Experience studies are used by actuaries to explore historical experience across blocks of business and to inform assumption setting activities. This package provides functions for preparing data, creating studies, visualizing results, and beginning assumption development. Experience study methods, including exposure calculations, are described in: Atkinson & McGarry (2016) "Experience Study Calculations" <https://www.soa.org/49378a/globalassets/assets/files/research/experience-study-calculations.pdf>. The limited fluctuation credibility method used by the exp_stats() function is described in: Herzog (1999, ISBN:1-56698-374-6) "Introduction to Credibility Theory".

Count transformation models featuring parameters interpretable as discrete hazard ratios, odds ratios, reverse-time discrete hazard ratios, or transformed expectations. An appropriate data transformation for a count outcome and regression coefficients are simultaneously estimated by maximising the exact discrete log-likelihood using the computational framework provided in package mlt', technical details are given in Siegfried & Hothorn (2020) <DOI:10.1111/2041-210X.13383>. The package also contains an experimental implementation of multivariate count transformation models with an application to multi-species distribution models <DOI:10.48550/arXiv.2201.13095>.

Collection of ancillary functions and utilities for Partial Linear Single Index Models for Environmental mixture analyses, which currently provides functions for scalar outcomes. The outputs of these functions include the single index function, single index coefficients, partial linear coefficients, mixture overall effect, exposure main and interaction effects, and differences of quartile effects. In the future, we will add functions for binary, ordinal, Poisson, survival, and longitudinal outcomes, as well as models for time-dependent exposures. See Wang et al (2020) <doi:10.1186/s12940-020-00644-4> for an overview.

This package provides an implementation of two-dimensional functional principal component analysis (FPCA), Marginal FPCA, and Product FPCA for repeated functional data. Marginal and Product FPCA implementations are done for both dense and sparsely observed functional data. References: Chen, K., Delicado, P., & Müller, H. G. (2017) <doi:10.1111/rssb.12160>. Chen, K., & Müller, H. G. (2012) <doi:10.1080/01621459.2012.734196>. Hall, P., Müller, H.G. and Wang, J.L. (2006) <doi:10.1214/009053606000000272>. Yao, F., Müller, H. G., & Wang, J. L. (2005) <doi:10.1198/016214504000001745>.

Fast, model-agnostic implementation of different H-statistics introduced by Jerome H. Friedman and Bogdan E. Popescu (2008) <doi:10.1214/07-AOAS148>. These statistics quantify interaction strength per feature, feature pair, and feature triple. The package supports multi-output predictions and can account for case weights. In addition, several variants of the original statistics are provided. The shape of the interactions can be explored through partial dependence plots or individual conditional expectation plots. DALEX explainers, meta learners ('mlr3', tidymodels', caret') and most other models work out-of-the-box.

Rank-based tests for enrichment of KOG (euKaryotic Orthologous Groups) classes with up- or down-regulated genes based on a continuous measure. The meta-analysis is based on correlation of KOG delta-ranks across datasets (delta-rank is the difference between mean rank of genes belonging to a KOG class and mean rank of all other genes). With binary measure (1 or 0 to indicate significant and non-significant genes), one-tailed Fisher's exact test for over-representation of each KOG class among significant genes will be performed.

This package contains sixteen moisture sorption isotherm models, which evaluate the fitness of adsorption and desorption curves for further understanding of the relationship between moisture content and water activity. Fitness evaluation is conducted through parameter estimation and error analysis. Moreover, graphical representation, hysteresis area estimation, and isotherm classification through the equation of Blahovec & Yanniotis (2009) <doi:10.1016/j.jfoodeng.2008.08.007> which is based on the classification system introduced by Brunauer et. al. (1940) <doi:10.1021/ja01864a025> are also included for the visualization of models and hysteresis.

This package provides a mixture model for clustering individuals (or sampling groups) into stocks based on their genetic profile. Here, sampling groups are individuals that are sure to come from the same stock (e.g. breeding adults or larvae). The mixture (log-)likelihood is maximised using the EM-algorithm after finding good starting values via a K-means clustering of the genetic data. Details can be found in: Foster, S. D.; Feutry, P.; Grewe, P. M.; Berry, O.; Hui, F. K. C. & Davies (2020) <doi:10.1111/1755-0998.12920>.

