Loading...

Supervised Learning: Datasets

Course video 9 of 35

This module delves into a wider variety of supervised learning methods for both classification and regression, learning about the connection between model complexity and generalization performance, the importance of proper feature scaling, and how to control model complexity by applying techniques like regularization to avoid overfitting. In addition to k-nearest neighbors, this week covers linear regression (least-squares, ridge, lasso, and polynomial regression), logistic regression, support vector machines, the use of cross-validation for model evaluation, and decision trees.

Applied Machine Learning in Python

Université du Michigan

Cours 3 de 5 dans Science des données appliquée avec Python Spécialisation

This course will introduce the learner to applied machine learning, focusing more on the techniques and methods than on the statistics behind these methods. The course will start with a discussion of how machine learning is different than descriptive statistics, and introduce the scikit learn toolkit through a tutorial. The issue of dimensionality of data will be discussed, and the task of clustering data, as well as evaluating those clusters, will be tackled. Supervised approaches for creating predictive models will be described, and learners will be able to apply the scikit learn predictive modelling methods while understanding process issues related to data generalizability (e.g. cross validation, overfitting). The course will end with a look at more advanced techniques, such as building ensembles, and practical limitations of predictive models. By the end of this course, students will be able to identify the difference between a supervised (classification) and unsupervised (clustering) technique, identify which technique they need to apply for a particular dataset and need, engineer features to meet that need, and write python code to carry out an analysis. This course should be taken after Introduction to Data Science in Python and Applied Plotting, Charting & Data Representation in Python and before Applied Text Mining in Python and Applied Social Analysis in Python.

Visualiser le programme de cours

Compétences que vous apprendrez

Python Programming, Machine Learning (ML) Algorithms, Machine Learning, Scikit-Learn

Enseigné par

Kevyn Collins-Thompson
Associate Professor

Transcription

To explore different supervised learning algorithms,

we're going to use a combination of small synthetic or

artificial datasets as examples, together with some larger real world datasets.

Psychit learn has a variety of methods in the SK learned

datasets library to create synthetic datasets.

The synthetic dataset will use, for

illustration purposes are typically low dimensional examples.

Because they only use a small number of features, typically one or two.

This makes them easy to explain and visualize.

Many real world datasets,

on the other hand have a higher dimensional feature space.

In other words they have dozens, hundreds, or even thousands, or

millions of features.

So some of the intuition we gain from looking at low dimensional examples

doesn't always translate to high dimensional datasets,

and we'll discuss that a bit more later.

For example, high dimensional data sets in some sense have most of their data in

corners with lots of empty space and that's kind of difficult to visualize.

We'll go through some examples later in the course.

But the low dimensional examples are still useful so

that we can understand things like how a model's complexity changes

with changes in some key parameters So for basic regression

we'll start with the simple problem that has one informative input variable.

One noisy linear output and 100 data set samples.

Here's a plot of a data set using scatter plot with each point

represented by one dot.

The x-axis shows the future value, and the y-axis shows the regression target.

To create this we use the make regression function in SK learned data sets.

Here is the code in the notebook.

To illustrate binary classification we will include a simple two class dataset

with two informative features.

Here's a scatterplot showing each data instance as a dot with the first feature

value corresponding to the x-axis.

And the second feature value corresponding to the y-axis.

The color of a point shows which class that data instance is labeled.

I'm calling this dataset simple because it has only two features,

both of which are informative.

In this case, these two classes are approximately linearly separable,

which means that a basic linear classifier placed between them does a pretty

good job of discriminating the points in the two classes.

So turning to the notebook, to create this data set we used to

make classification function in SK learn data sets.

Creating 100 points that roughly group the data samples into one cluster per class,

with here a 10% chance of randomly flipping the correct label

of any point just to make it a little more challenging the classifier.

We'll also look at a more complex binary classification problem

that uses two features.

But where the two classes are not really linearly separable,

instead forming into various clusters in different parts of the feature space.

This dataset was created in two steps.

First using the make_blobs function in SK learn datasets

to randomly generate 100 samples in 8 different clusters.

And then by changing the cluster label assigned by make_blobs, which is a number

from 1 to 8, to a binary number by converting it using a modulo 2 function.

Assigning the even index points to class 0 and odd index points to class 1.

To illustrate multi-class classification, we'll use our familiar fruits dataset,

which, as you may remember has four features and four possible target labels.

Here on the left, I'm showing the array of scatter plots that we saw in week one

that shows the relationship between all possible pairs of features and the class

labels, with the distribution of values for each feature along the diagonal.

To illustrate a real-world regression problem, we'll use a dataset

derived from the communities and crime dataset in the UCI repository.

Our dataset uses a subset of the original features and target values.

Which were originally created from combining several U.S.

government data sources, like the U.S. census.

Each data instance corresponds to a particular geographic area,

typically a town or a region of a city.

Our version of this dataset has 88 features

that encode various demographic and social economic properties of each location.

With 1994 location data instances.

The target value that we'll try to predict is the per capita violent crime rate.

To use this data set, we use the load_crime_dataset

function that's included with the share utilities module for this course

À propos de Coursera

Cours, Spécialisations et Diplômes en ligne enseignés par des enseignants du plus haut niveau provenant des meilleurs universités et établissements d'enseignement du monde.

Join a community of 40 million learners from around the world

Earn a skill-based course certificate to apply your knowledge

Gain confidence in your skills and further your career

Coursera propose un accès universel à la meilleure formation au monde, en partenariat avec des universités et des organisations du plus haut niveau, pour proposer des cours en ligne.

© 2019 Coursera Inc. Tous droits réservés.

Télécharger dans l'App Store

Disponible sur Google Play