Softpedia >Windows >Office tools >Other Office Tools >XLStat >  Changelog

XLStat Changelog

What's new in XLStat 2023.5.1 Build 1396

Feb 6, 2023
  • Friedman-Rafsky test:
  • The Friedman-Rafsky test has been added to our list of non-parametric tests. It allows to compare the distributions of two samples of quantitative data when they are described by more than one attribute/variable.
  • The interface gives you the choice between 4 different distances (Canberra, Chebychev, Euclidienne and Manhattan) and 3 algorithms to find the minimum spanning tree according to your data set (Chazelle using Soft-Heap, Kruskal and Boruvka).
  • An easy example: you know the weight and height of a Basketball team and a Football team. Use this test to determine if these two samples (teams) follow the same distribution. Click here to learn more.
  • This feature is available in all XLSTAT solutions (except Basic)
  • Similarity/Dissimilarity Matrices:
  • This feature allows you to calculate the similarities or dissimilarities between your observations or variables. XLSTAT offers many proximity measures to which we have added, in the context of binary data, two popular indices in supervised and unsupervised classification: the Rand and Adjusted Rand indices. These indices can be used to build a matrix before running a classification.
  • Moreover, it is now possible to enter large-volume data thanks to the file importation mode in XLSTAT.
  • ANOVA:
  • Several improvements have been added to our ANOVA functionality.
  • Improvement 1: Addition of groups
  • It is now possible to select a group variable in the General tab when your observations are associated with a group. This can be very helpful if your data is split into many sub-samples and you want to run an ANOVA for each subsample simultaneously.
  • Improvement 2: Add effect size measures
  • P-values tend to decrease when the sample size is increased. This may lead to falsely significant results in the case of large samples and thus to “wrong” interpretations.
  • For this reason, we have added effect size measures. They do not dependent on the sample size and allow better conclusions.
  • Improvement 3: Adding stars to the p-values
  • In order to quickly visualize the significance levels of the p-values, we have added stars to the ANOVA result tables.
  • Improvement 4: Sorting in the means charts
  • For an easier comparison of means, we have added an option in the Means sub-tab that allows you to sort means in descending order.

New in XLStat 2022.4 (Oct 20, 2022)

  • Sample size calculator for sensory discrimination tests in Excel
  • CATATIS method - an improvement of the usual CATA data processing:
  • Ability to have sessions
  • Tests of the subjects' weights
  • Consistency/significance tests of the subjects' homogeneity
  • Inter-subject Repeatability Index

New in XLStat 2022.3 (Jul 21, 2022)

  • Added new possibilities for data modeling and predictions using the
  • following methods:
  • Lasso Regression
  • Ridge Regression
  • Elastic Net Regression

New in XLStat 24.2.1289.0 (Apr 14, 2022)

  • Agglomerative Hierarchical Clustering: New truncation methods, extra
  • graphs, the ability to color dendrograms according to groups.
  • Statistical Process Control: New functionalities, more ergonomic UI, the ability to
  • select variables from a drop-down list or to call a dataset from a txt/csv file.

New in XLStat 22.5.1061 (Dec 12, 2020)

  • New features, advanced options & improvements:
  • Support Vector Regression
  • Clustatis
  • Screening Designs
  • K Nearest Neighbors (KNN)
  • Naive Bayes Classification

New in XLStat 22.5.1043 (Nov 5, 2020)

  • CLUSCATA (available in XLSTAT Sensory and Premium)
  • A new method for clustering assessors based on their perceptions of products is available. CLUSCATA can be seen as an adaptation of CLUSTATIS for CATA data. An interesting option is the creation of a "K+1" class in order to set aside assessors who do not conform to any class.
  • Access this new feature under the Sensory Data Analysis menu.
  • CATARACT (available in XLSTAT Sensory and Premium)
  • The CATARACT (CATA Rejection and ACceptation Tests) procedure offers a description of CATA surveys and allows you to extract information relative to the quality of the questionnaires. As part of this procedure, three new options are available in the CATA dialog box:
  • CATA data validation: to check the quality of the CATA data.
  • Independence of attributes: to determine whether the same attributes are checked by assessors for each product.
  • Handling of multiple sessions: to deal with unexpected cases where assessors evaluate a product several times.
  • Access this new feature under the Sensory Data Analysis menu.
  • Regression & Classification Trees (available in all XLSTAT solutions except Basic)
  • Users can now choose among three ways to define the tree parameters (parent size, son size, depth):
  • Manually enter the values of each parameter
  • Let XLSTAT search for the optimal values based on K-fold cross validation
  • Set a range of values and let XLSTAT choose the best combination of parameters.
  • Access this feature under the Machine Learning menu.
  • Nonlinear Regression Models (available in all XLSTAT solutions except Basic)
  • Boosted outputs and graphs:
  • 95% lower and upper bounds added to the model parameters table.
  • 95% confidence and prediction intervals curves displayed in the regression chart.
  • Four new enzyme kinetic equations (competitive, non-competitive, uncompetitive and mixed inhibition)
  • Access this feature under the Modeling Data menu.
  • Missing-data Imputation(available in all XLSTAT solutions)
  • New techniques and a better management of missing values:
  • Users can now simultaneously deal with quantitative and qualitative missing data.
  • The EM algorithm has been implemented for quantitative missing data.
  • The NIPALS algorithm and replacement by a given textual value are available for qualitative missing values.
  • Access this feature under the Preparing Data menu.

