Distance Models

15 Novembro 2016, 13:00 • João Marques Silva

This week's lecture focused on distance models. The first part of the lecture covered Minkowski's distance, and its relation with other well-known definitions of distance. In addition, we formalized the concepts of exemplars and neighbors, and centroids and medoids. The second part of the lecture covered k-nearest neighbors classification. Finally, the third part of the lecture covered clustering, defined scatter, described the K-means algorithm and variants, and investigated hierarchical clustering and its representation with dendrograms. For the different parts of the lecture, different examples were analyzed.