictions from a fitted point process model to new data

Checking assumptions:

diagnose.ppmlasso

Create diagnostic residual plots of ppmlasso object

envelope.ppmlasso

Create simulation envelope for goodness-of-fit checks on a ppmlasso object

Author(s)

Ian W. Renner

Maintainer: Ian W. Renner <Ian.Renner@newcastle.edu.au>

References

Renner, I.W. & Warton, D.I. (2013). Equivalence of MAXENT and Poisson point process models for species distribution modeling in ecology Biometrics 69, 274-281.

Renner, I.W. et al (2015). Point process models for presence-only analysis. Methods in Ecology and Evolution 6, 366-379.

Renner, I.W., Warton, D.I., & Hui, F.K.C. (2021). What is the effective sample size of a spatial point process? Australian & New Zealand Journal of Statistics 63, 144-158.

Examples

# Fit a regularisation path of Poisson point process models
data(BlueMountains)
sub.env = BlueMountains$env[BlueMountains$env$Y > 6270 & BlueMountains$env$X > 300,]
sub.euc = BlueMountains$eucalypt[BlueMountains$eucalypt$Y > 6270 & BlueMountains$eucalypt$X > 300,]
ppm.form = ~ poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2, raw = TRUE)
ppm.fit  = ppmlasso(ppm.form, sp.xy = sub.euc, env.grid = sub.env, sp.scale = 1, n.fits = 20,
writefile = FALSE)

# Fit a regularisation path of area-interaction models
data(BlueMountains)
ai.form  = ~ poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2, raw = TRUE)
ai.fit   = ppmlasso(ai.form, sp.xy = sub.euc, env.grid = sub.env, sp.scale = 1, 
family = "area.inter", r = 2, availability = BlueMountains$availability, n.fits = 20,
writefile = FALSE)

# Print a ppmlasso object
print(ppm.fit, out = "model")

# Residual plot of a ppmlasso object
diagnose(ppm.fit, which = "smooth", type = "Pearson")

# Make predictions
pred.mu = predict(ppm.fit, newdata = sub.env)

# Plot the intensity from a fitted ppmlasso object
plotFit(ppm.fit)

# Plot the regularisation path from a fitted ppmlasso object
plotPath(ppm.fit)

Blue Mountains eucalypt and environmental data.

Description

This data set contains the observed presence locations of a Sydney eucalypt (eucalypt), the values of four environmental variables and two variables related to site accessibility throughout the region at a spatial resolution of 500m (env), and a matrix indicating whether locations in the region are available to the species (availability).

Usage

data(BlueMountains)

Format

A list with three objects:

eucalypt

A data frame with a column X of UTM Easting coordinates (km) and a column Y of UTM Northing coordinates (km) of observed locations of a Sydney eucalypt

env

A data frame containing environmental data in the Blue Mountains region near Sydney

availability

A 301x201 matrix with UTM Northing and Easting locations stored in dimnames indicating whether locations are accessible or not

The data frame env contains the following environmental data:

X

UTM Easting coordinates (km)

Y

UTM Northing coordinates (km)

FC

Number of fires since 1943

D_MAIN_RDS

Distance from the nearest main road (m)

D_URBAN

Distance from the nearest urban area (m)

RAIN_ANN

Average annual rainfall (mm)

TMP_MAX

Average maximum temperature (degrees Celsius)

TMP_MIN

Average minimum temperature (degrees Celsius)

Methods for function diagnose

Description

Methods for function diagnose

Methods

signature(fit = "ppm")

Creates residual plots for a ppm object. See the help function for diagnose.ppm in spatstat for more details.

signature(fit = "ppmlasso")

Creates residual plots for a ppmlasso object. See the help function for diagnose.ppm in spatstat for more details.

Create diagnostic plots for a fitted point process model.

Description

This function is analogous to the diagnose.ppm function of the spatstat package.

Usage

## S3 method for class 'ppmlasso'
diagnose(object, ...)

Arguments

Details

See the help file for diagnose.ppm in the spatstat package for further details of diagnostic plots.

Author(s)

Ian W. Renner

References

Baddeley, A.J. & Turner, R. (2005). Spatstat: an R package for analyzing spatial point patterns. Journal of Statistical Software 12, 1-42.

