This package provides a Bayesian approach to using predictive probability in an ANOVA construct with a continuous normal response, when threshold values must be obtained for the question of interest to be evaluated as successful (Sieck and Christensen (2021) <doi:10.1002/qre.2802>). The Bayesian Mission Mean (BMM) is used to evaluate a question of interest (that is, a mean that randomly selects combination of factor levels based on their probability of occurring instead of averaging over the factor levels, as in the grand mean). Under this construct, in contrast to a Gibbs sampler (or Metropolis-within-Gibbs sampler), a two-stage sampling method is required. The nested sampler determines the conditional posterior distribution of the model parameters, given Y, and the outside sampler determines the marginal posterior distribution of Y (also commonly called the predictive distribution for Y). This approach provides a sample from the joint posterior distribution of Y and the model parameters, while also accounting for the threshold value that must be obtained in order for the question of interest to be evaluated as successful.

Takes as input a stable oxygen isotope (d18O) profile measured in growth direction (D) through a shell + uncertainties in both variables (d18O_err & D_err). It then models the seasonality in the d18O record by fitting a combination of a growth and temperature sine wave to year-length chunks of the data (see Judd et al., (2018) <doi:10.1016/j.palaeo.2017.09.034>). This modeling is carried out along a sliding window through the data and yields estimates of the day of the year (Julian Day) and local growth rate for each data point. Uncertainties in both modeling routine and the data itself are propagated and pooled to obtain a confidence envelope around the age of each data point in the shell. The end result is a shell chronology consisting of estimated ages of shell formation relative to the annual cycle with their uncertainties. All formulae in the package serve this purpose, but the user can customize the model (e.g. number of days in a year and the mineralogy of the shell carbonate) through input parameters.

This package provides functions for basic hydraulic calculations related to water flow in circular pipes both flowing full (under pressure), and partially full (gravity flow), and trapezoidal open channels. For pressure flow this includes friction loss calculations by solving the Darcy-Weisbach equation for head loss, flow or diameter, plotting a Moody diagram, matching a pump characteristic curve to a system curve, and solving for flows in a pipe network using the Hardy-Cross method. The Darcy-Weisbach friction factor is calculated using the Colebrook (or Colebrook-White equation), the basis of the Moody diagram, the original citation being Colebrook (1939) <doi:10.1680/ijoti.1939.13150>. For gravity flow, the Manning equation is used, again solving for missing parameters. The derivation of and solutions using the Darcy-Weisbach equation and the Manning equation are outlined in many fluid mechanics texts such as Finnemore and Maurer (2024, ISBN:978-1-264-78729-6). Some gradually- and rapidly-varied flow functions are included. For the Manning equation solutions, this package uses modifications of original code from the iemisc package by Irucka Embry.

Self-Consistent Field(SCF) calculation method is one of the most important steps in the calculation methods of quantum chemistry. Ehrenreich, H., & Cohen, M. H. (1959). <doi:10.1103/PhysRev.115.786> However, the most prevailing software in this area, Gaussian''s SCF convergence process is hard to monitor, especially while the job is still running, causing researchers difficulty in knowing whether the oscillation has started or not, wasting time and energy on useless configurations or abandoning the jobs that can actually work. M.J. Frisch, G.W. Trucks, H.B. Schlegel et al. (2016). <https://gaussian.com> SCFMonitor enables Gaussian quantum chemistry calculation software users to easily read the Gaussian .log files and monitor the SCF convergence and geometry optimization process with little effort and clear, beautiful, and clean outputs. It can generate graphs using tidyverse to let users check SCF convergence and geometry optimization processes in real-time. The software supports processing .log files remotely using with rbase::url(). This software is a suitcase for saving time and energy for the researchers, supporting multiple versions of Gaussian'.

