Very Large Knowledge Bases

This page lists some of the knowledge bases we are using with PARKA-DB.

Caper-20, Caper-100, Caper-200 - The CaPER knowledge bases are casebases from the CaPER Project. The casebases contain plans that were automatically generated using the Problem Generator for the transport logistics planning domain, UM Translog.

CIA World Factbook knowledge base contains geographical information that was extraced from the CIA World Factbook pages on the World-Wide Web using a web robot. The knowledge base also contains state and city data from the 1994 World Almanac and Book of Facts.

Knowledge Base Sizes:

Knowledge Base	Frames	Structural Links	Assertions	File Size (gzip)
Caper-20	11103	26404	118615	519k
Caper-100	56516	130526	807480	3210k
Caper-200	116297	266100	1635782	6477k
World Factbook	33335	18994	125913	563k

Knowledge Base Format

The format for the knowledge base files is a table of binary assertions. Each line contains a tab delineated binary assertion of the form:

(predicate#from-value#to-value)

For example:

(isa#Clyde#elephant)
(color#Clyde#white)

asserts that Clyde is an elephant, and the color of Clyde is white. The "A" indicates that the line contains an assertion.

This Sample knowledge base (8k gzip) contains 2000 assertions from the Caper-20 knowledge base.

The CIA World Fact Book can be downloaded here:

CIA World Fact Book - old version with cities and populations added.
CIA World Fact Book - new version with Uppercase letters, lists are kept together, and numbers have quantifiers (ft, mi, %).
CIA World Fact Book - new version with all lowercase, lists are split into separate assertions, and no quantifiers.
Contact Dr. Kilian Stoffel or Dr. James Hendler for more information and access to the complete knowledge bases.
Return to PARKA page
Web Accessibility