Method and system for creating ordered reading lists from unstructured document sets
AssigneeXerox Corporation
InventorsRobert J. St. Jacques, JR., Mary Catherine McCorkindale, Saurabh Prabhat
Life of this patent
Estimating the term…
Abstract
Systems and methods for organizing a repository of unstructured documents into groups of ordered reading lists, i.e., document trails, comprising an ordered list of documents that relate to each by subject matter. Text analytics and natural language processing sets group documents, chose the most important/relevant documents from each group, and organize the documents into a suggested reading order. Additionally, documents within each document trail may be marked up or highlighted to indicate which paragraphs therein contain novel or useful information, or information that is not useful or redundant
Claims17
A method for creating an ordered reading list for a set of documents, comprising:
- analyzing a corpus of the document set in accordance with preselected text analytics and natural language processing steps for identifying a grouping relationship and clustering the set into a plurality of cluster groups;
- prioritizing the documents in a one of the cluster groups in relation to importance to a topic of the cluster group;
- and
- organizing the documents in accordance with the prioritizing into the ordered reading list as a document trail for sequential access to a reader of the document set.
A system for processing a repository of documents comprising:
- a clustering processor for segregating the documents based on topics identified in the documents and relating a plurality of the documents into a cluster group;
- and,
- a prioritizing processor for organizing the documents in the cluster group based upon relative probability of relationship to the topic by sequencing the documents into an ordered reading list in accordance with the relative probability.
Dependencies are inferred from the claim text.