New in XLStat 2020.4 Build 22.4.1024 (Sep 16, 2020)

  • Support Vector Regression (available in all XLSTAT solutions except Basic):
  • Support Vector Machine (SVM) has been generalized and can be applied to regression problems in addition to classification ones. This method is commonly used in time series prediction. Our revamped interface now allows you to choose the best suited response variable type for your analysis (binary/multi-class classification or regression).
  • Access this feature under the Machine Learning menu.
  • Clustatis (available in XLSTAT Sensory and Premium):
  • A new clustering method for multiblock datasets, commonly used in sensory and consumer studies (e.g. projective mapping) has been added. Use this feature, to perform a cluster analysis of consumers or judges on the basis of their perceptions of products. This new feature is particularly useful when your data includes atypical configurations.
  • Access this new feature under the Sensory Data Analysis menu.
  • Screening Designs (available in XLSTAT Quality, Life Sciences and Premium):
  • Our revamped interface will allow you to easily generate an experiment design. It is now possible to select the data in a single tab instead of several ones. A new table has been added to the report sheet allowing you to define the optimization settings of the analysis or use the default software ones.
  • Access this feature under the DOE menu:
  • K Nearest Neighbors (KNN) (available in all XLSTAT solutions except Basic)
  • Users can now either manually define the number of neighbors or use the K-fold cross validation option to identify the optimal number of neighbors from a range of values. The model with the minimum cross-validation errors will apply. A summary table for each tested K, and the related chart will be displayed.
  • Access this feature under the Machine Learning menu.
  • Naive Bayes Classification (available in all XLSTAT solutions except Basic):
  • A Prediction tab has been added in the dialog box to allow you to select the variables of the prediction sample. As a result, you no longer need to select the prediction set in the General tab.
  • Access this feature under the Machine Learning menu.

New in XLStat 2020.1.3 Build 65333 (Apr 22, 2020)

  • Clustering after dimensionality reduction (available in all XLSTAT solutions):
  • A new button has been added to the XLSTAT output sheet which allows you to quickly run an Agglomerative Hierarchical Clustering (AHC) after conducting a Principant Component Analysis (PCA) or other dimension reduction methods (MFA, FA, MCA, etc).
  • 2-sample t-test and z-test (available in all XLSTAT solutions):
  • You can now select multiple columns of sample identifiers and carry out the independent test for each of the sample identifiers in one shot. Missing values in your data are no longer a limitation as an option for removing missing observations (tab Missing data) has been added.
  • This feature is accessible under the Parametric tests menu.
  • Ternary graphs with groups (available in all XLSTAT solutions):
  • Color observation points within your Ternary plot based on the group they belong to. Use the new field of the Ternary dialog box, named Groups, for this purpose.
  • Available under the Visualizing data menu.
  • Youden Plots (available in XLSTAT Life Sciences and Premium):
  • This new method validation tool consists of a bivariate scatter plot in which limits of acceptability and outliers are plotted. Commonly used in Laboratory Medicine to evaluate the performance of interlaboratory test procedures. XLSTAT implements methods as described in ISO 13528-2015-10 (algorithm A).
  • Access this feature under the Method Validation menu.
  • 4 or 5 parameter logistic regression (available in XLSTAT Life Sciences and Premium):
  • Powerful and faster computations are achieved due to the improvement of the Levenberg-Marquardt algorithm. Addition of the graph Residual values versus Fitted values for verifying the assumption of randomly distributed residuals with a homogeneous variance.
  • Access this feature under the Dose effect analysis menu.
  • STATIS / Free sorting (available in XLSTAT Sensory and Premium):
  • A chart has been added to project each object of each configuration on the axes determined by STATIS. In a sensory data analysis, this graph helps you evaluate, for a given product, how close the assessors are to each other and which ones stand out.
  • Access this feature under the Sensory data analysis menu.
  • Up and down arrow buttons (available in all XLSTAT solutions):
  • Convenient arrows now allow you to quickly move up and down any XLSTAT report sheet.