Calculation of distances, shortest paths and isochrones on weighted graphs using several variants of Dijkstra algorithm. Proposed algorithms are unidirectional Dijkstra (Dijkstra, E. W. (1959) <doi:10.1007/BF01386390>), bidirectional Dijkstra (Goldberg, Andrew & Fonseca F. Werneck, Renato (2005) <https://archive.siam.org/meetings/alenex05/papers/03agoldberg.pdf>), A* search (P. E. Hart, N. J. Nilsson et B. Raphael (1968) <doi:10.1109/TSSC.1968.300136>), new bidirectional A* (Pijls & Post (2009) <https://repub.eur.nl/pub/16100/ei2009-10.pdf>), Contraction hierarchies (R. Geisberger, P. Sanders, D. Schultes and D. Delling (2008) <doi:10.1007/978-3-540-68552-4_24>), PHAST (D. Delling, A.Goldberg, A. Nowatzyk, R. Werneck (2011) <doi:10.1016/j.jpdc.2012.02.007>). Algorithms for solving the traffic assignment problem are All-or-Nothing assignment, Method of Successive Averages, Frank-Wolfe algorithm (M. Fukushima (1984) <doi:10.1016/0191-2615(84)90029-8>), Conjugate and Bi-Conjugate Frank-Wolfe algorithms (M. Mitradjieva, P. O. Lindberg (2012) <doi:10.1287/trsc.1120.0409>), Algorithm-B (R. B. Dial (2006) <doi:10.1016/j.trb.2006.02.008>).

Fits ordinal regression models with elastic net penalty. Supported model families include cumulative probability, stopping ratio, continuation ratio, and adjacent category. These families are a subset of vector glm's which belong to a model class we call the elementwise link multinomial-ordinal (ELMO) class. Each family in this class links a vector of covariates to a vector of class probabilities. Each of these families has a parallel form, which is appropriate for ordinal response data, as well as a nonparallel form that is appropriate for an unordered categorical response, or as a more flexible model for ordinal data. The parallel model has a single set of coefficients, whereas the nonparallel model has a set of coefficients for each response category except the baseline category. It is also possible to fit a model with both parallel and nonparallel terms, which we call the semi-parallel model. The semi-parallel model has the flexibility of the nonparallel model, but the elastic net penalty shrinks it toward the parallel model. For details, refer to Wurm, Hanlon, and Rathouz (2021) <doi:10.18637/jss.v099.i06>.

This package provides functions for forward population genetic simulation in asexual populations, with special focus on cancer progression. Fitness can be an arbitrary function of genetic interactions between multiple genes or modules of genes, including epistasis, order restrictions in mutation accumulation, and order effects. Fitness (including just birth, just death, or both birth and death) can also be a function of the relative and absolute frequencies of other genotypes (i.e., frequency-dependent fitness). Mutation rates can differ between genes, and we can include mutator/antimutator genes (to model mutator phenotypes). Simulating multi-species scenarios and therapeutic interventions, including adaptive therapy, is also possible. Simulations use continuous-time models and can include driver and passenger genes and modules. Also included are functions for: simulating random DAGs of the type found in Oncogenetic Trees, Conjunctive Bayesian Networks, and other cancer progression models; plotting and sampling from single or multiple realizations of the simulations, including single-cell sampling; plotting the parent-child relationships of the clones; generating random fitness landscapes (Rough Mount Fuji, House of Cards, additive, NK, Ising, and Eggbox models) and plotting them.

Efficient sampling of truncated multivariate (scale) mixtures of normals under linear inequality constraints is nontrivial due to the analytically intractable normalizing constant. Meanwhile, traditional methods may subject to numerical issues, especially when the dimension is high and dependence is strong. Algorithms proposed by Li and Ghosh (2015) <doi: 10.1080/15598608.2014.996690> are adopted for overcoming difficulties in simulating truncated distributions. Efficient rejection sampling for simulating truncated univariate normal distribution is included in the package, which shows superiority in terms of acceptance rate and numerical stability compared to existing methods and R packages. An efficient function for sampling from truncated multivariate normal distribution subject to convex polytope restriction regions based on Gibbs sampler for conditional truncated univariate distribution is provided. By extending the sampling method, a function for sampling truncated multivariate Student's t distribution is also developed. Moreover, the proposed method and computation remain valid for high dimensional and strong dependence scenarios. Empirical results in Li and Ghosh (2015) <doi: 10.1080/15598608.2014.996690> illustrated the superior performance in terms of various criteria (e.g. mixing and integrated auto-correlation time).

Thematic quality indices are provided to facilitate the evaluation and quality control of geospatial data products (e.g. thematic maps, remote sensing classifications, etc.). The indices offered are based on the so-called confusion matrix. This matrix is constructed by comparing the assigned classes or attributes of a set of pairs of positions or objects in the product and the ground truth. In this package it is considered that the classes of the ground truth correspond to the columns and that the classes of the product to be valued correspond to the rows. The package offers two object classes with their methods: ConfMatrix (Confusion matrix) and QCCS (Quality Control Columns Set). The ConfMatrix class of objects offers more than 20 methods based on the confusion matrix. The QCCS class of objects offers a different perspective in which the ground truth is considered to allow the values of the column marginals to be fixed, see Ariza López et al. (2019) <doi:10.3390/app9204240> and Canran Liu et al. (2007) <doi:10.1016/j.rse.2006.10.010> for more details. The package was created with R6'.