See Also

envelope.ppmlasso, for other goodness-of-fit functions inherited from spatstat.

Examples

data(BlueMountains)
sub.env = BlueMountains$env[BlueMountains$env$Y > 6270 & BlueMountains$env$X > 300,]
sub.euc = BlueMountains$eucalypt[BlueMountains$eucalypt$Y > 6270 & BlueMountains$eucalypt$X > 300,]
ppm.form = ~poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2) + poly(D_MAIN_RDS, D_URBAN, degree = 2)
ppm.fit  = ppmlasso(ppm.form, sp.xy = sub.euc, env.grid = sub.env, sp.scale = 1, n.fits = 20,
writefile = FALSE)
diagnose(ppm.fit, which = "smooth", type = "Pearson")

Methods for function envelope

Description

Methods for function envelope

Methods

signature(Y = "ppmlasso")

Creates Monte Carlo simulation envelopes for a given function for a ppmlasso object. See the help function for envelope in spatstat for more details.

Calculates simulation envelopes for goodness-of-fit

Description

This function is analogous to the envelope function of the spatstat package.

Usage

## S3 method for class 'ppmlasso'
envelope(Y, fun = Kest, ...)

Arguments

Details

See the help file for envelope in the spatstat package for further details of simulation envelopes.

Author(s)

Ian W. Renner

References

Baddeley, A.J. & Turner, R. (2005). Spatstat: an R package for analyzing spatial point patterns. Journal of Statistical Software 12, 1-42.

See Also

diagnose.ppmlasso, for residual plots inherited from spatstat.

Examples

data(BlueMountains)
sub.env = BlueMountains$env[BlueMountains$env$Y > 6270 & BlueMountains$env$X > 300,]
sub.euc = BlueMountains$eucalypt[BlueMountains$eucalypt$Y > 6270 & BlueMountains$eucalypt$X > 300,]
ppm.form = ~poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2) + poly(D_MAIN_RDS, D_URBAN, degree = 2)
ppm.fit  = ppmlasso(ppm.form, sp.xy = sub.euc, env.grid = sub.env, sp.scale = 1, n.fits = 20,
writefile = FALSE)
envelope(ppm.fit, Kinhom, nsim = 20)

Choose spatial resolution for analysis

Description

This function produces a plot to choose the optimal spatial resolution for analysis. A point process model is calculated for each nominated spatial resolution and the log-likelihood of all fitted models are plotted against the spatial resolutions.

Usage

findRes(scales, lambda = 0, coord = c("X", "Y"), sp.xy, env.grid, 
formula, tol = 0.01, ...)

Arguments

Details

This function produces a plot which can be used to judge an optimal spatial resolution for analysis. As the spatial resolution gets finer, the log-likelihood tends to stabilise to a constant value. The largest spatial resolution at which the log-likelihood appears to stabilise may be considered optimal for model fitting.

Value

A plot of log-likelihood versus spatial resolution.

Author(s)

Ian W. Renner

References

Renner, I.W. et al (2015). Point process models for presence-only analysis. Methods in Ecology and Evolution 6, 366-379.

Examples

data(BlueMountains)
sub.env = BlueMountains$env[BlueMountains$env$Y > 6270 & BlueMountains$env$X > 300,]
sub.euc = BlueMountains$eucalypt[BlueMountains$eucalypt$Y > 6270 & BlueMountains$eucalypt$X > 300,]
scales = c(0.5, 1, 2, 4, 8, 16)
ppm.form = ~ poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2)
findRes(scales, formula = ppm.form, sp.xy = sub.euc, env.grid = sub.env)

Extract environmental data to presence locations

Description

Given a matrix of quadrature points and a list of species presences, this function extracts environmental data to presence locations using bilinear interpolation.

Usage

getEnvVar(sp.xy, env.grid, env.scale, coord = c("X", "Y"), envfilename = "SpEnvData", 
tol = 0.01, writefile = TRUE)

Arguments

Details

At a given species location with coordinates (x, y), the interpolated value of the environmental variable z is calculated as a weighted average of z at four reference quadrature points (x^{(1)}, y^{(1)}), (x^{(1)}, y^{(2)}), (x^{(2)}, y^{(1)}) and (x^{(2)}, y^{(2)}) that form a square of nominated side length env.scale surrounding (x, y).