New in XLStat 2020.1 Build 64377 (Feb 17, 2020)

  • A smart predictive tool that:
  • Identifies the appropriate models depending on your data type
  • Performs a comparison of the fitted models
  • Makes predictions of new observations
  • Discover this feaure under the XLSTAT.ai menu.
  • ANOVA (available in all XLSTAT solutions):
  • Our algorithm was optimized leading to a significant decrease in computational time. Run an ANOVA on large data sets including post hoc tests in just a few seconds. In addition, the Restricted ANOVA has been implemented. This model is useful when we have an interaction between a fixed factor and a random factor and assume that the sum of the interaction effects on the levels of the fixed factor is zero.
  • This feature is accessible under the Modeling data menu.
  • RANDOM FORESTS (available in all XLSTAT solutions except Basic):
  • It is now possible the split your data set into a training and a validation sample in order to confirm the fitted model. Available under the Machine learning menu.
  • TESTS ON CONTINGENCY TABLES (available in all XLSTAT solutions):
  • The p-values of Fisher's exact test of independence are now displayed per cell of the contingency table. Available under the Association tests menu.
  • FREE SORTING DATA ANALYSIS (available in XLSTAT Sensory and Premium):
  • Free Sorting tests are commonly used for the sensory characterization of products. Data obtained from such studies can be analyzed using our new feature with methods such as STATIS, MCA and CA. Access this feature under the Sensory data analysis menu.
  • TAGUCHI DESIGNS (available in XLSTAT Quality, Life Sciences & Premium):
  • Dynamic designs have been added to allow users to determine the best levels of control factors in order to improve the relationship between signal factors and output responses. A new option for adding interactions to the model is also available. This feature is available under the Designs of Experiments menu.
  • MANAGE ARCHIVED FILES TOOL (all XLSTAT solutions):
  • Automatically back-up your XLSTAT analyses and load previously saved files. This new tool is still in beta testing and is only available for Windows. Discover it under the Tools XLSTAT menu.

New in XLStat 2019.4.1 Build 62958 (Nov 26, 2019)

  • Demšar Significance Diagram:
  • Use the Demšar graph, also called Critical Differences plot, to visualize pairwise differences after performing a Friedman test followed by a post-hoc procedure. This feature can be found under the Visualizing data XLSTAT menu.
  • Latent Semantic Analysis (XLSTAT Marketing and Premium):
  • Often used in document classification or cross language retrieval, this text mining tool will help you to discover hidden semantics of words from a large set of documents. LSA uses a document-term matrix as input data which can be generated through the XLSTAT Feature Extraction. This feature is accessible under the Text mining XLSTAT menu.
  • Customer Long-term value (XLSTAT Marketing and Premium):
  • Understand your customers life cycle, identify periods of high churn risk, and estimate the profits generated by your customers over an extended time period. Here, compared to the classic CLV model, the retention rate varies over time and cash flows depend on the time of cancelation. Accessible under the Marketing Tools XLSTAT menu.
  • Taguchi experimental designs (XLSTAT Quality, Life Sciences and Premium):
  • Taguchi method is a popular engineering technique to optimize processes and improve the quality of products. It provides an improvement to full and fractional factorial designs. XLSTAT offers to study three parameters: the signal to noise ratios, the means and the standard deviations of the measured attributes.
  • Design of sensory discrimination tests (XLSTAT Sensory and Premium):
  • Our algorithm used to generate designs of sensory discimination tests (e.g. triangle, tetrad, etc) has been revised and improved. An optimal design can be now achieved.