This package provides a computational framework for analyzing mutations in immunoglobulin (Ig) sequences. Includes methods for Bayesian estimation of antigen-driven selection pressure, mutational load quantification, building of somatic hypermutation (SHM) models, and model-dependent distance calculations. Also includes empirically derived models of SHM for both mice and humans. Citations: Gupta and Vander Heiden, et al (2015) <doi:10.1093/bioinformatics/btv359>, Yaari, et al (2012) <doi:10.1093/nar/gks457>, Yaari, et al (2013) <doi:10.3389/fimmu.2013.00358>, Cui, et al (2016) <doi:10.4049/jimmunol.1502263>.

The LSTM (Long Short-Term Memory) model is a Recurrent Neural Network (RNN) based architecture that is widely used for time series forecasting. Min-Max transformation has been used for data preparation. Here, we have used one LSTM layer as a simple LSTM model and a Dense layer is used as the output layer. Then, compile the model using the loss function, optimizer and metrics. This package is based on Keras and TensorFlow modules and the algorithm of Paul and Garai (2021) <doi:10.1007/s00500-021-06087-4>.

This package provides a dedicated viral-explainer model tool designed to empower researchers in the field of HIV research, particularly in viral load and CD4 (Cluster of Differentiation 4) lymphocytes regression modeling. Drawing inspiration from the tidymodels framework for rigorous model building of Max Kuhn and Hadley Wickham (2020) <https://www.tidymodels.org>, and the DALEXtra tool for explainability by Przemyslaw Biecek (2020) <doi:10.48550/arXiv.2009.13248>. It aims to facilitate interpretable and reproducible research in biostatistics and computational biology for the benefit of understanding HIV dynamics.

New tools for the imputation of missing values in high-dimensional data are introduced using the non-parametric nearest neighbor methods. It includes weighted nearest neighbor imputation methods that use specific distances for selected variables. It includes an automatic procedure of cross validation and does not require prespecified values of the tuning parameters. It can be used to impute missing values in high-dimensional data when the sample size is smaller than the number of predictors. For more information see Faisal and Tutz (2017) <doi:10.1515/sagmb-2015-0098>.

Facilitates the use of data mining algorithms in classification and regression (including time series forecasting) tasks by presenting a short and coherent set of functions. Versions: 1.4.9 / 1.4.8 improved help, several warning and error code fixes (more stable version, all examples run correctly); 1.4.7 - improved Importance function and examples, minor error fixes; 1.4.6 / 1.4.5 / 1.4.4 new automated machine learning (AutoML) and ensembles, via improved fit(), mining() and mparheuristic() functions, and new categorical preprocessing, via improved delevels() function; 1.4.3 new metrics (e.g., macro precision, explained variance), new "lssvm" model and improved mparheuristic() function; 1.4.2 new "NMAE" metric, "xgboost" and "cv.glmnet" models (16 classification and 18 regression models); 1.4.1 new tutorial and more robust version; 1.4 - new classification and regression models, with a total of 14 classification and 15 regression methods, including: Decision Trees, Neural Networks, Support Vector Machines, Random Forests, Bagging and Boosting; 1.3 and 1.3.1 - new classification and regression metrics; 1.2 - new input importance methods via improved Importance() function; 1.0 - first version.

Alternating Manifold Proximal Gradient Method for Sparse PCA uses the Alternating Manifold Proximal Gradient (AManPG) method to find sparse principal components from a data or covariance matrix. Provides a novel algorithm for solving the sparse principal component analysis problem which provides advantages over existing methods in terms of efficiency and convergence guarantees. Chen, S., Ma, S., Xue, L., & Zou, H. (2020) <doi:10.1287/ijoo.2019.0032>. Zou, H., Hastie, T., & Tibshirani, R. (2006) <doi:10.1198/106186006X113430>. Zou, H., & Xue, L. (2018) <doi:10.1109/JPROC.2018.2846588>.

This package provides functions for testing if the covariance structure of 2-dimensional data (e.g. samples of surfaces X_i = X_i(s,t)) is separable, i.e. if covariance(X) = C_1 x C_2. A complete descriptions of the implemented tests can be found in the paper Aston, John A. D.; Pigoli, Davide; Tavakoli, Shahin. Tests for separability in nonparametric covariance operators of random surfaces. Ann. Statist. 45 (2017), no. 4, 1431--1461. <doi:10.1214/16-AOS1495> <https://projecteuclid.org/euclid.aos/1498636862> <arXiv:1505.02023>.

Flexible univariate count models based on renewal processes. The models may include covariates and can be specified with familiar formula syntax as in glm() and package flexsurv'. The methodology is described by Kharrat et all (2019) <doi:10.18637/jss.v090.i13> (included as vignette Countr_guide in the package). If the suggested package pscl is not available from CRAN, it can be installed with remotes::install_github("cran/pscl")'. It is no longer used by the functions in this package but is needed for some of the extended examples.

