• view each data point as an individual "cluster" with just that one point as a member

• calculate Euclidean distance between the centroids of all the clusters

• group the closest point pairs together

• repeat Step 2 and Step 3 until you reach a single cluster containing all the data in your set.

• plot a dendrogram

• decide on level

Compute distances

distances = dist(data, method = "euclidean")

applying Hierarchical clustering

clusterData = hclust(distances, method = "complete")

Plot the dendrogram

plot(clusterData)

Decision on Cluster Level

clusterGroups = cutree(clusterData, k = 10)

Agglomerative

• less complexity

• may fooled by local neighbors

• not see the larger implications of clusters

Divisive

• see the entire data distribution

• more accurate

• deeper complexity (can decrease the stability and increase the runtime)

k-mean

• simplicity

• instantiating random centroids and finding the closest points are time consuming

Hierarchical Cluster Analysis

• no need to pass in an explicit "k" number of clusters

• has more parameters to tweak

• clusters can be subjectively chosen through the evaluation of a dendrogram plot

Clustering Analysis for Recommender Systems:

• works at a group level

• generates less-personal recommendations

• often leads to worse accuracy than nearest neighbor algorithms

• works faster

• effective in shrinking the selection of relevant neighbors in a collaborative filtering algorithm

By Benjamin Johnston, Aaron Jones, Christopher Kruger May 2019, "Applied Unsupervised Learning with Python"

https://www.researchgate.net/publication/303870754_FHCC_A_Soft_Hierarchical_Clustering_Approach_for_Collaborative_Filtering_Recommendation

https://rpubs.com/kismetk/Netflix-recommendation

https://www.displayr.com/what-is-hierarchical-clustering/

https://towardsdatascience.com/understanding-the-concept-of-hierarchical-clustering-technique-c6e8243758ec

https://www.sciencedirect.com/topics/computer-science/hierarchical-cluster-analysis

https://www.bbvadata.com/recommender-systems-marketing-gets-personal/

https://subscription.packtpub.com/book/data/9781785884856/4/ch04lvl1sec25/clustering-techniques

https://towardsdatascience.com/unsupervised-learning-k-means-vs-hierarchical-clustering-5fe2da7c9554

https://subscription.packtpub.com/book/big_data_and_business_intelligence/9781789952292/2/ch02lvl1sec10/the-organization-of-hierarchy