LATEST UPDATE Posted by Ana Severiano and João Carriço 06.10.2011
- The Adjusted Wallace coefficient is now calculated by default. The Wallace coefficient is calculated when selected by the user.
Posted by Ana Severiano and João Carriço 21.09.2011
Posted by Ana Severiano and João Carriço 26.08.2011
- Check our recently published research paper:
- A graphic representation of results is now available by clicking "Diagram" above the results table. Please see the tutorial for an example on how to read the diagram.
- The user can now calculate a p-value when dealing with results from different data sets. This tool is found under Other Tools separator. See the tutorial for even more information.
Posted by Ana Severiano and João Carriço 29.03.2011
- The user can now calculate the Adjusted Wallace, a new measure corrected for chance agreement. Please see the tutorial for more information.
- Updated BioNumerics™ scripts are also available at Downloads.
Posted by Ana Severiano and João Carriço 16.02.2011
- 95% Confidence Intervals are now calculated using the jackknife pseudo-values method. The calculation of 95% Confidence Intervals using boostrap is no longer available by default. The user can have access to a version where the bootstrap calculation is possible.
- The p-values are now also calculated according to the jackknife pseudo-values method.
- Please read the tutorial for more information about these changes and for the alternative version link.
Posted by Ana Severiano and João Carriço 30.08.2010
- The user can now access how statistically significant is the difference between the calculated values using a p-value.
- Some warning messages are now shown when displaying the results (for example, when the 95% CI for Wallace includes the value in case of independence - Wi)
Posted by Ana Severiano and João Carriço 23.06.2010
- Some improvements in the layout of the web-pages have been made.
- You can now click "print version" (top right) to see a printer-friendly page.
- No changes have been made in the calculations.
Posted by Ana Severiano and João Carriço 11.12.2009
- You can now share this site with co-workers and friends! Just hit the button!
- A bug was found and fixed in the display of Wi (expected Wallace in the case of independence). This does not affect other coefficients calculation. Please confirm your previous Wi values.
- A bug was found and fixed in the calculation of Variation of Information upper limit. This does not change the coefficient value, but only the cell color in the displayed table.
- Please Contact us if you notice any other bugs or if you have suggestions.
Posted by Ana Severiano and João Carriço 28.10.2009
- New coefficients can now be calculated: Shannon's Index of Diversity, Mirkin Metric, Jaccard Index, Fowlkes and Mallows, Mutual Information, Normalized Mutual Information and Variation of Information. Although we recomend the use of Simpson's Index of Diversity as diversity measure and Adjusted Rand and Wallace coefficients for comparing partitions, we can not ignore the importance of a number of other coefficients in diferent research fields. We recommended caution when choosing the best coefficient for your research. For more information about these coefficients please read the tutorial and check the references section.
- The calculation of 95% Confidence Intervals using Bootstrap with Bias Corrected and accelerated method (BCa) is no longer available. However, you can still calculate the approximated intervals by Bootstrap using the percentile method. We removed it, since it's results were equivalent of those of the Percentile method and other bootstrap methods tested (such as Normal approximated interval with bootstrapped bias and standard error method or the bias corrected percentile method). We tested the performance of both bootstap methods (BCa and percentile) to calculate the 95% confidence intervals for Rand, Adjusted Rand and Wallace. (Results not shown and to be published).
- Tutorial section was extended and updated.
- Contact us section: Now you can access to our wikipage to report bugs, send us comments, ask for help or propose new features.
Posted by Ana Severiano and João Carriço 14.08.2009
- Non-approximated Simpson's 95% Confidence Interval
- 95% Confidence Interval calculation for Rand, Adjusted Rand and Wallace by bootstrap (Biased correction and acceleration methods and Percentile method
- Graphical bootstrap distribution visualization
- Tutorial updated
Posted by João Carriço 29.05.2007
- CSV data parsing: allows the use of comma separated values in the input area
- Data validation: detecting missing data in the text input
- Histograms: visualize your data distribution for each method
- Contigency tables: visualize your data distribution for each combination of two methods