Since their introduction by Bose and Nair (1939) <https://www.jstor.org/stable/40383923>, partially balanced incomplete block (PBIB) designs remain an important class of incomplete block designs. The concept of association scheme was used by Bose and Shimamoto (1952) <doi:10.1080/01621459.1952.10501161> for the classification of these designs. The constraint of resources always motivates the experimenter to advance towards PBIB designs, more specifically to higher associate class PBIB designs from balanced incomplete block designs. It is interesting to note that many times higher associate PBIB designs perform better than their counterpart lower associate PBIB designs for the same set of parameters v, b, r, k and lambda_i (i=1,2...m). This package contains functions named GETD() for generating m-associate (m>=2) class PBIB designs along with parameters (v, b, r, k and lambda_i, i = 1, 2,â ¦,m) based on Generalized Triangular (GT) Association Scheme. It also calculates the Information matrix, Average variance factor and canonical efficiency factor of the generated design. These designs, besides having good efficiency, require smaller number of replications and smallest possible concurrence of treatment pairs.

Different inference procedures are proposed in the literature to correct for selection bias that might be introduced with non-random selection mechanisms. A class of methods to correct for selection bias is to apply a statistical model to predict the units not in the sample (super-population modeling). Other studies use calibration or Statistical Matching (statistically match nonprobability and probability samples). To date, the more relevant methods are weighting by Propensity Score Adjustment (PSA). The Propensity Score Adjustment method was originally developed to construct weights by estimating response probabilities and using them in Horvitzâ Thompson type estimators. This method is usually used by combining a non-probability sample with a reference sample to construct propensity models for the non-probability sample. Calibration can be used in a posterior way to adding information of auxiliary variables. Propensity scores in PSA are usually estimated using logistic regression models. Machine learning classification algorithms can be used as alternatives for logistic regression as a technique to estimate propensities. The package NonProbEst implements some of these methods and thus provides a wide options to work with data coming from a non-probabilistic sample.

This package provides a collection of standard factor retention methods in Exploratory Factor Analysis (EFA), making it easier to determine the number of factors. Traditional methods such as the scree plot by Cattell (1966) <doi:10.1207/s15327906mbr0102_10>, Kaiser-Guttman Criterion (KGC) by Guttman (1954) <doi:10.1007/BF02289162> and Kaiser (1960) <doi:10.1177/001316446002000116>, and flexible Parallel Analysis (PA) by Horn (1965) <doi:10.1007/BF02289447> based on eigenvalues form PCA or EFA are readily available. This package also implements several newer methods, such as the Empirical Kaiser Criterion (EKC) by Braeken and van Assen (2017) <doi:10.1037/met0000074>, Comparison Data (CD) by Ruscio and Roche (2012) <doi:10.1037/a0025697>, and Hull method by Lorenzo-Seva et al. (2011) <doi:10.1080/00273171.2011.564527>, as well as some AI-based methods like Comparison Data Forest (CDF) by Goretzko and Ruscio (2024) <doi:10.3758/s13428-023-02122-4> and Factor Forest (FF) by Goretzko and Buhner (2020) <doi:10.1037/met0000262>. Additionally, it includes a deep neural network (DNN) trained on large-scale datasets that can efficiently and reliably determine the number of factors.

Publication bias, the fact that studies identified for inclusion in a meta analysis do not represent all studies on the topic of interest, is commonly recognized as a threat to the validity of the results of a meta analysis. One way to explicitly model publication bias is via selection models or weighted probability distributions. In this package we provide implementations of several parametric and nonparametric weight functions. The novelty in Rufibach (2011) is the proposal of a non-increasing variant of the nonparametric weight function of Dear & Begg (1992). The new approach potentially offers more insight in the selection process than other methods, but is more flexible than parametric approaches. To maximize the log-likelihood function proposed by Dear & Begg (1992) under a monotonicity constraint we use a differential evolution algorithm proposed by Ardia et al (2010a, b) and implemented in Mullen et al (2009). In addition, we offer a method to compute a confidence interval for the overall effect size theta, adjusted for selection bias as well as a function that computes the simulation-based p-value to assess the null hypothesis of no selection as described in Rufibach (2011, Section 6).

