12-12-2017 02:13 AM
I am new to Spark, and I am trying to find the best and simplest way for providing some simple forecast for transactions.
Imagine I have a dataset (CSV) with fields as in the picture below, but with hundreds of products and dozens of clients.
How can I forecast what product (and ideally amount) will be ordered next week.
I have loaded the dataset in spark, I was also able to transpose the columns with rows, so that columns are: week of year, product, amount. I thought this may help for using time-series analysis like in https://blog.cloudera.com/blog/2015/12/spark-ts-a-new-library-for-analyzing-time-series-data-with-ap... but I couldn't find how to forecast the next event (week of year) for a product/client.
Any help how to proceed will be appreciated.