Value

A matrix containing locations of species presences in the first two columns and the interpolated environmental data in the remaining columns.

Author(s)

Ian W. Renner

Examples

data(BlueMountains)
species.env = getEnvVar(BlueMountains$eucalypt, env.grid = BlueMountains$env, env.scale = 0.5,
envfilename = NA, writefile = FALSE)

Ensure that a geo-referenced matrix of environmental grids is rectangular

Description

This function ensures that the coordinates of the supplied geo-referenced matrix of environmental grids constitutes a rectangular grid.

Usage

griddify(envframe, tol = 0.01, coord = c("X", "Y"))

Arguments

Details

The functions in the ppmlasso package require a set of quadrature points along a rectangular grid. At times a set of quadrature points with a desired spatial resolution of x_\delta \times y_\delta will have some minor machine error in some coordinates such that the coordinates as supplied do not consistute a rectangular grid. The griddify function corrects this error as follows:

Let \{x_1, x_2, \ldots, x_n\} and \{y_1, y_2, \ldots, y_n\} be the supplied coordinates contained in envframe. The function first determines the spatial resolution x_\delta \times y_\delta based on the median of the differences in the unique values of x_i and y_i as well as the coordinates of a rectangular grid with this spatial resolution \{x^{grid}_1, x^{grid}_2, \ldots, x^{grid}_n\} and \{y^{grid}_1, y^{grid}_2, \ldots, y^{grid}_n\}. Given the tolerance \epsilon supplied to tol, any coordinate x_i for which 0 < \left|x_i - x^{grid}_i\right| \leq \epsilon \times x_\delta will be adjusted to x^{grid}_i. Likewise, any coordinate y_i for which 0 < \left|y_i - y^{grid}_i\right| \leq \epsilon \times y_\delta will be adjusted to y^{grid}_i.

Any environmental variables contained in envframe are left unchanged.

Value

A data frame containing the coordinates on a rectangular grid as well as any environmental variables left unchanged.

Author(s)

Ian W. Renner

Examples

X = seq(0, 5, 1)
Y = seq(1, 11, 2)
XY = expand.grid(X, Y) # generate 1 x 2 rectangular grid
names(XY) = c("X", "Y")
#move some coordinates off of rectangular grid
XY$X[1] = XY$X[1] - 0.01
XY$Y[1] = XY$Y[1] - 0.01
XY$X[7] = XY$X[7] + 0.01
XY$Y[7] = XY$Y[7] + 0.01

#generate environmental variables
XY$V1 = 0.1*XY$X + 0.2*XY$Y + rnorm(36, 0, 1)
XY$V2 = -0.2*XY$X + 0.1*XY$Y + 0.05*XY$X*XY$Y + rnorm(36, 0, 5)

XY_grid = griddify(XY)

Plot the predicted intensity of a fitted ppmlasso model

Description

This function produces a levelplot of the predicted intensity from a fitted ppmlasso model.

Usage

plotFit(fit, pred.data = data.frame(X = fit$x[fit$pres == 0], 
Y = fit$y[fit$pres == 0], 1, scale(fit$data[fit$pres == 0, -1], 
center = -fit$s.means/fit$s.sds, scale = 1/fit$s.sds)), 
coord = c("X", "Y"), asp = "iso", ylab = "", xlab = "", 
col.regions = heat.colors(1024)[900:1], cuts = length(col.regions), 
cex = 1.4, main.text = paste(toupper(fit$criterion), "fit"), cex.color = 1.4)

Arguments

Details

This function will compute the predicted intensity of a fitted ppmlasso object using the model within the regularisation path which optimises the criterion specified in the call to ppmlasso.

Author(s)

Ian W. Renner

Examples

data(BlueMountains)
sub.env = BlueMountains$env[BlueMountains$env$Y > 6270 & BlueMountains$env$X > 300,]
sub.euc = BlueMountains$eucalypt[BlueMountains$eucalypt$Y > 6270 & BlueMountains$eucalypt$X > 300,]
ppm.form = ~poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2) + poly(D_MAIN_RDS, D_URBAN, degree = 2)
ppm.fit  = ppmlasso(ppm.form, sp.xy = sub.euc, env.grid = sub.env, sp.scale = 1, n.fits = 20,
writefile = FALSE)
plotFit(ppm.fit)

Plot of the regularisation path of a ppmlasso model

Description

This function produces a trace plot of the coefficient estimates of a fitted ppmlasso object and identifies the models which optimise the various penalisation criteria.