New in XLStat 2019.3.2 Build 62451 (Nov 6, 2019)

  • In all XLSTAT solutions:
  • Parallel Coordinates Plot
  • New graphic capabilities are available for a more interactive visualization -
  • Display different scale values of quantitative variables on the Y axis
  • Use different line style and legends for descriptive statistics
  • Possibility to highlight a specific group on the chart
  • This feature can be found under the Visualizing data XLSTAT menu.
  • In XLSTAT Marketing & Premium:
  • MaxDiff Analysis
  • Including too many questions when designing a MaxDiff survey may lead to respondent fatigue or confusion. The latest XLSTAT version offers the possibility to use a different question set per group of respondents for your MaxDiff analysis.
  • This feature can be found under the Conjoint Analysis XLSTAT menu.
  • In XLSTAT Sensory & Premium:
  • Preference Mapping
  • The detection of possible ideal points and admission zone has been added. In addition, it is now possible to perform a preliminary transformation of the data based on PLS regression on the top of PCA and standardisation options.
  • Panel Analysis
  • The CAP (Control of Assessor Performances) table has been added. Display this table to synthesize the majority of the Panel analysis results and obtain a high amount of information on product discrimination, as well as on the agreement and repeatability of judges.
  • CATA Analysis
  • A new multiple pairwise procedure, based on Critical Difference (Sheskin), can be chosen when running Cochran's Q test on the Assessors x Products table. The results are displayed in a separate table for each attribute.
  • These features can be found under the Sensory data analysis XLSTAT menu.
  • In XLSTAT Life Sciences & Premium:
  • Life Table, Kaplan-Meier, Nelson-Aalen analyses
  • Possibility to import high volume csv or txt files and select your data from a list of variables.
  • The survival curves are now drawn as a step function in Life table analysis.
  • These features can be found under the Survival analysis XLSTAT menu.

New in XLStat 2019.2.2 Build 59417 (Jun 25, 2019)

  • Matrix operator:
  • A handy tool for engineers, scientists or teachers or anyone who want to carry out operations on matrices. The following operations can be performed:
  • Matrix Multiplication
  • Matrix Addition
  • Matrix Subtraction
  • Matrix Inversion
  • Transposition
  • The Matrix Operator tool can be found under the Mathematical Tools XLSTAT menu. Available in all XLSTAT-Solutions.
  • Cochran Q test:
  • The Critical difference (Sheskin) method has been added for pairwise comparisons. In addition, it is now possible to display a proportions chart as well as a distribution chart.
  • Cochran’s Q test can be accessed under the Nonparametric tests XLSTAT menu. Available in all XLSTAT-Solutions.
  • Passing Bablok:
  • A new estimation method (Part III) has been added. This method developed by Bablok et al. in 1988 is an improvement of the method known as Part I. It is more robust and can be used to compare two methods measured on different scales with possibly a negative correlation between X and Y.
  • In addition, you can now import large text file (csv or txt) and run a Passing Bablok regression on millions of data. Simply launch the Passing Bablok dialog box and click on the mouse button to change the data selection mode.
  • Passing Bablok regression can be accessed under the Method and Validation XLSTAT menu. Available in XLSTAT-Biomed and XLSTAT-Premium.

New in XLStat 2019.1.1 Build 56421 (Feb 8, 2019)

  • Data Selection: Run a Linear Regression or an ANOVA on millions of data points. Available under the Modeling data menu.
  • Test assumptions: Validate the hypothesis of normality and homogeneity of variances. Available under the Modeling data menu.
  • Influence diagnostics: Compute DFBetas, Mahalanobis distance and other influence statistics. Available under the Modeling data menu.
  • Variables characterization: Extra filtering and sorting options for a more customized output. It is also possible to use decimal weights in parametric tests. Available under the Describing data menu.
  • Nonlinear regression: New interface and additional built-in models which can be used in various fields such as pharmacology. Available under the Modeling data menu.
  • Scatterplots: Customize your graph choosing your own color for groups. Available under the Visualizing data menu.
  • Multiple Factor Analysis: You can now run this analysis on frequency tables such as count data of species. Available under the Sensory dataanalysis menu.
  • STATIS: This method can be particularly used in the case of projective mapping, conventional profiling, free choice profiling. Available under the Sensory data analysis menu.
  • DOE sensory: Improved designs for all sensory discrimination tests are now available. Available under the Sensory data analysis menu.
  • Customer Lifetime Value: A new and useful feature to assess the financial value of your customers. Available under the Marketing Tools menu.
  • Price Elasticity of Demand: Determine the price of your products at which the maximum revenue is generated. Available under the Marketing Tools menu.

