Similarity and Dissimilarity Measures
In this post we introduce some similarity and dissimilarity measures. Most content come from the Data Mining from Vipin Kumar at University of Minnesota.
Distances
- Euclidean Distance
- Minkowski Distance
- Mahalanobis Distance
Common Properties of a Distance
Similarity Between Binary Vectors
SMC versus Jaccard
Jaccard: Compute the overlap proportion.
Others
Cosine Similarity: a measure of similarity between two non-zero vectors of an inner product space that measures the cosine of the angle between them, independent of magnitude.
Correlations (linear relation)
Information based metrics
Mutual Information
Maximal Information Coefficient