Learn R Programming

rtemis Machine Learning and Visualization

A platform for advanced Machine Learning research and applications.
The goal of rtemis is to make data science efficient and accessible with no compromise on flexibility.

Online Documentation and vignettes

Installation

See here for more setup and installation instructions.

install.packages("remotes")
remotes::install_github("egenn/rtemis")

10-second intro to rtemis

Install dependencies if they are not already installed:

packages <- c("pbapply", "ranger")
.add <- !packages %in% installed.packages()
install.packages(packages[.add])

Load rtemis and get cross-validated random forest performance on the iris dataset:

library(rtemis)
mod <- elevate(iris)
mod$plot()

What's new

v. 0.78: First public release, April 2019

Features

Visualization
- Static: mplot3 family (base graphics)
- Dynamic: dplot3 family (plotly)
Unsupervised Learning
- Clustering: u.*
- Decomposition: d.*
Supervised Learning
- Classification, Regression, Survival Analysis: s.*
Cross-Decomposition
- Sparse Canonical Correlation / Sparse Decomposition: x.*
Meta-Models
[Have been temporarily removed for updating]
- Model Stacking: metaMod()
- Modality Stacking: metaFeat()
- Group-weighted Stacking: metaGroup()

Ongoing work

Novel algorithms developed in rtemis will generally be added to this public repository as soon as the corresponding papers or preprints are published.
R Documentation is ongoing and should be completed soon.
rtemis is under active development with many enhancements and extensions in the works

2019 Efstathios D. Gennatas

Copy Link

Version

Version

0.79

License

GPL (>=3)

Maintainer

Efstathios D. Gennatas

Last Published

May 21st, 2021

Functions in rtemis (0.79)

rtemis internal: Recursive function to build Additive Tree

as.boost Place model in boost structure

rtemis internal: Low-level Additive Tree procedure

as.addtboost Place model in addtboost structure

as.cartLiteBoostTV

Place model in cartLiteBoostTV structure

as.addboost Place model in addtboost structure

Check for constant columns

rtemis internal: Gradient Boosting of Additive Trees

as.cartLinBoostTV

Place model in cartLinBoostTV structure

as.data.tree.rpart

Convert rpart rules to data.tree object

as.glmLiteBoostTV

Place model in glmLiteBoostTV structure

Balanced Accuracy

Adjust HSV Color

Area under the Curve by pairwise concordance

Bare bones decision tree derived from rpart

Collapse data.frame to vector by getting column max

cartLiteBoostTV

Boost an rtemis learner for regression

Area under the ROC Curve

Concatenate Vectors

Bootstrap Resampling

Boost an rtemis learner for regression

Bag an rtemis learner for regression or classification [C, R]

Boost an rtemis learner for regression

Clustering with rtemis

Non-negative Matrix Factorization (NMF)

Class Imbalance

Color gradient for continuous variable

Principal Component Analysis

CUR Decomposition

rtemis internal: Dependencies check

Format Numbers for Printing

Pastelify a color (make a color more pastel)

Matrix Decomposition with rtemis

Set Dynamic Range

Sparse Principal Component Analysis

Singular Value Decomposition

Autoencoder using H2O

Extract coefficients from Additive Tree leaves

Create an alternating sequence of graded colors

Select rtemis Clusterer

Extract coefficients from Additive Tree leaves

Generalized Low-Rank Models (GLRM) on H2O

Independent Component Analysis

rtemis-internals: errorSummary

expand.glmLiteBoostTV

Expand boosting series

expand.cartLiteBoostTV

Expand boosting series

B-Spline matrix from dataset

expand.addtboost

Expand boosting series

distillTreeRules

Distill rules from trained RF and GBM learners

Create n-degree polynomial from data frame

Tune, Train, and Test an rtemis Learner by Nested Resampling

Error functions

Dynamic Plots (plotly)

rtemis internal: Grid check

gridSearchLearn

rtemis internal: Grid Search for Hyperparameter Tuning of rtemis Learners

Factor Analysis

Supervised Learning with rtemis

rtemis-internals: intro

factorHarmonize

Factor harmonize

Say No to NULL

Write Data to a Nifti File

Check if vector is constant

likelihoodMediboostChooseFeat

rtemis internal: likelihoodMediboostChooseFeat Selects a feature with maximum information gain and provides the decision values and column index for the chosen feature

