# RELEASE NOTES ## spatstat 1.47-0 ### 12 October 2016 We thank Marcel Austenfeld, Guy Bayegnak, Tilman Davies, Cenk Icos, Jorge Mateu, Frederico Mestre, Mehdi Moradi, Virginia Morera Pujol, Suman Rakshit and Sven Wagner for contributions. ### OVERVIEW * Non-Gaussian smoothing kernels. * Important bug fix in linearK, linearpcf * Changed internal format of linnet and lpp objects. * Faster computation in linear networks. * Bias correction techniques. * Bounding circle of a spatial object. * Minkowski sum also applicable to point patterns and line segment patterns. * Option to plot marked points as arrows. * Kernel smoothing accelerated. * Workaround for bug in some graphics drivers affecting image orientation. * Bug fixes and improvements. * Version nickname: 'Responsible Gambler' ### NEW FUNCTIONS * anyNA.im Method for 'anyNA' for pixel images. * bc Bias correction (Newton-Raphson) for fitted model parameters. See also 'rex'. * boundingcircle, boundingcentre Find the smallest circle enclosing a window or point pattern. * "[.linim" Subset operator for pixel images on a linear network. * mean.linim, median.linim, quantile.linim The mean, median, or quantiles of pixel values in a pixel image on a linear network. * rex Richardson extrapolation for numerical integrals and statistical model parameter estimates. * weighted.median, weighted.quantile Median or quantile of numerical data with associated weights. ### SIGNIFICANT USER-VISIBLE CHANGES * linear networks The internal format of a 'linnet' (linear network) object has been changed. Existing datasets of class 'linnet' and 'lpp' are still supported. However, computation will be faster if they are converted to the new format. To convert a linnet object L to the new format, use L <- as.linnet(L). To convert an lpp object X to the new format, use X <- as.lpp(X). * density.ppp, Smooth.ppp New argument 'kernel' allows the user to specify the smoothing kernel. * density.ppp, Smooth.ppp Argument 'weights' can now be a pixel image. * MinkowskiSum, %(+)% Now accepts arguments which are point patterns or line segment patterns as well as windows. * sumouter New argument 'y' allows computation of asymmetric outer products. * linearKinhom, linearpcfinhom New argument 'normpower'. * rmh.default, rmh.ppm New arguments 'nsim', 'saveinfo'. * symbolmap, plot.ppp, plot.lpp New option: shape="arrows" * rcellnumber New argument 'mu'. * lengths.psp New argument 'squared'. * plot.linfun Now passes arguments to the function being plotted. * as.linnet.psp If the line segment pattern has marks, then the resulting linear network also carries these marks in the $lines component. * summary.owin, summary.im The fraction of frame area that is occupied by the window/image is now reported. * density.ppp, Smooth.ppp Computation accelerated by about 15% in the case where at='points' and kernel='gaussian'. * linearK, linearpcf Accelerated by about 40%. * pixellate.ppp Accelerated in the case where weights are given * density.ppp Accelerated in the cases where weights are given or 'diggle=TRUE' * dilation.ppp Improved geometrical accuracy. Now accepts arguments to control resolution of polygonal approximation. * discs New argument 'npoly'. Accelerated in some cases. * plot.pp3 New arguments 'box.front', 'box.back' control plotting of the box. * grow.rectangle New argument 'fraction'. * nnfun.lpp New argument 'k'. * bw.ppl New argument 'sigma'. * lppm New argument 'random' controls placement of dummy points. * rhohat.lpp New argument 'random' controls placement of dummy points. * quadrat.test.ppm Accelerated in the case where the original window is a rectangle. * kppm, mincontrast, cauchy.estpcf, lgcp.estpcf, matclust.estpcf, thomas.estpcf, vargamma.estpcf A warning about infinite values of the summary function no longer occurs when the default settings are used. * circumradius This function is now deprecated, in favour of 'boundingradius' * print.quad More information is printed. ### BUG FIXES 1. linearK, linearpcf, and relatives: These functions were sometimes greatly underestimated when the network had segments shorter than 10 coordinate units. [Bug introduced in spatstat 1.44-0, december 2015.] Fixed. 2. integral.linim, integral.linfun Results were slightly inaccurate because of a bias in the distribution of sample points. [Bug introduced in spatstat 1.41-0, february 2015.] Fixed. 3. intensity.ppm Result was incorrect for Gibbs models if the model was *exactly* equivalent to a Poisson process (i.e. if all interaction coefficients were exactly zero). [Bug introduced in spatstat 1.28-1, june 2012.] Fixed. 4. rSSI Sometimes terminated prematurely. [Spotted by Frederico Mestre.] Fixed. 5. perspPoints Crashed if the image Z contained NA (i.e. if Z was only defined on a subset of the bounding frame). Spotted by Guy Bayegnak. Fixed. 6. plot.ppp, plot.lpp Crashed if the argument 'shape' was given. Fixed. 7. plot.kppm Crashed if the model was not fitted by minimum contrast. Fixed. 8. superimpose Crashed if the argument was a 'solist' containing line segment patterns. Fixed. 9. Jest Crashed sometimes, depending on the shape of the observation window. [Spotted by Cenk Icos.] Fixed. 10. plot.studpermutest Crashed when the summary statistic was a multitype pair correlation function or multitype K function. [Spotted by Sven Wagner.] Fixed. 11. pool.anylist Crashed with a message about buffer size, if the list was longer than about 100 items. Fixed. 12. diagnose.ppm, plot.diagppm Crashed in some cases when cumulative=FALSE. Fixed. 13. leverage.ppm, influence.ppm, dfbetas.ppm Crashed sometimes with a message about wrong replacement length. [Spotted by Virginia Morera Pujol.] Fixed. 14. as.linnet.psp Crashed with marked segment patterns, if any segments were very short. [Spotted by Suman Rakshit.] Fixed. 15. stieltjes Returned NA if some values of f were not finite. Fixed. 16. plot.symbolmap If a new plot window was initialised, it was sometimes too small to contain the geometric figures (circles, squares etc) in the symbol map. Fixed. 17. plot.ppp, plot.im Ignored xlim, ylim. Fixed. 18. rhohat.lpp Ignored nd, eps. Fixed. 19. nnfun.lpp Print method gave incorrect information about the point pattern. Fixed. 20. "[.fv" The default plot formula was not updated. Fixed. 21. fitted.ppm The result was sometimes a 1-dimensional array rather than a numeric vector. Fixed.