# Christoph Bergmeir

#### 11 packages on CRAN

The Stuttgart Neural Network Simulator (SNNS) is a library containing many standard implementations of neural networks. This package wraps the SNNS functionality to make it available from within R. Using the 'RSNNS' low-level interface, all of the algorithmic functionality and flexibility of SNNS can be accessed. Furthermore, the package contains a convenient high-level interface, so that the most common neural network topologies and learning algorithms integrate seamlessly into R.

An implementation of various learning algorithms based on fuzzy rule-based systems (FRBSs) for dealing with classification and regression tasks. Moreover, it allows to construct an FRBS model defined by human experts. FRBSs are based on the concept of fuzzy sets, proposed by Zadeh in 1965, which aims at representing the reasoning of human experts in a set of IF-THEN rules, to handle real-life problems in, e.g., control, prediction and inference, data mining, bioinformatics data processing, and robotics. FRBSs are also known as fuzzy inference systems and fuzzy models. During the modeling of an FRBS, there are two important steps that need to be conducted: structure identification and parameter estimation. Nowadays, there exists a wide variety of algorithms to generate fuzzy IF-THEN rules automatically from numerical data, covering both steps. Approaches that have been used in the past are, e.g., heuristic procedures, neuro-fuzzy techniques, clustering methods, genetic algorithms, squares methods, etc. Furthermore, in this version we provide a universal framework named 'frbsPMML', which is adopted from the Predictive Model Markup Language (PMML), for representing FRBS models. PMML is an XML-based language to provide a standard for describing models produced by data mining and machine learning algorithms. Therefore, we are allowed to export and import an FRBS model to/from 'frbsPMML'. Finally, this package aims to implement the most widely used standard procedures, thus offering a standard package for FRBS modeling to the R community.

An implementation of an algorithm family for continuous optimization called memetic algorithms with local search chains (MA-LS-Chains). Memetic algorithms are hybridizations of genetic algorithms with local search methods. They are especially suited for continuous optimization.

Implementations of algorithms for data analysis based on the rough set theory (RST) and the fuzzy rough set theory (FRST). We not only provide implementations for the basic concepts of RST and FRST but also popular algorithms that derive from those theories. The methods included in the package can be divided into several categories based on their functionality: discretization, feature selection, instance selection, rule induction and classification based on nearest neighbors. RST was introduced by ZdzisĹ‚aw Pawlak in 1982 as a sophisticated mathematical tool to model and process imprecise or incomplete information. By using the indiscernibility relation for objects/instances, RST does not require additional parameters to analyze the data. FRST is an extension of RST. The FRST combines concepts of vagueness and indiscernibility that are expressed with fuzzy sets (as proposed by Zadeh, in 1965) and RST.

Provides a simple R interface to the OPUS Miner algorithm (implemented in C++) for finding the top-k productive, non-redundant itemsets from transaction data. The OPUS Miner algorithm uses the OPUS search algorithm to efficiently discover the key associations in transaction data, in the form of self-sufficient itemsets, using either leverage or lift. See <http://i.giwebb.com/index.php/research/association-discovery/> for more information in relation to the OPUS Miner algorithm.

An implementation of a number of Global Trend models for time series forecasting that are Bayesian generalizations and extensions of some Exponential Smoothing models. The main differences/additions include 1) nonlinear global trend, 2) Student-t error distribution, and 3) a function for the error size, so heteroscedasticity. The methods are particularly useful for short time series. When tested on the well-known M3 dataset, they are able to outperform all classical time series algorithms. The models are fitted with MCMC using the 'rstan' package.

Provides a collection of self-labeled techniques for semi-supervised classification. In semi-supervised classification, both labeled and unlabeled data are used to train a classifier. This learning paradigm has obtained promising results, specifically in the presence of a reduced set of labeled examples. This package implements a collection of self-labeled techniques to construct a classification model. This family of techniques enlarges the original labeled set using the most confident predictions to classify unlabeled data. The techniques implemented can be applied to classification problems in several domains by the specification of a supervised base classifier. At low ratios of labeled data, it can be shown to perform better than classical supervised classifiers.

Methods and tools for displaying and analysing univariate time series forecasts including exponential smoothing via state space models and automatic ARIMA modelling.

Learning the structure of graphical models from datasets with thousands of variables. More information about the research papers detailing the theory behind Chordalysis is available at <http://www.francois-petitjean.com/Research> (KDD 2016, SDM 2015, ICDM 2014, ICDM 2013). The R package development site is <https://github.com/HerrmannM/Monash-ChoR>.

The 1001 time series from the M-competition (Makridakis et al. 1982) <DOI:10.1002/for.3980010202> and the 3003 time series from the IJF-M3 competition (Makridakis and Hibon, 2000) <DOI:10.1016/S0169-2070(00)00057-1>.

We provide an outlier robust alternative of the function ets() in the 'forecast' package of Hyndman and Khandakar (2008) <DOI:10.18637/jss.v027.i03>. For each method of a class of exponential smoothing variants we made a robust alternative. The class includes methods with a damped trend and/or seasonal components. The robust method is developed by robustifying every aspect of the original exponential smoothing variant. We provide robust forecasting equations, robust initial values, robust smoothing parameter estimation and a robust information criterion. The method is described in more detail in Crevits and Croux (2016) <DOI:10.13140/RG.2.2.11791.18080>.