| Jason Brinkley |
Clustering algorithms have become very popular in recent years as a means to extract new information from data of all types. A clustering algorithm is data driven and partitions the observed data up into a non-overlapping set of clusters with clusters having similar values in the original data. Often cluster algorithms follow two paths: quantitative and qualitative.
Quantitative clustering algorithms often use continuous data and statistical measurements of 'distance' to group similar items together to form clusters. Qualitative clustering looks to take data with only a finite and fixed number of levels and group data with similar levels together. This hands on workshop will explain in detail both quantitative and qualitative clustering from both a theoretical and code driven perspective. Topics include Ward and centroid based clustering, hierarchical clustering, dendrograms and other multivariate visualizations that assist in developing or showcasing the results of a clustering algorithm. |