|
Past 3.x - the Past of the Future
|
Version list
3.02
- New module for PCA of Elliptical Fourier shapes
- New module for size removal from distance measures (Burnaby's method etc.)
- Box's M test for equal covariance matrices now handles more than two groups
- Clustering results can be saved as NEXUS tree files
- Two-sample and several-sample univariate tests: Now accepts samples either in separate columns or in a single column
plus a Group column
- Bug fix, clustering: Data were corrupted when bootstrapping repeatedly
- Bug fix, Raup-Crick similarity: Missing data were treated incorrectly
- Bug fix, data pasted into spreadsheet were truncated if the spreadsheet had too few rows
- Conversion of value pairs to data matrix
- Calculation of row percentages
- Max number of group columns increased to six (useful for taxonomic distinctness)
- Added arc sine (asin) to Evaluate Expression module
3.03
- Bug fix: Insert rows and columns did not work in 3.02
- Bug fix: File extension had to be typed in when saving graphic
- Bug fix: Polynomial regression did not run
- Bug fix: Crash in LOESS when bootstrapping with missing data
- Bug fix: Symbols in plots were dashed when background grid was switched on
- New option for showing group names in many plots
- Added Number of polymorphic sites to Genetic sequence statistics
3.04
- One-sample t test and one-sample Wilcoxon test
- Test for single-sample proportion
- Pie and doughnut charts
- Font name and size can now be changed in Graph Settings
- Geographical distance measure
- Date/time conversion tool
- Edit buttons added to spreadsheet tool bar
- Bug fix: Pearson's R correlation inaccurate for very large data sets
- Bug fix: Graphics overwrote part of text on some computers
- Bug fix: Concentration ellipses did not show correctly on some computers
- Bug fix: Select all from Edit menu did not work
- Very large SVG paths did not read correctly in Illustrator - Past now splits them into pieces
3.05
- F and t tests from parameters
- REDFIT now does not start until hitting "Compute" button
- Raup-Crick similarity now runs 50% faster
- User control over color map in matrix plot, gridding etc.
- New export file formats: NTSYS, TPS, RASC, RTF
- Sum added to univariate summary statistics
- Bug fix: Copy data did not work in some modules.
- Bug fix: MANOVA did not update when constraints value was changed
- Bug fix: Crash in spherical analysis/stereoplot
- Bug fix: PCA loadings table showed raw loadings, not correlations when correlations selected in loadings plot
- Bug fix: Filled convex hulls had no transparency
- Bug fix: Broken Undo after Paste
- Bug fix: Nonlinear fit: von Bertalanffy curve correct, but small typo in the equation shown above the plot
- Bug fix: Coefficient of variation given as 0 when mean was negative
- Improved handling of non-integer data for the Diversity module
- Small fix in spatial autocorrelation
3.06
- Clopper-Pearson confidence interval for single proportion
- Confidence interval for many proportions, with plotting
- Two-way repeated measures ANOVA
- Two-way ANOSIM without replication
- Stacked bar and area chart
- Missing data supported in group variable for two-sample univariate tests etc.
- Confidence interval for sample mean in one-sample t test
- Wavelet analysis: Sampling interval can now be specified
- Bug fix: Faulty selection of area in the spreadsheet
- Bug fix: All columns were included in many analyses when disjunct columns were selected
- Bug fix: Occasional crash in bivariate linear regression module
- Bug fix: Find (search) function did not work
- Bug fix: Garbled selection boxes in the REDFIT module
- Bug fix: Discriminant analysis: Too many axes generated when number of groups exceeded number of variables+1
- Bug fix: Crash in one-way ANOSIM if a group contained only one row
- Bug fix: Landmark PCA deformation expansion factors not shown for negative PC scores
3.07
- 3D scatter and 4D bubble plot
- Mann-Kendall test for trend
- Fitting of species abundance along gradient to Gaussian packing model
- Mantel-Cochran-Haenzsel test on multiple 2x2 contingency tables
- Confidence interval on stratigraphic range
- Bug fix: NMDS gave inaccurate results if environmental variables were included (introduced in 3.0)
- Bug fix: Pairwise Mahalanobis in MANOVA gave 0 instead of 'fail' for errors
- Bug fix: Procrustes analysis overwrote group columns
- Bug fix: K-means clustering failed for very large numbers in input
- Bug fix: Missing numerical output in abundance distribution modelling
- Bug fix: Excessive computation times in some modules in Mac version
- Save image added to surface plot
- Improved robustness of Gaussian fit in nonlinar fit module
3.08
- Powerful scripting language, to be extended in future versions
- Mood's median test for two samples
- Unconditional confidence interval for individual rarefaction
- Tamura genetic sequence distance
- Option for using group column for box plot
- Sorting moved to menus, and other improvements in sorting
- Bug fix: Discriminant analysis showed blank plot for data with 3 groups
- Bug fix: Linear regression crashed if started twice (introduced in 3.07)
- Bug fix: Occasional crash on program exit (introduced in 3.07)
- Bug fix: Column data type menu could get garbled (introduced in 3.07)
- Bug fix: Small inconvenience in Row colors/symbols
3.09
- ARMA and intervention analysis for time series
- Tests for equality of regression slopes in bivariate fitting (OLS, MA and RMA)
- Insolation model, 250 Ma to Recent (La2004)
- Copy-paste now works also on graphics, but only bitmap format (you must still export to
SVG for publication quality)
- New in Transform menu: Convert to ranks
- Many new functions in the scripting language
- Bug fix Unitary Associations: Copy-paste did not work for some output
- Bug fix Moran’s I: Incorrect significance line for very large sample sizes
- Bug fix: Hidden windows now pop to front when selected again from menu
3.10
- Circular spectral analysis for point events
- Equal axes button in graph preferences
- Linear regression now reports standard error of estimate
- Scripting: New functions, including user-defined windows (forms) with labels, buttons, edits, check boxes.
