In DataSystemsGroupUT/SmartML: Machine Learning Automation

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

SmartML:

Curently, SmartML is an R-Package representing a meta learning-based framework for automated selection and hyperparameter tuning for machine learning algorithms. Being meta-learning based, the framework is able to simulate the role of the machine learning expert. In particular, the framework is equipped with a continuously updated knowledge base that stores information about the meta-features of all processed datasets along with the associated performance of the different classifiers and their tuned parameters. Thus, for any new dataset, SmartML automatically extracts its meta features and searches its knowledge base for the best performing algorithm to start its optimization process. In addition, SmartML makes use of the new runs to continuously enrich its knowledge base to improve its performance and robustness for future runs.

SmartML Contribution Points and Goals:

The goal of SmartML is to automate the process of classifier algorithm selection, and hyper-parameter tuning in supervised machine learning using a modified version of SMAC bayesian optimization that prefers explitation more than exploration thanks to Meta-Learning. 1. SmartML is the first R package to deal with the sueprvised machine learning automation, and it is built over 16 different classifier algorithms from different R packages.
2. In addition, we offer different data preprocessing, and feature engineering algorithms that can be specified by user and applied on tabular datasets of either CSV or ARFF extensions easily. 3. SmartML has a collaborative knowledge base that grows by time as more users are using our tool. 4. Finally, SmartML has the ability to do some model interpretability plots for feature importance and interaction by help of iml package for ML model interpretability. 5. SmartML has a web service for the tool with a simple R Shiny interface that can be found HERE , and demonstration for how to use the web service can be found HERE.

Installation

You can install the released version of SmartML from Github with:

install_github("DataSystemsGroupUT/SmartML")

---

User Manual

Manual for the SmartML R package can be found HERE

---

Example

---

Contribution GuideLines to SmartML

To Contribute to SmartML, Please Follow these GuideLines

---

Publication

For More details, you can view our publication about SmartML. SmartML has been accepted as a DEMO paper at EDBT 19 in Lisbon Portugal [PDF]:

Mohamed Maher, Sherif Sakr.,SMARTML: A Meta Learning-Based Framework for Automated Selection and Hyperparameter Tuning for Machine Learning Algorithms (2019). Advances in Database Technology-EDBT 2019: 22nd International Conference on Extending Database Technology, Lisbon, Portugal, March 26-29.

---

Funding:

This work is funded by the European Regional Development Funds via the Mobilitas Plus programme (grant MOBTT75).

---

Licence:

© 2019, Data Systems Group at University of Tartu

This work is licensed under the terms of the GNU General Public License, version 3.0 (GPLv3)

DataSystemsGroupUT/SmartML documentation built on Nov. 24, 2020, 1:31 p.m.