12 Types of Non-Linear Dimensionality Reduction (NLDR) Techniques in Machine Learning
A complete guide to manifold learning methods in unsupervised learning
Dimensionality reduction plays an important role in machine learning as many use cases exist.
Dimensionality reduction means that we reduce the number of features in the dataset to get so many advantages without losing much information.
Dimensionality reduction algorithms fall under the unsupervised learning category in which we train the algorithm with unlabelled data.
Even if there are so many types of dimensionality reduction methods, all of them fall under two major categories: Linear and Non-linear.
The linear methods linearly project data (hence the name, linear) from a high-dimensional space to a low-dimensional space. Examples include PCA and LDA.
The non-linear methods provide an approach to perform non-linear dimensionality reduction (NLDR). We often perform NLDR to discover the non-linear structure of the original data. NLDR is also useful when the original data is not linearly separable.
In some contexts, non-linear dimensionality reduction is also known as manifold learning.
