Reference no: EM131447893

Load the auto-mpg sample dataset into the Orange application - ensure that the origin ?eld is set as a target attribute type, as it will be used as a class label. Also ensure that the mpg ?eld is set as a feature attribute type, as it will be used as a feature within the clustering. Impute any missing values via an Average Most Frequent method, and calculate Distances. Perform a Hierarchical Clustering using Linkage set to Average, with Pruning set to a Max Depth of 5. Also, set Selection to Top N with a value of 3. This will result in a shallow tree of depth 5, and a ?nal cut resulting in 3 clusters. Examine the resulting clusters (C1,C2,C3) via Distributions analysis - is there a clear relationship between the cluster assignment and class label (1,2,3)? What are the probabilities calculated for each value of origin for each cluster? Does changing the Max Depth a?ect the results in any way?