Incorporates functions for image preprocessing, filtering and image recognition. The package takes advantage of RcppArmadillo to speed up computationally intensive functions. The histogram of oriented gradients descriptor is a modification of the findHOGFeatures function of the SimpleCV computer vision platform, the average_hash(), dhash() and phash() functions are based on the ImageHash python library. The Gabor Feature Extraction functions are based on Matlab code of the paper, "CloudID: Trustworthy cloud-based and cross-enterprise biometric identification" by M. Haghighat, S. Zonouz, M. Abdel-Mottaleb, Expert Systems with Applications, vol. 42, no. 21, pp. 7905-7916, 2015, <doi:10.1016/j.eswa.2015.06.025>. The SLIC and SLICO superpixel algorithms were explained in detail in (i) "SLIC Superpixels Compared to State-of-the-art Superpixel Methods", Radhakrishna Achanta, Appu Shaji, Kevin Smith, Aurelien Lucchi, Pascal Fua, and Sabine Suesstrunk, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 34, num. 11, p. 2274-2282, May 2012, <doi:10.1109/TPAMI.2012.120> and (ii) "SLIC Superpixels", Radhakrishna Achanta, Appu Shaji, Kevin Smith, Aurelien Lucchi, Pascal Fua, and Sabine Suesstrunk, EPFL Technical Report no. 149300, June 2010.

This package provides a retake on the Genie algorithm (Gagolewski, 2021 <DOI:10.1016/j.softx.2021.100722>) - a robust hierarchical clustering method (Gagolewski, Bartoszuk, Cena, 2016 <DOI:10.1016/j.ins.2016.05.003>). Now faster and more memory efficient; determining the whole hierarchy for datasets of 10M points in low dimensional Euclidean spaces or 100K points in high-dimensional ones takes only 1-2 minutes. Allows clustering with respect to mutual reachability distances so that it can act as a noise point detector or a robustified version of HDBSCAN* (that is able to detect a predefined number of clusters and hence it does not dependent on the somewhat fragile eps parameter). The package also features an implementation of inequality indices (the Gini, Bonferroni index), external cluster validity measures (e.g., the normalised clustering accuracy and partition similarity scores such as the adjusted Rand, Fowlkes-Mallows, adjusted mutual information, and the pair sets index), and internal cluster validity indices (e.g., the Calinski-Harabasz, Davies-Bouldin, Ball-Hall, Silhouette, and generalised Dunn indices). See also the Python version of genieclust available on PyPI', which supports sparse data, more metrics, and even larger datasets.

Neighbour-balanced designs ensure that no treatment is disadvantaged unfairly by its surroundings. The treatment allocation in these designs is such that every treatment appears equally often as a neighbour with every other treatment. Neighbour Balanced Designs are employed when there is a possibility of neighbour effects from treatments used in adjacent experimental units. In the literature, a vast number of such designs have been developed. This package generates some efficient neighbour balanced block designs which are balanced and partially variance balanced for estimating the contrast pertaining to direct and neighbour effects, as well as provides a function for analysing the data obtained from such trials (Azais, J.M., Bailey, R.A. and Monod, H. (1993). "A catalogue of efficient neighbour designs with border plots". Biometrics, 49, 1252-1261 ; Tomar, J. S., Jaggi, Seema and Varghese, Cini (2005)<DOI: 10.1080/0266476042000305177>. "On totally balanced block designs for competition effects"). This package contains functions named nbbd1(),nbbd2(),nbbd3(),pnbbd1() and pnbbd2() which generates neighbour balanced block designs within a specified range of number of treatment (v). It contains another function named anlys()for performing the analysis of data generated from such trials.

The permubiome R package was created to perform a permutation-based non-parametric analysis on microbiome data for biomarker discovery aims. This test executes thousands of comparisons in a pairwise manner, after a random shuffling of data into the different groups of study with a prior selection of the microbiome features with the largest variation among groups. Previous to the permutation test itself, data can be normalized according to different methods proposed to handle microbiome data ('proportions or Anders'). The median-based differences between groups resulting from the multiple simulations are fitted to a normal distribution with the aim to calculate their significance. A multiple testing correction based on Benjamini-Hochberg method (fdr) is finally applied to extract the differentially presented features between groups of your dataset. LATEST UPDATES: v1.1 and olders incorporates function to parse COLUMN format; v1.2 and olders incorporates -optimize- function to maximize evaluation of features with largest inter-class variation; v1.3 and olders includes the -size.effect- function to perform estimation statistics using the bootstrap-coupled approach implemented in the dabestr (>=0.3.0) R package. Current v1.3.2 fixed bug with "Class" recognition and updated dabestr functions.