- Increased max number of items in clustering from 900 to 3500
- Improved algorithm for single linkage clustering of large data sets
- Drag-drop files onto spreadsheet
- Bug fix: Remove column or row did not work after moving it
- Bug fix: Freeze after error condition in Evaluate expression
- Bug fix for Mac OSX: Copy-paste into spreadsheet did not work properly
- Minor algorithm improvement in Unitary Associations
- New application icon
3.11
- New: Interface to Google Maps, with plotting of points, polygons and lines (Windows only)
- New: “Graphics basket” for managing and combining plots
- Bug fix: Copy-paste from Excel broken under Mac OSX
- New: Nested ANOVA
- Improved: Selection of file format for graphics export
- New: Support for reading RASC files
- Small fixes in Unitary Associations
- New: Pooled and unpooled z ratio for risk difference
- Replace now only operates in selected cells
- Omega2 in ANOVA set to zero if negative
- User-defined N for Monte Carlo normality tests
- Sinusoidal fit: Frequency resolution increased from 1/1000 to 1/500,000 of Nyquist frequency
- Thin-plate spline deformation grid can now be extended in x and y directions
- Scripting: New functions
- Beta diversity now takes samples in columns, not rows, to be consistent with other modules
- Bug fix: Bivariate linear regression: Slope was not shown in the Statistics tab for RMA or MA.
- Bug fix: Save data in other than the Past format broken in some recent version
- Bug fix: 3D scatter plot crashed when axis ranges were very different
3.12
- New “detach tab” button allows viewing several tabs (subwindows) at the same time
- PCA of 3D landmarks, with 3D plotting
- Use your own images as symbols in plots
- Arrow plot, alternative to rose plot in circular analysis
- Plotting of residuals in bivariate linear regression
- Combined box and jitter plot
- Plotting of displacement vectors in thin-plate spline deformation module
- Contingency table analysis now includes table of residuals etc.
- Z and t tests for probability of a single observation taken from a given sample
- Scripting: New functions
- Bug fix: Wrong labels on points in ternary bubble plot
- Bug fix: Occasional error message when exiting program on Mac
- Numerous small improvements in user interface
3.13
- Graphics can now be exported also in the PDF format
- Bug fix: Excessive computation time for two-sample tests on Mac OSX
- Bug fix: Mann-Kendall test, incorrect p values for negative trend and n<11
- Bug fix: Anderson-Darling normality test, incorrect p values for extremely large data sets
- Bug fix: Cluster analysis, save Nexus file, faulty when using comma decimal separator
- Max string length for x axis labels increased to 20
3.14
- New: Point pattern statistics for counts in quadrats (Morisita’s index etc.)
- New: Dunn’s post hoc test for Kruskal-Wallis (in addition to pairwise Mann-Whitney)
- Bug fix: Export graphic file type selection did not work on Mac OSX
- Bug fix: Search and replace stopped after first hit
- Bug fix: Group labels not shown for small groups (n<3) in x-y plots
- Bug fix, NMDS: Missing values (?) in environmental variables not handled correctly
- REDFIT: Maximal oversampling increased to 64
- Neighbour joining: Root (outgroup) can now be selected in menu
- Normality tests now recognize samples marked as groups
- Survival analysis: Censoring events can now be plotted on the Kaplan-Meier curves
3.15
- Tukey’s post-hoc tests for two-way ANOVA
- Permutation test for correlation and one-way ANOVA
- Mosaic plot for 2-way and 3-way contingency tables
- Stem plot
- Distribution-free confidence intervals for stratigraphic ranges
- Max N for wavelets increased to 130,000
- Improved histogram plots
- Univariate summary statistics now supports groups
- New in Regular interpolation: Option for antialiasing (sinc) interpolation
- Sorting now supports both types of decimal separator
- Remove uninformative rows/columns: New option to remove constant-valued rows/columns
- Bug fix, Ward’s method clustering: Missing data not handled properly
- Bug fix: Setting row colors/symbols required user to select some data in addition to a group column
- Bug fix: Scatter plot labels sometimes disappeared at edges of plot
- Bug fix, CCA: Table of scores did not update when changing the scaling type