How one can match similar sounding features of two items

I am trying to build simple app which will compare two products from provideid URLs based on their features. We will be showing the products in tabular form with columns being each product and rows being features. The problem here is that we want to accept URLs from different shopping websites. We are able to fetch features from given URLs and get them as key – value dictionary, with key being feature name (‘RAM Size’ for laptop) and value be feature value (‘8GB’). Difficulty is how can we match different features names from different shopping sites with same meaning? Like how we can map ‘RAM Size’ from one shopping site to ‘Memory size’ from other shopping site? We want to build app for several categories which have varying features like a pressure cooker might have volumne specified in litres. How can we generalise such matching of features for different kinds categories?

Do we have to go through the route of natural language processing? Or we will be better off doing manual mapping, even though a bit tedious? Is there any NLP library which can ease NLP based implementation?