In this demo, we show how to train a factor graph model using structured SVM. We will go through the basic knowledge of factor graph and structured output learning. Next, the corresponding APIs in Shogun will be covered. For testing the scalability, we show an experiment on a real OCR dataset for handwritten character recognition.
This notebook is about learning and using Gaussian Mixture Models (GMM) in Shogun. Below, we demonstrate how to use them for sampling, for density estimation via Expectation Maximisation (EM), and for clustering.
This notebook is intended to introduce the user to document classification and help him overcome the difficulties encountered in large collections. In order to achieve that we will explain the idea behind the hashing trick and we will show, through examples, how easy it is to use it with Shogun!
In this notebook we are going to see how Shogun Machine Learning Toolbox can be used for clustering with KMeans. In particular, we will be discussing the various options/choices provided to a user by the KMeans implementation in Shogun.
This notebook details the K-Nearest Neighbors (KNN) algorithm. It is a very simple but effective algorithm for solving multi-class classification problems.
In this notebook we are going to see how metric learning can be used for classification and feature selection using the Shogun Machine Learning Toolbox. In particular, will we be dealing with an algorithm for metric learning called Large Margin Nearest Neighbour, or just LMNN, for short.
This notebook illustrates how to train a binary support vector machine (SVM) classifier with shogun. A classifier attempts to distinguish objects of different type. In case of of binary classification there are just two types of objects that we want to distinguish.
In this notebook we will be doing some unsupervised learning using the suite of dimensionality reduction algorithms available in Shogun. Shogun provides access to all these algorithms using Tapkee, a C++ library especialized in dimensionality reduction.
In this notebook I am going to show you how we can do Blind Source Separation (BSS) using algorithms available in the Shogun Machine Learning Toolbox. What is Blind Source Separation? BSS is the separation of a set of source signals from a set of mixed signals.
In this notebook I want to demonstrate how we can apply the Independent Component Analysis (ICA) algorithms in Shogun to do Blind Source Separation (BSS) on images rather than typical 1D signals. This example program is going to be very similar to the bss_audio notebook only using images rather than audio files.
An interesting application of Independent Component Analysis (ICA) is extracting a baby's Electrocardiogram (ECG) by performing Blind Source Separation (BSS) on several time synchronised ECG's of the baby's mother.
This notebook is about Bayesian regression and classification models with Gaussian Process (GP) priors in Shogun. After providing a semi-formal introductions, we illustrate how to efficiently train them, use them for predictions, and automatically learn parameters .
This notebook illustrates how to estimate large-scale sparse Gaussian densities. It first introduces the reader into the mathematical background and then shows how one can do the estimation with shogun on a number of real-world data sets.
This notebook describes Shogun's framework for statistical hypothesis testing. We begin by giving a brief outline of the problem setting and then describe various implemented algorithms.
Since binary classification problems are one of the most thoroughly studied problems in machine learning, it is very appealing to consider reducing multiclass problems to binary ones. Then many advanced learning and optimization techniques as well as generalization bound analysis for binary classification can be utilized.
Naive Bayes is a simple and fast algorithm for multiclass learning. Formally, it predict the class by computing the posterior probability of each class k after observing the input x.
This notebook illustrates how to evaluate prediction algorithms in Shogus using cross-validation, and how to select their parameters using grid-search. Cross-validation estimates the expected value of a chosen loss function (for example testing error) via dividing data into disjoint partitions and then running the algorithms on a number of combinations for training and testing. Grid-search is a way to compare a number of registered parameters using cross-validation. We demonstrate this on a number of different algorithms within Shogun.