Uniform quantization

Uniform quantization involves partitioning the range of a continuous variable into equally spaced intervals, and then mapping each value to the interval that it overlaps. For quantizing vectors, this implies defining -dimensional regions.

Uniform quantization is extremely common for continuous variables. Along with quantile discretization, uniform quantization is often called “binning” or “bucketing” in data manipulation tools like Pandas. It is less common for vectors, as it results in regions, each of which is likely to be sparse. Clustering-based approaches are often used in these situations. (See vector quantization.)