Tree edit distance for hierarchical data compatible with HMIL paradigm

In contemporary data analysis in industrial and academic research we often need to work with data that has a hierarchical structure. The analysis of such data is naturally more difficult than the analysis of data with a flat structure because the schema of the hierarchically organized dataset may possess important information which is lost if we ignore it. A common task of a dataset analysis is to evaluate the difference between two of its instances. If the dataset has a flat structure and consists of numerical vectors, appropriate distance function from vector spaces can be used. Similarly, we can utilize the Levenshtein edit distance[1] for comparison of two strings, etc.