Ingestion plan based on table uniqueness

Patent metadata

Field	Value
Publication number	US-9720945-B2
Application number	US-201615140976-A
Country	US
Kind code	B2
Filing date	Apr 28, 2016
Priority date	Nov 5, 2015
Publication date	Aug 1, 2017
Grant date	Aug 1, 2017

Abstract

Embodiments of the present invention disclose a method for processing tabular data. In various embodiments, an electronic document is received through a network, along with associated metadata. A plurality of table markers, or tabular data markers, are identified, in response to analyzing the received electronic document for said markers. References and citations associated with the plurality of tabular data markers are identified. A graphical representation of the relationship between identified tabular data markers and the identified references is generated. A uniqueness score is calculated, based on the generated graph and an ingestion plan is generated for the received electronic documents based on the calculated uniqueness score value.

First claim

What is claimed is:

1. A computer implemented method for processing tabular data, the method comprising:
receiving a plurality of electronic documents on a computer through a network, the plurality of electronic documents being stored on a remote server, the network being an internet connection;
receiving a plurality of metadata from the remote server through the network, the metadata being a plurality of identifying information associated with the received plurality of electronic documents;
indexing the received plurality of electronic documents and the received plurality of metadata in a data store;
identifying a plurality of tabular data markers, in response to analyzing the received electronic document and associated metadata;
identifying references for association with the identified plurality of tabular data markers by natural language analysis;
generating a graphical representation of the relationship between the identified tabular data markers and identified references, the graphical representation comprising a plurality of inbound directional edges and a plurality of vertices, wherein the directional edges are based on the identified references having an amplitude based on a count of identified references, and the vertices of the plurality of vertices are tabular data of the identified references;
calculating a uniqueness score value based on the generated graphical representation, the uniqueness score comprising a first value based on the plurality of inbound directional edges and a second value based on the plurality of vertices;
modifying the calculated uniqueness score based on one or more of:
a first count based on the directional edges to a vertex in the graphical representation;
multiplying the uniqueness score by zero, in response to the count of direction edges not exceeding a threshold; and
in response to input by a user, a second count based on the of vertexes in the graphical representation exceeding a threshold; and

generating an ingestion plan for the received electronic documents for display based on the calculated uniqueness score value, the ingestion plan comprising an ordered list of the received plurality of electronic documents.

Assignees

• IBM

Inventors

• Bastide Paul R
• Broomhall Matthew E
• Byron Donna K
• Loredo Robert E
• Pikovsky Alexander

Classifications

• G06F16/2228

  Indexing structures · CPC title

• G06V30/412 Primary

  Layout analysis of documents structured with printed lines or input boxes, e.g. business forms or tables · CPC title

• G06F18/2323

  based on graph theory, e.g. minimum spanning trees [MST] or graph cuts · CPC title

• G06F17/30477

  Physics · mapped topic

• G06F17/30958

  Physics · mapped topic