Basics of machine learning
OverviewQuestions:Objectives:
What is machine learning?
Why is it useful?
What are its different approaches?
Requirements:
Provide the basics of machine learning and its variants.
Learn how to do classification using the training and test data.
Learn how to use Galaxy’s machine learning tools.
Time estimation: 30 minutesSupporting Materials:Last modification: Oct 18, 2022
Introduction
Machine learning uses techniques from statistics, mathematics and computer science to make computer programs learn from data. It is one of the most popular fields of computer science and finds applications in multiple streams of data analysis such as classification, regression, clustering, dimensionality reduction, density estimation and many more. Some real-life applications are spam filtering, medical diagnosis, autonomous driving, recommendation systems, facial recognition, stock prices prediction and many more. The following image shows a basic flow of any machine learning task. Data is provided by a user to a machine learning algorithm for analysis.
There are multiple ways in which machine learning can be used to perform data analysis. They depend on the nature of data and the kind of data analysis. The following image shows the most popular ones. In supervised learning techniques, the categories of data records are known beforehand. But in unsupervised learning, the categories of data records are not known.
In general, machine learning can be used in multiple real-life tasks by applying its variants as depicted in the following image.
The following image shows how a classification task is performed. The complete data is divided into training and test sets. The training set is used by a classifier to learn features. It results in a trained model and its robustness (of learning) is evaluated using the test set (unseen by the classifier during the training).
This tutorial shows how to use a machine learning module implemented as a Galaxy tool. The data used in this tutorial is available at Zenodo.
AgendaPerforming a machine learning task (classification) using a tool involves the following steps:
Data upload
The datasets required for this tutorial contain 9 features of breast cells which include the thickness of clump, cell-size, cell-shape and so on (more information). In addition to these features, the training dataset contains one more column as target
. It has a binary value (0 or 1) for each row. 0
indicates no breast cancer and 1
indicates breast cancer. The test dataset does not contain the target
column.
Hands-on: Data upload
Create a new history for this tutorial
Click the new-history icon at the top of the history panel.
If the new-history is missing:
- Click on the galaxy-gear icon (History options) on the top of the history panel
- Select the option Create New from the menu
Import the following datasets and choose the type of data as
tabular
.https://zenodo.org/record/1401230/files/breast-w_train.tsv https://zenodo.org/record/1401230/files/breast-w_test.tsv
- Copy the link location
Open the Galaxy Upload Manager (galaxy-upload on the top-right of the tool panel)
- Select Paste/Fetch Data
Paste the link into the text field
Press Start
- Close the window
Rename datasets to
breast-w_train
andbreast-w_test
.
- Click on the galaxy-pencil pencil icon for the dataset to edit its attributes
- In the central panel, change the Name field
- Click the Save button
The datasets should look like these:
Train a classifier
In this step, we will use the SVM (support vector machine) classifier for training on the breast-w_train
dataset. The classifier learns a mapping between each row and its category. SVM is a memory efficient classifier which needs only those data points which lie on the decision boundaries among different classes to predict a class for a new sample. The rest of the data points can thrown away. We will use the LinearSVC
variant of SVM which is faster. Other variants SVC
and NuSVC
have high running time for large datasets. The last column of the training dataset contains a category/class for each row. The classifier learns a mapping between data row and its category which is called a trained model. The trained model is used to predict the categories of the unseen data.
Hands-on: Train a classifierSupport vector machines (SVMs) for classification tool with the following parameters to train:
- “Select a Classification Task”:
Train a model
- “Classifier type”:
Linear Support Vector Classification
- “Select input type”:
tabular data
- param-file “Training samples dataset”:
breast-w_train
tabular file- “Does the dataset contain header”:
Yes
- “Choose how to select data by column”:
All columns EXCLUDING some by column header name(s)
- “Type header name(s)”:
target
- param-file “Dataset containing class labels or target values”:
breast-w_train
tabular file- “Does the dataset contain header”:
Yes
- “Choose how to select data by column”:
Select columns by column header name(s)
- “Type header name(s):”:
target
Predict using a trained model
The previous step produced a trained model (zip
file) which we will use to predict classes for the test data (breast-w_test
).
