Chapter9 ML

Unsupervised Learning

A type of machine learning that deals with data sets without labeled responses, aiming to find patterns and relationships in the data.

Clustering

The process of grouping similar instances together into clusters based on feature similarity.

Anomaly Detection

The task of identifying instances that deviate significantly from the normal pattern, often considered outliers.

Density Estimation

The estimation of the probability density function that represents the distribution of the dataset.

K-Means Algorithm

A popular unsupervised learning algorithm that partitions data into k clusters by assigning instances to the nearest cluster centroid.

DBSCAN

A density-based clustering algorithm that can identify clusters of arbitrary shapes by finding dense regions separated by low-density regions.

Gaussian Mixture Model (GMM)

A probabilistic model that assumes that all instances are generated from a mixture of several Gaussian distributions with unknown parameters.

Silhouette Score

A metric used to evaluate the quality of a clustering by measuring how similar an instance is to its own cluster compared to other clusters.

Mini-Batch K-Means

A variant of K-Means that processes small random batches of data rather than the entire dataset at once, improving speed and efficiency.

Label Propagation

A technique in semi-supervised learning where the labels from a small number of labeled instances are spread to the unlabeled instances within the same cluster.

Bayesian Gaussian Mixture Model

A variant of GMM that can automatically adjust the number of clusters based on the data, allowing for more flexible modeling of the data.

Expectations-Maximization Algorithm

An iterative method to find maximum likelihood estimates of parameters in statistical models, commonly used in GMM.

Inertia Metric

A measure of how well the clustering algorithm has performed, calculated as the average squared distance between instances and their closest centroid.

Hierarchical Clustering

A type of clustering that builds a hierarchy of clusters, typically represented as a dendrogram.

One-Class SVM

A variant of Support Vector Machines suited for novelty detection, where the model learns to recognize the normal instances and flags others as anomalies.

Active Learning

An iterative process where the model requests labels for instances from a human expert when it is uncertain about them, improving learning efficiency.

Dimensionality Reduction

The process of reducing the number of features in a dataset, which can improve computational efficiency and help in visualization.

Affinity Propagation

A clustering algorithm that uses message passing between data points to identify exemplars and clusters.

Mean-Shift Clustering

A clustering algorithm that iteratively shifts data points towards the mode (highest density point) of data points in their neighborhood.

Gaussian Radial Basis Function

A function used as a kernel in clustering algorithms, particularly measuring the distance between instances and centroids.

Elbow Method

A heuristic used to determine the optimal number of clusters in K-Means clustering by plotting the inertia against the number of clusters.

Cut-off Distance in Clustering

The threshold distance used to determine which instances in a cluster are labeled when performing semi-supervised learning.