Usage

plotPath(fit, colors = c("gold", "green3", "blue", "brown", "pink"), logX = TRUE)

Arguments

Details

A fitted ppmlasso object contains a matrix called betas which stores the coefficient estimates of each of the n.fits models fitted. This function produces a traceplot of these coefficient estimates for each environmental variable and highlights the models which optimise each of the penalisation criteria.

Author(s)

Ian W. Renner

Examples

data(BlueMountains)
sub.env = BlueMountains$env[BlueMountains$env$Y > 6270 & BlueMountains$env$X > 300,]
sub.euc = BlueMountains$eucalypt[BlueMountains$eucalypt$Y > 6270 & BlueMountains$eucalypt$X > 300,]
ppm.form = ~poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2) + poly(D_MAIN_RDS, D_URBAN, degree = 2)
ppm.fit  = ppmlasso(ppm.form, sp.xy = sub.euc, env.grid = sub.env, sp.scale = 1, n.fits = 20,
writefile = FALSE)
plotPath(ppm.fit)

Calculate point interactions for area-interaction models

Description

This function calculates point interactions at presence locations and quadrature points required for fitting a regularisation path of area-interaction models.

Usage

pointInteractions(dat.ppm, r, availability = NA)

Arguments

Details

Theoretically, the point interaction t(y) at a point y is calculated as the proportion of available area in a circular region Y or radius r centred at y that overlaps with circles of radius r centred at other presence locations (Baddeley & Turner, 2005).

This function discretises the study region at the same spatial resolution as availability by defining the matrix occupied, a fine grid of locations spanning the study region initialised to zero. The values of occupied within a distance of r of each presence location are then augmented by 1, such that occupied then contains the total number of presence locations with which each grid location interacts. To prevent unavailable areas from being included in the calculation of point interactions, the values of occupied at grid locations for which availability = 0 are set to zero.

t(y) is then estimated as the proportion of available grid locations within Y that overlap circular regions around other presence locations.

The availability matrix is particularly useful for regions that have inaccessible areas (e.g. due to the presence of ocean or urban areas).

Finer resolutions of the availability matrix will yield more precise estimates but at a cost of greater computation time.

Value

A vector of point interactions corresponding to the locations contained in the dat.ppm argument.

Author(s)

Ian W. Renner

References

Baddeley, A.J. & Turner, R. (2005). Spatstat: an R package for analyzing spatial point patterns. Journal of Statistical Software 12, 1-42.

See Also

ppmlasso for fitting a regularisation path of area-interaction models

Examples

data(BlueMountains)
species.ppm = ppmdat(sp.xy = BlueMountains$eucalypt, back.xy = BlueMountains$env, 
sp.scale = 1, datfilename = NA, writefile = FALSE) # generate design matrix
species.int = pointInteractions(species.ppm, 2, BlueMountains$availability)

Prepare data for model fitting

Description

This function prepares the data for model fitting. In particular, it determines observation weights and sets up the design matrix required for fitting a regularisation path.

Usage

ppmdat(sp.xy, sp.scale, back.xy, coord = c("X","Y"), sp.dat = getEnvVar(sp.xy = sp.xy, 
env.scale = sp.scale, env.grid = back.xy, coord = coord, writefile = writefile), 
sp.file = NA, quad.file = NA, datfilename = "PPMDat", writefile = TRUE)

Arguments

Details

This function will call the sampleQuad and getEnvVar functions to generate a quadrature scheme and interpolate environmental data to presence locations. Alternatively, the quadrature scheme may be directly supplied to the quad.file argument, and the matrix of presence locations and associated environmental data may be directly supplied to either the sp.dat argument (as an object in the workspace) or to the sp.file argument (as the name of a saved file containing this matrix).

Value

A matrix dat.ppm with columns representing the latitude and longitude of presence locations and quadrature points along with the associated environmental data, as well as a column Pres indicating whether either point corresponds to a presence location or a quadrature point, and a column wt of observation weights.

