Skip to content

Latest commit

 

History

History
103 lines (69 loc) · 3.9 KB

NEWS.md

File metadata and controls

103 lines (69 loc) · 3.9 KB
Raw

sppt 0.2.3

  • Added generalized S-index

  • Touch-ups to vignettes, helper functions, and documentation due to generalized S-index

sppt 0.2.2

  • Minor updates to code to ensure compatibility with dplyr 1.0.0 and higher.

  • Renamed function summary.sppt to summary_sppt to remove confusion wrt S3method.

sppt 0.2.1

  • Added checks for SpatialPolygons and SpatialPoints objects. For the spatial units of analysis, SpatialPolygons can now be used as they are automatically coerced to SpatialPolygonsDataframe. If provided objects are not of type SpatialPolygons or SpatialPoints, functions stop and give error message.

sppt 0.2.0

Important changes

  • Function sppt_diff() changed to have "Chi2_Nmin1" as the default test. This means that a function call to sppt_diff() will give different results than the same call up to version 0.2.0, as the latter had "Yates" as default.

  • Function sppt_diff() updated to have a variety of tests.

  • Function sppt_diff() now uses the p-values throughout (instead of confidence intervals) to decide on statistical significance. This makes it easy to adjust the p-values (and decisions to reject the null) for multiple comparisons, the default setting.

  • Function sppt_diff() no longer uses exact2x2 package as fisher.test reports the same p-value (only the confidence intervals were different, which are no longer used.)

  • Arguments of sppt_boot() changed. Instead of 'replacement' now 'bootstrap' is used. 'percpoints' default is set to 85, but is ignored if bootstrap = TRUE. This means that in the default behavior, a bootstrap sample (i.e. 100% of points are samples with replacement) is drawn for both Base and Test. If bootstrap = FALSE, the percpoints (default 85) is used in a subsampling method for both Base and Test (i.e. as in Andresen's original test, but also for Base).

  • Added argument 'bootstrap' to sppt(). In the default behavior, bootstrap = FALSE and sppt() will be Andresen's original test (i.e. subsampling Test only with percpoints sampling without replacement). When bootstrap = TRUE, 'percpoints' is ignored and a bootstrapped sample (i.e. 100% of Test points sampled WITH replacement) is used for Test.

Misc

  • Vignette sppt renamed into sppt_intro

  • Vignette sppt_diff updated with new information.

  • LICENSE file added

  • README updated

sppt 0.1.6

  • Adjusted function sppt_diff() to use exact2x2::exact2x2() instead of stats::fisher.test as the former matches the different two-sided conditional exact tests for 2x2 tables with the appropriate confidence intervals. See https://cran.r-project.org/package=exact2x2 for details.

  • Added information on the exact2x2::exact2x2() to vignette 'Proportion difference tests'.

  • README updated

  • Moved helper functions such as summary.sppt() to separate R file.

sppt 0.1.5

New features

  • New function sppt_diff() added. This function calculates the difference in proportions between Base and Test points using standard 'proportion difference' tests. Having zero points in an areal unit will not lead to a confidence interval of 0% to 0% (as in sppt() and sppt_boot()), making is a more conservative test. See Wheeler, A., Steenbeek, W., & Andresen, M.A. (2018). Testing for Similarity in Area-Based Spatial Patterns: Alternative Methods to Andresen's Spatial Point Pattern Test. (SSRN: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3111822)

  • New function sppt_boot() added. This is a full bootstrapping procedure to measure the difference in proportions between Base and Test points. In contrast to the original sppt(), the results of the test are not dependent on which dataset is selected as Base and which is selected as Test.

Misc

  • Changed README.md and vignettes to facilitate easier building of vignettes on local machine.

sppt 0.1.4

  • First published version of the sppt package. Includes function sppt().