Machine Learning - Unsupervised Learning - Dimensionality Reduction

dimensionality reductions is taking a dataset with a higher dimensional complexity and reducing the number of dimensions, if the data suggests doing so. For example, if the data set is 2 dimensional and the data points starts to show an arrangement of a one dimensional manifold (arrangement that appears that the data could be on a single dimension), then a function may be created that would rearrange the data to represent a single (or reduction) of dimensions. It is important that the data maintains its structure after the reduction has occurred.