Author(s)

Ian W. Renner

See Also

sampleQuad for generating a regular grid of quadrature points.

getEnvVar for interpolating environmental data to species presence locations.

Examples

## Do not run because of NOTE Examples with CPU time > 2.5 times elapsed time
#data(BlueMountains)
#species.ppm = ppmdat(sp.xy = BlueMountains$eucalypt, back.xy = BlueMountains$env, 
#sp.scale = 1, datfilename = NA, writefile = FALSE)

Fit point process models with LASSO penalties

Description

The ppmlasso function fits point process models (either Poisson or area-interaction models) with a sequence of LASSO, adaptive LASSO or elastic net penalties (a "regularisation path").

Usage

ppmlasso(formula, sp.xy, env.grid, sp.scale, coord = c("X", "Y"), 
data = ppmdat(sp.xy = sp.xy, sp.scale = sp.scale, back.xy = env.grid, coord = c("X","Y"),
sp.file = NA, quad.file = NA, datfilename = "PPMDat", writefile = writefile), lamb = NA,
n.fits = 200, ob.wt = NA, criterion = "bic", alpha = 1, family = "poisson", tol = 1.e-9, 
gamma = 0, init.coef = NA, mu.min = 1.e-16, mu.max = 1/mu.min, r = NA, interactions = NA, 
availability = NA, max.it = 25, min.lamb = -10, standardise = TRUE, n.blocks = NA, 
block.size = sp.scale*100, seed = 1, writefile = TRUE)

Arguments

Details

This function fits a regularisation path of point process models provided a list of species locations and a geo-referenced grid of environmental data. It is assumed that Poisson point process models (Warton & Shepherd, 2010) fit intensity as a log-linear model of environmental covariates, and that area-interaction models (Widom & Rowlinson, 1970; Baddeley & van Lieshout, 1995) fit conditional intensity as a log-linear model of environmental covariates and point interactions. Parameter coefficients are estimated by maximum likelihood for Poisson point process models and by maximum pseudolikelihood (Besag, 1977) for area-interaction models. The expressions for both the likelihood and pseudolikelihood involve an intractable integral which is approximated using a quadrature scheme (Berman & Turner, 1992).

Each model in the regularisation path is fitted by extending the Osborne descent algorithm (Osborne, 2000) to generalised linear models with penalised iteratively reweighted least squares.

Three classes of penalty p(\beta) are available for the vector of parameter coefficients \beta:

For the LASSO (Tibshirani, 1996), p(\beta) = \lambda*\sum_{j = 1}^p |\beta_j|

For the adaptive LASSO (Zou, 2006), p(\beta) = \lambda*\sum_{j = 1}^p w_j*|\beta_j|, where w_j = 1/|\hat{\beta}_{init, j}|^\gamma for some initial estimate of parameters \hat{\beta}_{init}.

For the elastic net (Zou & Hastie, 2005), \alpha*\lambda*\sum_{j = 1}^p |\beta_j| + (1 - \alpha)*\lambda*\sum_{j = 1}^p (\beta_j)^2. Note that this form of the penalty is a restricted case of the general elastic net penalty.

There are various criteria available for managing the bias-variance tradeoff (Renner, 2013). The default choice is BIC, the Bayesian Information Criterion, which has been shown to have good performance.

An alternative criterion useful when data are sparse is MSI, the maximum score of the intercept model (Renner, in prep). For a set of m presence locations, the MSI penalty is \lambda_{MSI} = \lambda_{max}/\sqrt{m}, where \lambda_{max} is the smallest penalty that shrinks all environmental coefficients to zero. The MSI penalty differs from the other criteria in that does not require an entire regularisation path to be fitted.

It is also possible to control the magnitude of the penalty by spatial cross validation by setting the criterion argument to "blockCV". The study region is then divided into square blocks with edge lengths controlled by the block.size argument, which are assigned to one of a number of cross validation groups controlled by the n.groups argument. The penalty which maximises the predicted log-likelihood is chosen.

Value

An object of class "ppmlasso", with elements:

Author(s)

Ian W. Renner

References

Baddeley, A.J. & van Lieshout, M.N.M. (1995). Area-interaction point processes. Annals of the Institute of Statistical Mathematics 47, 601-619.