This project provides a group of new functions to calculate the outputs of the two main components of the Canadian Forest Fire Danger Rating System (CFFDRS) Van Wagner and Pickett (1985) <https://cfs.nrcan.gc.ca/publications?id=19973>) at various time scales: the Fire Weather Index (FWI) System Wan Wagner (1985) <https://cfs.nrcan.gc.ca/publications?id=19927> and the Fire Behaviour Prediction (FBP) System Forestry Canada Fire Danger Group (1992) <https://cfs.nrcan.gc.ca/pubwarehouse/pdfs/10068.pdf>. Some functions have two versions, table and raster based.

Tool collection for common and not so common data science use cases. This includes custom made algorithms for data management as well as value calculations that are hard to find elsewhere because of their specificity but would be a waste to get lost nonetheless. Currently available functionality: find sub-graphs in an edge list data.frame, find mode or modes in a vector of values, extract (a) specific regular expression group(s), generate ISO time stamps that play well with file names, or generate URL parameter lists by expanding value combinations.

Conducts sensitivity analyses for unmeasured confounding, selection bias, and measurement error (individually or in combination; VanderWeele & Ding (2017) <doi:10.7326/M16-2607>; Smith & VanderWeele (2019) <doi:10.1097/EDE.0000000000001032>; VanderWeele & Li (2019) <doi:10.1093/aje/kwz133>; Smith & VanderWeele (2021) <arXiv:2005.02908>). Also conducts sensitivity analyses for unmeasured confounding in meta-analyses (Mathur & VanderWeele (2020a) <doi:10.1080/01621459.2018.1529598>; Mathur & VanderWeele (2020b) <doi:10.1097/EDE.0000000000001180>) and for additive measures of effect modification (Mathur et al., under review).

Build a map of path-based geometry, this is a simple description of the number of parts in an object and their basic structure. Translation and restructuring operations for planar shapes and other hierarchical types require a data model with a record of the underlying relationships between elements. The gibble() function creates a geometry map, a simple record of the underlying structure in path-based hierarchical types. There are methods for the planar shape types in the sf and sp packages and for types in the trip and silicate packages.

Mapper-based survival analysis with transcriptomics data is designed to carry out. Mapper-based survival analysis is a modification of Progression Analysis of Disease (PAD) where survival data is taken into account in the filtering function. More details in: J. Fores-Martos, B. Suay-Garcia, R. Bosch-Romeu, M.C. Sanfeliu-Alonso, A. Falco, J. Climent, "Progression Analysis of Disease with Survival (PAD-S) by SurvMap identifies different prognostic subgroups of breast cancer in a large combined set of transcriptomics and methylation studies" <doi:10.1101/2022.09.08.507080>.

This package implements the GAMbag, GAMrsm and GAMens ensemble classifiers for binary classification (De Bock et al., 2010) <doi:10.1016/j.csda.2009.12.013>. The ensembles implement Bagging (Breiman, 1996) <doi:10.1023/A:1010933404324>, the Random Subspace Method (Ho, 1998) <doi:10.1109/34.709601> , or both, and use Hastie and Tibshirani's (1990, ISBN:978-0412343902) generalized additive models (GAMs) as base classifiers. Once an ensemble classifier has been trained, it can be used for predictions on new data. A function for cross validation is also included.

The Gene Ontology (GO) Consortium <https://geneontology.org/> organizes genes into hierarchical categories based on biological process (BP), molecular function (MF) and cellular component (CC, i.e., subcellular localization). Tools such as GoMiner (see Zeeberg, B.R., Feng, W., Wang, G. et al. (2003) <doi:10.1186/gb-2003-4-4-r28>) can leverage GO to perform ontological analysis of microarray and proteomics studies, typically generating a list of significant functional categories. To capture the benefit of all three ontologies, I developed HTGM3D', a three-dimensional version of GoMiner'.

An implementation of classifier chains (CC's) for multi-label prediction. Users can employ an external package (e.g. randomForest', C50'), or supply their own. The package can train a single set of CC's or train an ensemble of CC's -- in parallel if running in a multi-core environment. New observations are classified using a Gibbs sampler since each unobserved label is conditioned on the others. The package includes methods for evaluating the predictions for accuracy and aggregating across iterations and models to produce binary or probabilistic classifications.