Simple Color Operations

checkpoint.earlyStopping

Early stopping checkpoint

Leave-one-out Resampling

Connectivity Matrix to Edge List

Plot extended confusion matrix for binary classification

Mass-univariate Analysis

matchCasesByRules

Match Rules to Cases

mplot3.decision

mplot3: Decision boundaries

likelihoodMediboostSplitNode

rtemis internal: likelihoodMediboostSplitNode

Select rtemis Learner

Error Metrics for Supervised Learning

mplot3: Boxplot

mplot3: Barplot

Get the mode of a factor or integer

Locally Linear Embedding

Multidimensional Scaling

Dynamic Heatmap

rtemis-internals: dataPrepare

Plot variable importance using plotly

rtemis-internals: dataSummary

Logit transform

Check if variable is discrete (factor or integer)

K-fold Resampling

rtemis internal: Get Variable Name from Arguments

mplot3: Univariate plots: index, histogram, density, QQ-line

Log Loss for a binary classifier

Boost an rtemis learner for regression

Bayesian Gaussian Processes [R]

Convert text for label printing

mplot3: ADSR Plot

mplot3: Survival Plots

mplot3 ADDTREE trees

Classification Error

mplot3: Image (False color 2D)

mplot3: XY Scatter and line plots

Write Matrix to Multiple Nifti Files

mplot3 Scatter plot with marginal density and/or histogram

Message with provenance

Predict method for boost object

Plot CART Decision Tree

Convert data frame columns to list elements

Kernel Principal Component Analysis

predict.cartLinBoostTV

Predict method for cartLinBoostTV object

Print method for addTree object

mplot3: Variable Importance

t-distributed Stochastic Neighbor Embedding

Select rtemis Decomposer

Delay and Reverb Time Calculator

Uniform Manifold Approximation and Projection

expand.cartLinBoostTV

Expand boosting series

Get terms of a formula

Expand boosting series

rtemis internal: Ridge and Stump

Create a color gradient

rtemis internal: Ridge and Stump

Bare bones decision tree derived from rpart

rtemis-internals: gridSummary

mplot3 ROC curves

mplot3 Plot resample

print.addtboost

Print method for boost object

predict.addTreeRaw

Predict method for addTreeLite object

predict.addtboost

Predict method for addtboost object

plot method for resample object

Greater-than Table

print.cartLinBoostTV

Print method for boost object

Construct an n-length vector of letters

Print method for boost object

Random rtemis art

predict.rtTLS: predict method for rtTLS object

Write rtemis model to RDS file

R6 class for rtemis cross-decompositions

Plot HSV color range

Linear and Quadratic Discriminant Analysis [C]

predict.rtBSplines

Predict S3 method for rtBSplines

Mass-univariate CART prediction and variable importance

Linear Model Coefficients

Logistic function

Mass-univariate GAM Analysis

Print data frame

print.survError

Print survError object

Prune ADDTREE tree

preorderTree.addtree

rtemis-internal Traverse ADDTREE tree by preorder

Data preprocessing

One hot encoding

rtemis static plotting

mplot3: data.tree trees

Plot confusion matrix

mplot3.harmonograph

Plot a harmonograph

predict.nullmod

rtemis internal: predict for an object of class nullmod

rtemis-internals: outro

Predict method for nlareg object

Print regError object

print method for resample object

predict.ruleFeat

predict method for ruleFeat object

Artificial Neural Network [C, R]

ReLU - Rectified Linear Unit

mplot3: Guitar Fretboard

mplot3: True vs. Fitted plot

mplot3 Heatmap (image; modified heatmap)

prune.rpart experimental replacement

Variable Selection by Random Forest

Random Normal Matrix

rtemis Meta Model Class

rtMeta S3 methods

ExtraTrees [C, R]

Evolutionary Learning of Globally Optimal Trees [C, R]

rtemis internal: Resample Learn

Collapse data.frame to vector by getting row max

Resampling methods

Heatmap with plotly

rtemis internal: NonLinear Activation regression (NLAreg)

predict.glmLite

Predict method for glmLite object

predict.addTree

Predict method for addTree object

print.cartLiteBoostTV

Print method for boost object

predict.glmLiteBoostTV

Predict method for glmLiteBoostTV object

print.classError

Print classError

Logistic Regression

Local Polynomial Regression (LOESS) [R]

Gradient Boosting Machine on H2O [C, R]

Classify Images using pre-trained Inception network with mxnet [C]

Coefficient of Variation (Relative standard deviation)

rtModBag-methods

rtModBag S3 methods

rtemis Bagged Supervised Model Class

Population Standard Deviation

rtemis Cross-Validated Supervised Model Class

Reverse factor levels

rtModCV-methods

S3 methods for rtModCV class that differ from those of the rtMod superclass

rtemis Supervised Model Log Class

Reduce List to one value per element

Random Forest on H2O [C, R]

Nadaraya-Watson kernel regression [R]

Nonlinear Least Squares (NLS) [R]

rtemis model logger

Conditional Inference Trees [C, R, S]

GLM with Elastic Net Regularization [C, R, S]

