Browse Prior Art Database

Method and system of building data cube by automatically constructing the hierarchy of dimension

IP.com Disclosure Number: IPCOM000235523D
Publication Date: 2014-Mar-06
Document File: 4 page(s) / 87K

Publishing Venue

The IP.com Prior Art Database

Abstract

This disclosure proposes method and system of building data cube by automatically constructing the hierarchy of dimension. The dimension in data cube is a kinds of hierarchy structure, e.g. The city name of Shanghai. There are several possible dimensions: 1) Economic dimension; 2) Location dimension: city->province->country; 3) Footbal dimension etc. In many situations, the hierarchy structure is hidden and not obviously stored in database. This disclosure aims to 1) automatically build the hierarchy of dimension; 2) based on the hierarchy, detect the outliers for decision support. It leverages the entity relationship built in the knowledge graph to discover the underlying semantic connections and hierarchy automatically.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 90% of the total text.

Page 01 of 4

Method and system of building data cube by automatically constructing the hierarchy of dimension


Background


1) The dimension in data cube is a kinds of hierarchy structure, e.g.


2) Time dimension: day->month->quarter->year


3) Location dimension: city->province->country


Problem


1) In many situations, the hierarchy structure is hidden and not obviously stored in database.


2) In an finance system, the column in a table only stores company name but no the hierarchy of company industry (e.g. Software, Computer hardware, Management Consulting Services).


3) In healthcare system, the column stores the disease names but no the disease hierarchy.


4) This disclosure aims to automatically build the hierarchy of dimension.


Core idea:


1) Automatically build the dimension hierarchy using open knowledge graph


2) Automatic select the relevant dimensions from target KPI from factor table

1


Page 02 of 4

2


Page 03 of 4


Main Step

3


Page 04 of 4


1) Choose the table columns


2) For each entity in table column, find the corresponding entity in knowledge base.


- Group based method to disambiguate the entity


3) Build the hierarchy from the entity set.


- Extract parent-child relationships


Claiming Points


1) automatic data cube hierarchy discovery and building that uses open-access structured data set to select and discover dimension hierarchy;
-entity detection & matching between a collection of text from a field and structured dataset from knowledge graph -dimension selection and filtering...