Berman, M. & Turner, T.R. (1992). Approximating point process likelihoods with GLIM. Journal of the Royal Statistics Society, Series C 41, 31-38.

Besag, J. (1977). Some methods of statistical analysis for spatial data. Bulletin of the International Statistical Institute 47, 77-91.

Osborne, M.R., Presnell, B., & Turlach, B.A. (2000). On the lasso and its dual. Journal of Computational and Graphical Statistics 9, 319-337.

Renner, I.W. & Warton, D.I. (2013). Equivalence of MAXENT and Poisson point process models for species distribution modeling in ecology. Biometrics 69, 274-281.

Renner, I.W. (2013). Advances in presence-only methods in ecology. https://unsworks.unsw.edu.au/fapi/datastream/unsworks:11510/SOURCE01

Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society, Series B 58, 267-288.

Warton, D.I. & Shepherd, L.C. (2010). Poisson point process models solve the "pseudo-absence problem" for presence-only data in ecology. Annals of Applied Statistics 4, 1383-1402.

Widom, B. & Rowlinson, J.S. (1970). New model for the study of liquid-vapor phase transitions. The Journal of Chemical Physics 52, 1670-1684.

Zou, H. (2006). The adaptive lasso and its oracle properties. Journal of the American Statistical Association 101, 1418-1429.

Zou, H. & Hastie, T. (2005). Regularization and variable selection via the elastic net. Journal of the Royal Statistical Society, Series B 67, 301-320.

See Also

print.ppmlasso for printing features of the fitted regularisation path.

predict.ppmlasso for predicting intensity for a set of new data.

envelope.ppmlasso for constructing a K-envelope of the model which optimises the given criterion from the spatstat package.

diagnose.ppmlasso for diagnostic plots from the spatstat package.

Examples

# Fit a regularisation path of Poisson point process models
data(BlueMountains)
sub.env = BlueMountains$env[BlueMountains$env$Y > 6270 & BlueMountains$env$X > 300,]
sub.euc = BlueMountains$eucalypt[BlueMountains$eucalypt$Y > 6270 & BlueMountains$eucalypt$X > 300,]
ppm.form = ~ poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2)
ppm.fit  = ppmlasso(ppm.form, sp.xy = sub.euc, env.grid = sub.env, sp.scale = 1, n.fits = 20,
writefile = FALSE)

#Fit a regularisation path of area-interaction models
data(BlueMountains)
ai.form  = ~ poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2)
ai.fit   = ppmlasso(ai.form, sp.xy = sub.euc, env.grid = sub.env, sp.scale = 1, 
family = "area.inter", r = 2, availability = BlueMountains$availability, n.fits = 20,
writefile = FALSE)

Class "ppmlasso"

Description

A class ppmlasso which represents a point process model with a LASSO-type penalty.

Methods

diagnose

signature(object = "ppmlasso"): Produce diagnostic plots for a fitted point process model.

envelope

signature(Y = "ppmlasso"): Produce a Monte Carlo simulation envelope for a summary function of a fitted point process model.

predict

signature(object = "ppmlasso"): Calculate the predicted intensity for a fitted point process model to a set of data.

print

signature(x = "ppmlasso"): Print the details of a fitted point process model.

Author(s)

Ian W. Renner

Examples

showClass("ppmlasso")

Internal ppmlasso functions

Description

These are internal functions called by the main functions of ppmlasso, but are not generally used directly by the user.

