Hi dear Avi hope you find this article helpful.

http://www.imm.dtu.dk/~perbb/MAS/ST116/module02/index.html

https://support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_cluster_sect021.htm

Can check on the rapidminer tool , I think it is better then R and suits for your work.

Best regards,

Mahi

http://finzi.psych.upenn.edu/library/bio3d/html/rmsd.html

https://www.youtube.com/watch?v=G8j8KAJtJlw

https://www.youtube.com/watch?v=-2Koi-caSZw

https://rstudio-pubs-static.s3.amazonaws.com/31867_8236987cf0a8444e962ccd2aec46d9c3.html

http://vancouverdata.blogspot.com/2010/11/text-analytics-with-rapidminer-part-4.html

https://rapidminer.com/

Our data set have 2447 object and 42 attribute of disease and meteorological.

In classification method, disease and meteorological attribute were converted to classification.  Then, have established the model by two-thirds of the package (training set) and the model’s accuracy is then estimated in the test set (one-thirds of the package).  

For cluster analysis, was used numerical tuple. Before cluster analysis method, Hopkins Statistic coefficient (H) is calculated, because H is 0.1, the tuple has statistically significant clusters.

By k-medoids method, because silhouette coefficient of created clusters are 0.4, 0.4, 0.4, clusters have statistically significant clusters. Then we are nominating k-medoids method.