Mahout common similarity measurement (note)

Mahout is a commonly used similarity measurement based on Recommendation Systems, classification, and clustering algorithms:

PearsoncorrelationsimilarityPearson distance

EuclideandistancesimilarityEuclidean distance

CosinemeasuresimilarityCosine distance (0.7ChangedUncenteredcosinesimilarity)

SpearmancorrelationsimilarityThe Pearson distance after sorting.

TanimotocoefficientsimilarityGrain correlation coefficient, based on Boolean preference

LoglikelihoodsimilarityThe maximum likelihood estimation, also known as the most approximate similarity estimation, is a statistical method used to obtain the relevant probability density function parameters of a sample set.Generally betterTanimotocoefficientsimilarity 

CityblocksimilarityBased on Manhattan distance

Reference: mahout recommendation algorithm Basics

This article from the "Wandering footsteps" blog, please be sure to keep this source http://now51jq.blog.51cto.com/3474143/1539515