Identifying commodity groups at the intermediate level of aggregation to preserve maximum granularity in the output and trade data

Abstract

There are many studies successfully investigating international trade and output produced within a country. However, most of papers deal solely with the former or the latter type of data. Those few works digging into both sources of data often have quite a high-light description of how two data sources were benchmarked one against the other. Same time there is a limitation for those interested in tackling this subject area. There is no publicly available classifier with the unambiguous matching in-between two sources, while the publicly available ones are based on the many-to-many matching principle. The absence of such unambiguous matching at the levels higher than the very granular one requires solving a problem of finding homogenous commodity groups at the interim level. But every researcher has to pass this way on his own. Thus, the findings of different papers can be hardly compared to each other. We are closing this gap by justifying transparent and reproducible approach to such matching. Moreover, we readily disclose the outcome of such a procedure in the form of a final classifier. More so, the value of our solution lies in that it allows to recreate the classifier subject to any changes in the initial data hierarchies. Such an outcome is guaranteed by our unique solution on how to map the raw data at the interim level of aggregation which is formed on the many‑to‑many principle at the level more aggregate, than the very granular one.