1-machine_learning

Icon made by Freepik from www.flaticon.com

Machine Learning overview

Python Workshop - Leeds Institute for Data Analytics (LIDA)

The workshop has some additional materials:

1. Presentation
2. Linear Regression notebook
3. Classification notebook
4. Hidden Markov notebook

Necessary modules

• sklearn
• numpy
• pandas
• matplotlib.pyplot
• tensorflow
• tensorflow_probability

General definitions

1. Artificial Intelligence:
   • The effort to automate intellectual tasks normally performed by humans.
   • Very good AI is a very good set of rules!
2. Machine Learning
   • Rather than giving the program the rules, an algorithm finds (figures out) the rules for us.
   • Classical programming: Data+rules = answers
   • ML: Data + answers = rules.
3. Neural networks
   • A form of machine learning that uses layered representation of data
   • a lot of layers

Features

Information that is given to the model (input).

Labels

Is what we are trying to predicting (output)

TensorFlow

TensorFlow useful methods

• tf.Variable(itens, itens-type): return a tensor
• tf.rank(tensor): return the tensor's rank
• tf.shape(variable): return the tensor's shape
• tf.zeros(shape): returns a tensor with values 0
• tf.ones(shape): returns a tensor with values 1
• tf.reshape(new-shape): returns a tensor with a different shape

TensorFlow: core learning algorithms

• Linear regression: It analyses the relationship between independent and dependent variables to make predictions.
• Classification: It separates data points into separate categories.
• Clustering: A way of grouping the data points into different clusters, consisting of similar data points.
• Hidden Markov models: The HiddenMarkovModel distribution implements a (batch of) discrete hidden Markov models where the initial states, transition probabilities and observed states are all given by user-provided distributions.