This package provides an all-in-one solution for automatic classification of sound events using convolutional neural networks (CNN). The main purpose is to provide a sound classification workflow, from annotating sound events in recordings to training and automating model usage in real-life situations. Using the package requires a pre-compiled collection of recordings with sound events of interest and it can be employed for: 1) Annotation: create a database of annotated recordings, 2) Training: prepare train data from annotated recordings and fit CNN models, 3) Classification: automate the use of the fitted model for classifying new recordings. By using automatic feature selection and a user-friendly GUI for managing data and training/deploying models, this package is intended to be used by a broad audience as it does not require specific expertise in statistics, programming or sound analysis. Please refer to the vignette for further information. Gibb, R., et al. (2019) <doi:10.1111/2041-210X.13101> Mac Aodha, O., et al. (2018) <doi:10.1371/journal.pcbi.1005995> Stowell, D., et al. (2019) <doi:10.1111/2041-210X.13103> LeCun, Y., et al. (2012) <doi:10.1007/978-3-642-35289-8_3>.

Systematic conservation prioritization using mixed integer linear programming (MILP). It provides a flexible interface for building and solving conservation planning problems. Once built, conservation planning problems can be solved using a variety of commercial and open-source exact algorithm solvers. By using exact algorithm solvers, solutions can be generated that are guaranteed to be optimal (or within a pre-specified optimality gap). Furthermore, conservation problems can be constructed to optimize the spatial allocation of different management actions or zones, meaning that conservation practitioners can identify solutions that benefit multiple stakeholders. To solve large-scale or complex conservation planning problems, users should install the Gurobi optimization software (available from <https://www.gurobi.com/>) and the gurobi R package (see Gurobi Installation Guide vignette for details). Users can also install the IBM CPLEX software (<https://www.ibm.com/products/ilog-cplex-optimization-studio/cplex-optimizer>) and the cplexAPI R package (available at <https://github.com/cran/cplexAPI>). Additionally, the rcbc R package (available at <https://github.com/dirkschumacher/rcbc>) can be used to generate solutions using the CBC optimization software (<https://github.com/coin-or/Cbc>). For further details, see Hanson et al. (2024) <doi:10.1111/cobi.14376>.

Non-Domestic VAERS vaccine data for 01/01/2016 - 06/14/2016. If you want to explore the full VAERS data for 1990 - Present (data, symptoms, and vaccines), then check out the vaersND package from the URL below. The URL and BugReports below correspond to the vaersND package, of which vaersNDvax is a small subset (2016 only). vaersND is not hosted on CRAN due to the large size of the data set. To install the Suggested vaers and vaersND packages, use the following R code: devtools::install_git("https://gitlab.com/iembry/vaers.git", build_vignettes = TRUE) and devtools::install_git("https://gitlab.com/iembry/vaersND.git", build_vignettes = TRUE)'. "VAERS is a national vaccine safety surveillance program co-sponsored by the US Centers for Disease Control and Prevention (CDC) and the US Food and Drug Administration (FDA). VAERS is a post-marketing safety surveillance program, collecting information about adverse events (possible side effects) that occur after the administration of vaccines licensed for use in the United States." For more information about the data, visit <https://vaers.hhs.gov/index>. For information about vaccination/immunization hazards, visit <http://www.questionuniverse.com/rethink.html/#vaccine>.

pathwayPCA is an integrative analysis tool that implements the principal component analysis (PCA) based pathway analysis approaches described in Chen et al. (2008), Chen et al. (2010), and Chen (2011). pathwayPCA allows users to: (1) Test pathway association with binary, continuous, or survival phenotypes. (2) Extract relevant genes in the pathways using the SuperPCA and AES-PCA approaches. (3) Compute principal components (PCs) based on the selected genes. These estimated latent variables represent pathway activities for individual subjects, which can then be used to perform integrative pathway analysis, such as multi-omics analysis. (4) Extract relevant genes that drive pathway significance as well as data corresponding to these relevant genes for additional in-depth analysis. (5) Perform analyses with enhanced computational efficiency with parallel computing and enhanced data safety with S4-class data objects. (6) Analyze studies with complex experimental designs, with multiple covariates, and with interaction effects, e.g., testing whether pathway association with clinical phenotype is different between male and female subjects. Citations: Chen et al. (2008) <https://doi.org/10.1093/bioinformatics/btn458>; Chen et al. (2010) <https://doi.org/10.1002/gepi.20532>; and Chen (2011) <https://doi.org/10.2202/1544-6115.1697>.

