I need a product-to-product input-output (I-O) table to conduct environmentally extended I-O analysis. I have supply and use tables at basic prices relating different industries and products, but there are a larger number of products (~200) than industries (~100). I also have a square I-O table relating different industries (~100 X 100). What I want to do is to use the supply and use tables to build a disaggregated product-to-product table, i.e., the products should not be aggregated into industries in order to avoid losing resolution. I have found a couple of methods to do this but (e.g., Almon's method), as far as I know, all of them do aggregate different products to industries. Is there any statistically sound approach to build this type of table?