What does the letter 'L' in L1-norm, L2-norm, Lp-norm metric mean? Is it an abbreviation of a word or just a convention?
Previous literature has simply written about L1 or L2 paradigms without explaining what L means. For example:
- Kwak N. Principal component analysis based on L1-norm maximization[J]. IEEE transactions on pattern analysis and machine intelligence, 2008, 30(9): 1672-1680.
- Jajuga K. L1-norm based fuzzy clustering[J]. Fuzzy Sets and Systems, 1991, 39(1): 43-50.
The $L$ stands for Lebesgue. This terminology originates in functional analysis where it is closely tied up with Lebesgue integration.