# RELEASE NOTES ## spatstat 1.55-0 ### 29 January 2018 We thank 'AdriMaz' and Nicholas Read for contributions. ### OVERVIEW * Lurking variable plot for models fitted to several point patterns. * Improvements to code for class 'mppm'. * Improvements to leverage and influence diagnostics. * Improved summary information for funxy and distfun objects. * Bug fixes and improvements. * Removed old warnings and deprecated functions. * Nickname: "Stunned Mullet" ### NEW FUNCTIONS * contour.leverage.ppm Method for 'contour' for leverage functions of class 'leverage.ppm' * lurking New generic function for lurking variable plots. * lurking.ppp, lurking.ppm These are equivalent to the original function 'lurking'. They are now methods for the new generic 'lurking'. * lurking.mppm New method for class 'mppm' Lurking variable plot for models fitted to several point patterns. * print.lurk Prints information about the object returned by the function 'lurking' representing a lurking variable plot. * model.matrix.mppm Method for 'model.matrix' for models of class 'mppm'. * test.crossing.psp, test.selfcrossing.psp Previously undocumented functions for testing whether segments cross. ### SIGNIFICANT USER-VISIBLE CHANGES * predict.ppm Now recognises the arguments 'dimyx' and 'eps' for specifying the resolution of the grid of prediction points. * leverage.ppm, dfbetas.ppm Increased the default resolution of the pixel images. Spatial resolution can now be controlled by the arguments 'dimyx', 'eps'. * ppmInfluence The result now belongs to class 'ppmInfluence', for which there are methods for 'leverage', 'influence', 'dfbetas' which extract the desired component. * plot.leverage.ppm New argument 'what'. * persp.leverage.ppm New arguments 'zlab' and 'what'. * as.im.leverage.ppm New argument 'what'. * summary.funxy, summary.distfun Printed information now includes a summary of the function values. * lurking.ppm Accelerated. * "[.psp" Accelerated. * clf.test, conspire, bounding.box, ksmooth.ppp, mpl, superimposePSP, eval.hyper, smooth.fv, smooth.ppp, smooth.msr, rtoro, plot.kstest These deprecated functions have now been removed. * bermantest This deprecated function has now been removed. Use berman.test instead. * kstest This deprecated function has now been removed. Use cdf.test instead. * plot.ppp A very old warning, about the interpretation of the mark scale as the circle diameter, is no longer printed. ### BUG FIXES 1. nnmap, nnmark Values were incorrect if the resulting pixel image had unequal numbers of rows and columns. Fixed. 2. vcov.mppm Format was incorrect (rows/columns were omitted) in some cases. Fixed. 3. model.matrix.ppm, model.frame.ppm Values were sometimes incorrect when applied to the result of subfits(). To be precise, if 'fit' was an mppm object fitted to a hyperframe that included 'design covariates' (covariates that take a constant value in each row of the hyperframe), and if 'futs <- subfits(fit)', then model.matrix(futs[[i]]) gave incorrect values in the columns corresponding to the design covariates. Fixed. 4. model.matrix.ppm The attribute 'assign' was omitted, in some cases. Fixed. 5. simulate.dppm, simulate.detpointprocfamily In dimensions higher than 2, the result was shifted so that it was centred at the origin. Fixed. 6. Smooth.ppp Crashed if geometric=TRUE and there were several columns of marks. Fixed. 7. simulate.dppm, simulate.detpointprocfamily Crashed if nsim > 1 and the spatial dimension was not equal to 2. Fixed. 8. plot.leverage.ppm Contour line was annotated, which was not intended. Fixed. 9. leverage.ppm The leverage function was oversmoothed, when the model was fitted with method="logi". Fixed.