Hands-on: Predict using a trained modelSupport vector machines (SVMs) for classification tool with the following parameters
- “Select a Classification Task”:
Load a model and predict
- param-file “Models”:
Zipped
file (output of Support vector machines (SVMs) for classification tool)- param-file “Data (tabular)”:
breast-w_test
file- “Does the dataset contain header”:
Yes
- “Select the type of prediction”:
Predict class labels
See predictions
The last column of the predicted dataset shows the category of each row. A row either gets 0
(no breast cancer) or 1
(breast cancer) as its predicted category.
Hands-on: See the predicted column
- Click on
view data
link of the dataset created after executing the previous step.- The last column of the
tabular
data shows the predicted category (target
) for each row.
Read more about machine learning using scikit-learn here.
Key points
Machine learning algorithms learn features from data.
It is used for multiple tasks such as classification, regression, clustering and so on.
Multiple learning tasks can be performed using Galaxy’s machine learning tools.
For the classification and regression tasks, data is divided into training and test sets.
Each sample/record in the training data has a category/class/label.
A machine learning algorithm learns features from the training data and do predictions on the test data.
Frequently Asked Questions
Have questions about this tutorial? Check out the tutorial FAQ page or the FAQ page for the Statistics and machine learning topic to see if your question is listed there. If not, please ask your question on the GTN Gitter Channel or the Galaxy Help ForumFeedback
Did you use this material as an instructor? Feel free to give us feedback on how it went.
Did you use this material as a learner or student? Click the form below to leave feedback.
Citing this Tutorial
- Anup Kumar, 2022 Basics of machine learning (Galaxy Training Materials). https://training.galaxyproject.org/training-material/topics/statistics/tutorials/machinelearning/tutorial.html Online; accessed TODAY
- Batut et al., 2018 Community-Driven Data Analysis Training for Biology Cell Systems 10.1016/j.cels.2018.05.012
Congratulations on successfully completing this tutorial!@misc{statistics-machinelearning, author = "Anup Kumar", title = "Basics of machine learning (Galaxy Training Materials)", year = "2022", month = "10", day = "18" url = "\url{https://training.galaxyproject.org/training-material/topics/statistics/tutorials/machinelearning/tutorial.html}", note = "[Online; accessed TODAY]" } @article{Batut_2018, doi = {10.1016/j.cels.2018.05.012}, url = {https://doi.org/10.1016%2Fj.cels.2018.05.012}, year = 2018, month = {jun}, publisher = {Elsevier {BV}}, volume = {6}, number = {6}, pages = {752--758.e1}, author = {B{\'{e}}r{\'{e}}nice Batut and Saskia Hiltemann and Andrea Bagnacani and Dannon Baker and Vivek Bhardwaj and Clemens Blank and Anthony Bretaudeau and Loraine Brillet-Gu{\'{e}}guen and Martin {\v{C}}ech and John Chilton and Dave Clements and Olivia Doppelt-Azeroual and Anika Erxleben and Mallory Ann Freeberg and Simon Gladman and Youri Hoogstrate and Hans-Rudolf Hotz and Torsten Houwaart and Pratik Jagtap and Delphine Larivi{\`{e}}re and Gildas Le Corguill{\'{e}} and Thomas Manke and Fabien Mareuil and Fidel Ram{\'{\i}}rez and Devon Ryan and Florian Christoph Sigloch and Nicola Soranzo and Joachim Wolff and Pavankumar Videm and Markus Wolfien and Aisanjiang Wubuli and Dilmurat Yusuf and James Taylor and Rolf Backofen and Anton Nekrutenko and Björn Grüning}, title = {Community-Driven Data Analysis Training for Biology}, journal = {Cell Systems} }