Using R in Robotics applications with ROS

For those of you who are interested in using R alongside Robotics applications and want to use ROS (Robot Operating System) together with R.
Enjoy the slides of our presentation on this topic during the last RBelgium meetup.

If you are interested in applying real-time analysis of your data streams or sensor data with R and ROS, we can help - get in touch at index.php/contact/get-in-touch

Host a CRAN mirror using Docker

CRAN mirrors are the backbone to everyday common R usage. They provide the R website and most of the R packages today. Currently there are about 104 official CRAN mirrors. Hosting a CRAN mirror is one step to help the R community and is explained here.

cran mirror docker

To ease that process, at BNOSAC, we have created a Docker image which sets up a CRAN mirror.
That Docker image can be found and is available for download at the following docker registry: https://registry.hub.docker.com/u/bnosac/cran-mirror

For people who don't know Docker, it is basically a tool which allows developers to containerise an application. In this case, the application is to run a CRAN mirror.
How does it work. 3 steps:

1. Install Docker on your computer or server as explained here, if you haven't done this already.
2. Pull the docker image: docker pull bnosac/cran-mirror
3. Run the CRAN mirror: docker run -p 22:22 -p 80:80 -v /home/bnosac/CRAN/:/var/www/html -d bnosac/cran-mirror

That's it! It started synching CRAN on /home/bnosac/CRAN and will synch every day at 02h30 UTC. You can now go to 0.0.0.0 in your browser or find the ip address where it is running and go to that address in your browser to see the R website (see https://registry.hub.docker.com/u/bnosac/cran-mirror for more info).

Now what can you do with it?

  • have a local CRAN mirror in your company
  • install.packages("data.table", repos = "mylocalmirror")
  • serve the community with another mirror if a closeby mirror is not readily available

Machine learning with R & Advanced R programming course

This year, BNOSAC offers 2 R courses in cooperation with the Leuven Statistics Research Center. The courses are part of the Leuven STATistics STATe of the Art Training Initiative and are given in Leuven (Belgium).

lstat courses

For R users and data scientists we offer a 2 short courses on R programming & statistical learning. Namely

* Advanced R programming topics (November 3-4, 2014)
* Statistical Machine Learning with R (November 27-28, 2014):

You can download the brochure with the courses here: http://lstat.kuleuven.be/training/HR_BRO_LSTAT_2014-2015.pdf

Interested? Registration can be done here.

RMOA package for running streaming classifcation & regression models now at CRAN

Last week, we released the RMOA package at CRAN (http://cran.r-project.org/web/packages/RMOA). It is an R package to allow building streaming classification and regression models on top of MOA.
MOA is the acronym of 'Massive Online Analysis' and it is the most popular open source framework for data stream mining which is being developed at the University of Waikato: http://moa.cms.waikato.ac.nz. Our RMOA package interfaces with MOA version 2014.04 and focusses on building, evaluating and scoring streaming classification & regression models on data streams.

rmoa
Classification & regression models which are possible through RMOA are:

- Classification trees:
  * AdaHoeffdingOptionTree
  * ASHoeffdingTree
  * DecisionStump
  * HoeffdingAdaptiveTree
  * HoeffdingOptionTree
  * HoeffdingTree
  * LimAttHoeffdingTree
  * RandomHoeffdingTree
- Bayesian classification:
  * NaiveBayes
  * NaiveBayesMultinomial
- Active learning classification:
  * ActiveClassifier
- Ensemble (meta) classifiers:
  * Bagging
      + LeveragingBag
      + OzaBag
      + OzaBagAdwin
      + OzaBagASHT
  * Boosting
      + OCBoost
      + OzaBoost
      + OzaBoostAdwin
  * Stacking
      + LimAttClassifier
  * Other
      + AccuracyUpdatedEnsemble
      + AccuracyWeightedEnsemble
      + ADACC
      + DACC
      + OnlineAccuracyUpdatedEnsemble
      + TemporallyAugmentedClassifier
      + WeightedMajorityAlgorithm

- Regression modelling:
  * AMRulesRegressor
  * FadingTargetMean
  * FIMTDD
  * ORTO
  * Perceptron
  * RandomRules
  * SGD (Stochastic Gradient Descent)
  * TargetMean

Interfaces are implemented to model data in standard files (csv, txt, delimited), ffdf data (from the ff package), data.frames and matrices.

Documentation of MOA directed towards RMOA users can be found at http://jwijffels.github.io/RMOA
Examples on the use of RMOA can be found in the documentation, on github at https://github.com/jwijffels/RMOA or e.g. by viewing the showcase at http://bnosac.be/index.php/blog/16-rmoa-massive-online-data-stream-classifications-with-r-a-moa

If you need support on building streaming models on top of your large dataset. Get into contact.