Table 2.
Publically available software for machine learning applications [websites last viewed in October 2020].
| Application Name | Primary machine learning models | URL |
|---|---|---|
| Scikit-learna | Classification, regression, clustering | https://scikit-learn.org/stable/index.html |
| WEKA | Classification, regression, clustering | https://www.cs.waikato.ac.nz/ml/weka/ |
| KNIME | Classification, regression, clustering | https://www.knime.com/ |
| TensorFlowa | Neural networks | https://www.tensorflow.org/ |
| Kerasa | Neural networks | https://keras.io/ |
| Microsoft Cognitive Toolkita | Deep learningb | https://www.microsoft.com/en-us/cognitive-toolkit/ |
| PyTorcha | Deep learningb | https://pytorch.org/ |
| Theanoa | Deep learningb | http://www.deeplearning.net/software/theano/ |
| Caffe | Deep learningb | http://caffe.berkeleyvision.org/ |
Python frameworks (a collection of libraries, which are specific files containing pre-written code that can be imported into user's Python code).
Deep Learning is essentially working with large neural networks (‘deep’ typically refers to the number of layers).
WEKA—a collection of ML algorithms that can either be applied directly to a dataset or called from Java code written by a user.
KNIME—a GUI-based workflow platform that allows a user to drag-and-drop various pre-built machine learning modules without writing programming code. However, user code written in R and/or Python can be integrated in a KNIME analytical workflow.
Caffe—an open-source Deep Learning framework that supports interfaces such as C, C++, Python, MATLAB, and Command Line interfaces (CLI). However, familiarity with C++ is required.