Machine Learning Basics: Decision Tree Classifier

Machine Learning: demo a decision tree model to classify a dataset of Iris flowers. -Reviews an iris dataset of petal dimensions and classify the species of iris.

• Fundamental concepts in machine learning
• Supervised versus Unsupervised learning
• Machine learning frameworks
• Machine learning using Python and scikit-learn

Dependencies

• Python (>= 2.7 or >= 3.4),
• NumPy (>= 1.8.2),
• SciPy (>= 0.13.3)
• scikit-learn
• Iris Flower Dataset (included with scikit-learn)

Setup

• Clone repo
• Install scikit-learn and all dependencies: pip install -U scikit-learn[alldeps]
• For Anaconda Python users, load the "environmental.yml" file for dependencies.
• run, python iris-machine-learning

The model classifies the first data as label "1" (species Iris versicolor) and the second as label "0" (Iris setosa)

Notes

• A classifier looks at a piece of data and trys to classify it.
• DecisionTreeClassifier "is a class capable of performing multi-class classification on a dataset."
• The "Iris Flower Data Set" is commonly used to test machine learning
• The classifier often produces different results each time it is run.
  • This is because it uses probalistic behavior by randomly chooses a feature that might make the best comparison
• Load the dataset, then write the classifier.

from sklearn.datasets import load_iris      # Load iris flower dataset
iris = load_iris()  #assign dataset to a variable
print(list(iris.target_names))    #print labels (target names)

• Use the decision tree model to classify the dataset

classifier = tree.DecisionTreeClassifier()      # Build a classifier

• Supervised Learning: Machine intelligence uses models to predicat a category or quantity.
• Unsupervised Learning: A computer analyzes unlabeled data and attempts to recognize patterns.Unsupervised
• An example is a single element in a dataset.
• A feature is one characteristic of an example.

Review/Kata

• Write code to create a classifier.
• Write code to load the iris flower dataset.

Sources

• Machine Learning Basics
• scikit-learn.org
• Wikipedia: Machine Learning
• Docs: Classification
• Docs: sklearn.tree.DecisionTreeClassifier
• Introduction to Anaconda
• Wikipedia Iris Flower Data Set
• Docs: load_iris()
• Docs: fit()