I have a big database of many items of retail company. If I would like to find the items which are similar to any particular item, can I use pearson correlation in Spark ML to do that? Is there any other better algorithm to do it? How do I make sure the machine also learns as it evolves?
Edit - I implemented Mapreduce program to find distance between various features. But how can I make it as Machine learning solution? Suppose if I let the program identify the correct neighbor, how can the program make use of this learning for next time?