JASPAR (https://jaspar.elixir.no/) is a widely-used open-access database presenting manually curated high-quality and non-redundant DNA-binding profiles for transcription factors (TFs) across taxa. In this 10th release and 20th-anniversary update, the CORE collection has expanded with 329 new profiles. We updated three existing profiles and provided orthogonal support for 72 profiles from the previous release UNVALIDATED collection. Altogether, the JASPAR 2024 update provides a 20 percent increase in CORE profiles from the previous release. A trimming algorithm enhanced profiles by removing low information content flanking base pairs, which were likely uninformative (within the capacity of the PFM models) for TFBS predictions and modelling TF-DNA interactions. This release includes enhanced metadata, featuring a refined classification for plant TFs structural DNA-binding domains. The new JASPAR collections prompt updates to the genomic tracks of predicted TF-binding sites in 8 organisms, with human and mouse tracks available as native tracks in the UCSC Genome browser. All data are available through the JASPAR web interface and programmatically through its API and the updated Bioconductor and pyJASPAR packages. Finally, a new TFBS extraction tool enables users to retrieve predicted JASPAR TFBSs intersecting their genomic regions of interest.

This package provides a system designed for detecting concept drift in streaming datasets. It offers a comprehensive suite of statistical methods to detect concept drift, including methods for monitoring changes in data distributions over time. The package supports several tests, such as Drift Detection Method (DDM), Early Drift Detection Method (EDDM), Hoeffding Drift Detection Methods (HDDM_A, HDDM_W), Kolmogorov-Smirnov test-based Windowing (KSWIN) and Page Hinkley (PH) tests. The methods implemented in this package are based on established research and have been demonstrated to be effective in real-time data analysis. For more details on the methods, please check to the following sources. KobyliÅ ska et al. (2023) <doi:10.48550/arXiv.2308.11446>, S. Kullback & R.A. Leibler (1951) <doi:10.1214/aoms/1177729694>, Gama et al. (2004) <doi:10.1007/978-3-540-28645-5_29>, Baena-Garcia et al. (2006) <https://www.researchgate.net/publication/245999704_Early_Drift_Detection_Method>, Frà as-Blanco et al. (2014) <https://ieeexplore.ieee.org/document/6871418>, Raab et al. (2020) <doi:10.1016/j.neucom.2019.11.111>, Page (1954) <doi:10.1093/biomet/41.1-2.100>, Montiel et al. (2018) <https://jmlr.org/papers/volume19/18-251/18-251.pdf>.

Distance measures (GDM1, GDM2, Sokal-Michener, Bray-Curtis, for symbolic interval-valued data), cluster quality indices (Calinski-Harabasz, Baker-Hubert, Hubert-Levine, Silhouette, Krzanowski-Lai, Hartigan, Gap, Davies-Bouldin), data normalization formulas (metric data, interval-valued symbolic data), data generation (typical and non-typical data), HINoV method, replication analysis, linear ordering methods, spectral clustering, agreement indices between two partitions, plot functions (for categorical and symbolic interval-valued data). (MILLIGAN, G.W., COOPER, M.C. (1985) <doi:10.1007/BF02294245>, HUBERT, L., ARABIE, P. (1985) <doi:10.1007%2FBF01908075>, RAND, W.M. (1971) <doi:10.1080/01621459.1971.10482356>, JAJUGA, K., WALESIAK, M. (2000) <doi:10.1007/978-3-642-57280-7_11>, MILLIGAN, G.W., COOPER, M.C. (1988) <doi:10.1007/BF01897163>, JAJUGA, K., WALESIAK, M., BAK, A. (2003) <doi:10.1007/978-3-642-55721-7_12>, DAVIES, D.L., BOULDIN, D.W. (1979) <doi:10.1109/TPAMI.1979.4766909>, CALINSKI, T., HARABASZ, J. (1974) <doi:10.1080/03610927408827101>, HUBERT, L. (1974) <doi:10.1080/01621459.1974.10480191>, TIBSHIRANI, R., WALTHER, G., HASTIE, T. (2001) <doi:10.1111/1467-9868.00293>, BRECKENRIDGE, J.N. (2000) <doi:10.1207/S15327906MBR3502_5>, WALESIAK, M., DUDEK, A. (2008) <doi:10.1007/978-3-540-78246-9_11>).