New in XLStat 2018.7 Build 54867 (Nov 23, 2018)

  • Analytic Hierarchy Process:
  • Use this decision aid method to analyze complex multi-criteria problems. The method used is to simplify the problem by breaking it down into a hierarchical system. Thomas Saaty created it in the 1970s. Available under the Decision Aid menu.
  • Load large CSV/text files:
  • You can now upload large data files which exceed the standard Excel worksheet size (1,048,576 rows by 16,384 columns in Excel 2016). This new data import tool is currently available within the Data Management Meature. Available under the Preparing Data menu.
  • Classification Trees:
  • The Information Gain (entropy) quality measure, used to split a node, has been added along with the complexity parameter (CP) stop criterion for C&RT classification trees. Available under the Machine Learning menu.

New in XLStat 2018.6 Build 53533 (Oct 1, 2018)

  • New statistical features and options:
  • Fuzzy K-means: An unsupervised clustering algorithm for organising large data sets into groups. It can be used for document classification and many other applications. Available in all XLSTAT solutions under the Machine Learning menu.
  • Conjoint Survey Design: It is now possible to generate different question sets for different respondent groups in conjoint surveys. Available in XLSTAT-Marketing under the Conjoint Analysis menu.
  • Johnson Transformation: Transform your data into a normal distribution using the Johnson method. It can be used for data including zero and negative values. Available in all XLSTAT solutions under the Preparing Data menu.
  • Temporal Dominance of Sensations: New options are available for defining either automatically or manually the smoothness of a TDS curve. Available in the XLSTAT-Sensory under the Text Mining menu.

New in XLStat 2018.5 Build 51886 (Jun 27, 2018)

  • New statistical features and options:
  • Factorial analysis of mixed data: Explore a data table composed of quantitative and qualitative variables using the PCAmix method.
  • ELECTRE 3: Commonly used in decision-making, this multicriteria analysis method classifies a set of solutions from the best to the worst.
  • Importing data into Excel: SPSS, Minitab, SAS and other data format files can be now imported into Excel in few clicks.
  • Comparison plots: Combine the power of boxplots and p-values to test the difference between two samples within a single chart.
  • Search box: A handy tool which allows you to quickly find methods and functions within the XLSTAT menu.
  • Multiple Correspondence Analysis: It is now possible to use the Burt table as input as well to link categories of a variable on a factorial map.

New in XLStat 2018.3 Build 50896 (May 7, 2018)

  • New statistical features and options:
  • Feature extraction: Transform a collection of text documents into feature vectors based on the Bag-Of-Words model. The output can be used to compare corpus in the Word Cloud tool or to run a classification. Available in XLSTAT-Marketing under the Text mining menu.
  • Probability Plot: This new graphical feature allows you to visually control if a sample comes from a population that follows a given distribution. Available in all XLSTAT solutions under the Visualizing data menu.
  • Correlation tests: It is now possible to compute the lower and upper bound confidence intervals for Pearson, Spearman and Kendall correlation coefficients. Available in all XLSTAT solutions under the Correlations/Associations tests menu.
  • ANOVA: Lower and upper confidence intervals can now be reported for multiple comparisons tests when running a one-way or multi-way ANOVA. Available in all XLSTAT solutions under the Modeling data menu.

New in XLStat 2018.2 Build 50198 (Mar 25, 2018)

  • New statistical features and options:
  • Word cloud: A powerful text visualization tool for quickly identifying important words in a document. Colors can be defined by the user for a more customized output. Available in all XLSTAT solutions under the Visualizing data menu.
  • ELECTRE I: Commonly used in decision making, this method aims to detect, compare and rank solutions to a multi-criteria problem. Available in all XLSTAT solutions under the Decision aid menu.
  • Conjoint design: It is now possible to take into account prohibited pairs when creating a full-profile or a choice-based conjoint design. Available in XLSTAT-Marketing & XLSTAT-Premium under the Conjoint analysis menu.
  • Random Forests: Several new options including Random Input method and Mean Decrease Accuracy for measuring variable importance. Available in all XLSTAT solutions under the Machine Learning menu.

New in XLStat 2015.5 (Oct 9, 2015)

  • This version brings several new features (MANOVA, marginal effects in Logistic regression, Confidence ellipses in Correspondence analysis, new results in Conjoint analysis), and a major breakthrough in statistical computing speed with the first function being GPU-enabled. Running Monte Carlo simulations for nonparametric tests can now be done on NVIDIA graphic cards, faster than ever. From now on, XLSTAT features will be gradually GPU-upgraded. XLSTAT version 2015.5 is compatible with the new Excel 2016 (Windows only, Mac to be announced later).