We have a large collection of documents (D), each accompanied by a set of metadata (M). Within this collection, some documents act as parent documents and have multiple child documents. Both parent and child documents are part of the document set D. The number of child documents can vary for each parent document. In the past, humans have manually sorted the child documents of every parent document based on the discretion of the parent and child metadata. Our objective is to develop a machine learning (ML) model that can learn this sorting criteria and predict the sequence of child documents attached to a parent document, utilizing both parent and child metadata (M). Essentially, we aim to infer the relative ordering of child documents associated with a parent.
Currently, we possess a dataset structured as M(Parent), M(Children), Sort_Order. However, we can regenerate/rearrange the dataset to meet the required format. Given this scenario, what strategy should we employ to address this problem?