Usage

blocks(n.blocks, block.scale, dat, seed = 1)
calculateGCV(y, X, ob.wt, b.lasso, lambda, alpha = alpha, 
unp.likelihood = unp.likelihood, penalty = TRUE, family = "poisson", 
mu.min = 1.e-16, mu.max = 1.e16, eta.min = log(1.e-16), eta.max = log(1.e16),
tol = 1.e-9, area.int = FALSE)
calculateLikelihood(X, family, ob.wt, mu, y, alpha, lambda, beta, penalty = FALSE)
calculateUnpenalisedLikelihood(ob.wt, y, mu)
catConvert(env.frame)
catFrame(cat.mat)
decimalCount(x, max.dec = max(10, max(nchar(x))), tol = 0.1)
etaFromMu(mu, family, mu.min = 1.e-16, mu.max = 1/mu.min, eta.min, eta.max)
interp(sp.xy, sp.scale, f, back.xy, coord = c("X", "Y"))
irlsUpdate(y, X, ob.wt, is.in, signs, eta, mu, alpha, lambda, beta.old, 
penalty = FALSE, family, mu.min = 1.e-16, mu.max = 1/mu.min, eta.min, 
eta.max, tol = tol)
makeMask(dat.ppm)
muFromEta(eta, family, mu.min = 1.e-16, mu.max = 1/mu.min)
polyNames(X)
ppmSpatstat(fit)
scoreIntercept(y, X.des, ob.wt = rep(1, length(y)), area.int = FALSE, int = NA, family)
singleLasso(y, X, lamb, ob.wt = rep(1, length(y)), alpha = 1, b.init = NA,
intercept = NA, family = "gaussian", tol = 1.e-9, gamma = 0, init.coef = NA, 
mu.min = 1.e-16, mu.max = 1/mu.min, area.int = FALSE, interactions, max.it = 25,
standardise = TRUE)
spatRes(env.grid, coord = c("X", "Y"))
standardiseX(mat)
weights(sp.xy, quad.xy, coord = c("X", "Y"))
zapCoord(x, numdec = decimalCount(x))

Details

These are generally not called by the user.

Author(s)

Ian W. Renner

Methods for function predict

Description

Methods for function predict

Methods

signature(object = "ppmlasso")

Creates a vector of predicted intensities for an object of class ppmlasso.

Prediction to new data from a fitted regularisation path

Description

Given a fitted regularisation path produced by ppmlasso, this function will predict the intensity for a new set of data.

Usage

## S3 method for class 'ppmlasso'
predict(object, ..., newdata, interactions = NA)

Arguments

Value

A vector of predicted intensities corresponding to the environmental data provided in the newdata argument.

Author(s)

Ian W. Renner

See Also

ppmlasso for fitting a regularisation path of point process models.

Examples

data(BlueMountains)
sub.env = BlueMountains$env[BlueMountains$env$Y > 6270 & BlueMountains$env$X > 300,]
sub.euc = BlueMountains$eucalypt[BlueMountains$eucalypt$Y > 6270 & BlueMountains$eucalypt$X > 300,]
ppm.form = ~ poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2, raw = TRUE)
ppm.fit  = ppmlasso(ppm.form, sp.xy = sub.euc, env.grid = sub.env, sp.scale = 1, n.fits = 20,
writefile = FALSE)
pred.mu  = predict(ppm.fit, newdata = sub.env)

Methods for function print

Description

Methods for function print

Methods

signature(x = "ppmlasso")

Prints output for a ppmlasso object with details controlled by arguments of the print.ppmlasso function.

Print a fitted regularisation path

Description

This function prints output from a fitted regularisation path.

Usage

## S3 method for class 'ppmlasso'
print(x, ..., output = c("all", "path", "model", "interaction"))

Arguments

Value

N/A

Author(s)

Ian W. Renner

See Also

ppmlasso for fitting regularisation paths.

Examples

# Fit a regularisation path of Poisson point process models
data(BlueMountains)
ppm.form = ~ poly(FC, TMP_MIN, TMP_MAX, RAIN_ANN, degree = 2)
ppm.fit  = ppmlasso(ppm.form, sp.xy = BlueMountains$eucalypt, 
env.grid = BlueMountains$env, sp.scale = 1, n.fits = 20, writefile = FALSE)
print(ppm.fit)

Generate regular grid of quadrature points with environmental data

Description

This function generates a regular grid of quadrature points and associated environmental data at a nominated spatial resolution.

Usage

sampleQuad(env.grid, sp.scale, coord = c("X", "Y"), quadfilename = "Quad", tol = 0.01, 
writefile = TRUE)

Arguments

Value

The output is a matrix of quadrature points at the spatial resolution supplied to sp.scale. If a vector of resolutions is supplied, the output is a list of matrices for each spatial resolution.

Author(s)

Ian W. Renner

Examples

data(BlueMountains)
quad.1 = sampleQuad(env.grid = BlueMountains$env, sp.scale = 1, quadfilename = NA, 
writefile = FALSE)