Random Uniform Matrix

rtemis-internals: Convert rules from cutoffs to median/mode and range

Generalized Linear Model [C, R]

Classification and Regression Trees [C, R, S]

Multinomial Logistic Regression

Neural Network with mxnet [C, R]

rtemis Color Palettes

Build an ROC curve

Random Forest Classification and Regression [C, R]

Quantile Regression Neural Network [R]

rtModCVclass-class

rtemis Cross-Validated Classification Model Class

Layout for plotting on multiple panels

rtModLite-class

rtemis Lite Supervised Model Class

R6 class for rtemis clustering

rtemis Classification Model Class

rtClust-methods

rtClust S3 methods

rtModLite-methods

rtModLite S3 methods

Iterative Random Forest [C, R]

parameterSummary

rtemis-internals: parameterSummary

rtemis internal: lle::calc_k function adapted to work with pbapply

rtemis default-setting functions

print.glmLiteBoostTV

Print method for boost object

print.gridSearch

print method for gridSearch object

predict.cartLiteBoostTV

Predict method for cartLiteBoostTV object

predict.cartLite

Predict method for cartLite object

Pretty print list

Print 1 x N data frame

rtDecomLearn-class

R6 class for rtemis decomLearn

R6 Class for rtemis Decompositions

rtemis Supervised Model Class

rtemis: Machine Learning and Visualization

XGBoost with linear booster (wrapper for s.XGB)

Projection Pursuit Regression (BRUTO) [R]

Generalized Additive Model (GAM) C, R

Multivariate adaptive regression splines (MARS) [C, R]

Stochastic Gradient Descent (SGD) [C, R]

Random Forest Classification and Regression [C, R]

C5.0 Decision Trees and Rule-Based Models [C]

Spark MLlib Random Forest [C, R]

Random Forest for Classification, Regression, and Survival [C, R, S]

Regularized Generalized Additive Model (GAMSEL) [C, R]

Sparse Partial Least Squares Regression [C, R]

Write data to RDS file

Set type of columns

Adaboost Binary Classifier [C]

rtMod S3 methods

rtemis Color Palettes

Boosting of Additive Trees [R]

Additive Tree with Linear Nodes [R]

Additive Tree: Tree-Structured Boosting [C]

Gradient Boosting Machine [C, R, S]

Bayesian Additive Regression Trees [C, R]

Total Least Squares Regression [R]

Partitioning Around Medoids

Strict assignment by class or type

modified PMA::CCA.permute for parallel execution

rtemis-internals Project Variables to First Eigenvector

Partitioning Around Medoids with k Estimation

XGboost Classification and Regression [C, R]

Sparsify a vector

Fuzzy C-means Clustering

k-Nearest Neighbors Classification and Regression [C, R]

Gradient Boosting Machine [C, R, S]

x.CCA.permute.both

modified PMA:::CCA.permute.both for parallel execution

Hieararchical Ordered Partitioning and Collapsing Hybrid

Select rtemis cross-decomposer

Sparse CCA with Initialization By Resampling

Clustering by Hard Competitive Learning

Generalized Least Squares [R]

Polynomial Regression

Projection Pursuit Regression (PPR) [R]

Deep Learning on H2O [C, R]

Multivariate adaptive polynomial spline regression (POLYMARS) [C, R]

Projection Pursuit Tree Classification [C]

Parametric Survival Regression [S]

Size of matrix or vector

Quadratic Discriminant Analysis

Robust linear model

Naive Bayes Classifier [C]

Generalized Additive Model (GAM) C, R

Linear Discriminant Analysis

NonLinear Activation unit Regression (NLA) [R]

Generalized Additive Model (GAM) [C, R]

Extract standard error of fit from rtemis model

Softmax function

Standard Error of the Mean

Softplus function

ruleFeat [C, R]

Neural Network with tensorflow [C, R]

Sigmoid function

Sequence generation with automatic cycling

Support Vector Machines [C, R]

Resample using Stratified Subsamples

Spectral Clustering

Stratified Bootstrap Resampling

update.addtboost

Update boost object's fitted values

update.glmLiteBoostTV

rtemis internals: Update glmLiteBoostTV object's fitted values

update.rtMod.boost

rtemis internals: Update rtMod boost object's fitted values in-place

update.cartLinBoostTV

rtemis internals: Update cartLinBoostTV object's fitted values

Expectation Maximization Clustering

K-Means Clustering on H2O

update.cartLiteBoostTV

rtemis internals: Update cartLiteBoostTV object's fitted values

sparseVectorSummary

Sparseness and pairwise correlation of vectors

Synthesize Simple Regression Data

K-means Clustering

Neural Gas Clustering

Variable Selection by Variable Importace

Sparse Canonical Correlation Analysis (CCA)