Fit Interpretable Models and Explain Blackbox Machine Learning (interpret)
Machine Learning package for training interpretable models and explaining blackbox systems. Historically, the most intelligible models were not very accurate, and the most accurate models were not intelligible. Microsoft Research has developed an algorithm called the Explainable Boosting Machine (EBM) which has both high accuracy and intelligibility. EBM uses modern machine learning techniques like bagging and boosting to breathe new life into traditional GAMs (Generalized Additive Models). This makes them as accurate as random forests and gradient boosted trees, and also enhances their intelligibility and editability. Details on the EBM algorithm can be found in the paper by Rich Caruana, Yin Lou, Johannes Gehrke, Paul Koch, Marc Sturm, and Noemie Elhadad (2015, <doi:10.1145/2783258.2788613>).

Bayesian Logistic Regression with Heavy-Tailed Priors (HTLR)
Efficient Bayesian multinomial logistic regression based on heavy-tailed (hyper-LASSO, non-convex) priors. The posterior of coefficients and hyper-parameters is sampled with restricted Gibbs sampling for leveraging the high-dimensionality and Hamiltonian Monte Carlo for handling the high-correlation among coefficients. A detailed description of the method: Li and Yao (2018), JSCS, 88:14, 2827-2851, <arXiv:1405.3319>.

Fitting RT-MPT Models (rtmpt)
Fit response-time extended multinomial processing tree (RT-MPT) models by Klauer and Kellen (2018) <doi:10.1016/j.jmp.2017.12.003>. The RT-MPT class not only incorporate frequencies like traditional multinomial processing tree (MPT) models, but also latencies. This enables it to estimate process completion times and encoding plus motor execution times next to the process probabilities of traditional MPTs. ‘rtmpt’ is a Bayesian framework and posterior samples are sampled using a Metropolis-Gibbs sampler like the one described in the Klauer and Kellen (2018), but with some modifications. Other than in the original C++ program we use the free and open source GNU Scientific Library (GSL). There is also the possibility to suppress single process completion times.

Modeling of Ordinal Random Variables via Softmax Regression (ohenery)
Supports the modeling of ordinal random variables, like the outcomes of races, via Softmax regression, under the Harville <doi:10.1080/01621459.1973.10482425> and Henery <doi:10.1111/j.2517-6161.1981.tb01153.x> models.

Color Manipulation Tools (prismatic)
Manipulate and visualize colors in a intuitive, low-dependency and functional way.

Advertisements