We are looking at using R to enhance our spend analysis engine – we have a massive body of training data that we use to classify invoices lines from companies to specific categories (using the UNSPSC taxonomy). We want to see if we can improve the process by using R statistical approach.
The data quality that we get is patchy but all data should have a supplier name and description that can then be matched against our vast training set
we are looking to develop a prototype of this and would like a practical academically minded person to help develop this (or reuse what already exists). We have done some research and understand that the R would be a good tool to use existing libraries to effect a quick result
We are based in London and would prefer to work in our own time zone (give or take a few hours) if possible with a prospective freelancer – but we would not need you to travel to London to do the job (it can be done ‘from home’)