Browse Prior Art Database

Inductive Inference Method of Time Serial Pattern Extraction From Time Serial Information

IP.com Disclosure Number: IPCOM000040364D
Original Publication Date: 1987-Nov-01
Included in the Prior Art Database: 2005-Feb-02
Document File: 2 page(s) / 13K

Publishing Venue

IBM

Related People

Oda, N: AUTHOR

Abstract

This article describes an algorithm for determining a common time-serial event pattern from a group of qualitative time-serial event information items. The algorithm uses an inductive inference method to determine the pattern, and has the capability for the evaluation of any information items which include contradictions in them. This algorithm is more efficient and simpler than using statistics or graph theory. DEFINITION Time-Serial Information (TSI) A list of independent facts aligned with their time-serial appearance sequence observed from one observation object. Non-Time-Serial Facts (NTSF) A list of facts whose appearance sequence can not be identified by observation.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 66% of the total text.

Page 1 of 2

Inductive Inference Method of Time Serial Pattern Extraction From Time Serial Information

This article describes an algorithm for determining a common time-serial event pattern from a group of qualitative time-serial event information items. The algorithm uses an inductive inference method to determine the pattern, and has the capability for the evaluation of any information items which include contradictions in them. This algorithm is more efficient and simpler than using statistics or graph theory. DEFINITION Time-Serial Information (TSI) A list of independent facts aligned with their

time-serial

appearance sequence observed from one observation

object. Non-Time-Serial Facts (NTSF)

A list of facts whose appearance sequence can not be

identified by observation.

Example of TSI and NTSF

{f1 (f2 f3) f4 f5} <--- TSI

(f2 f3) <-------------- NTSF Common Time-Serial
Pattern (CTSP) A time-serial pattern which includes all facts of

the group of TSIs and has no contradiction with

them. If there are common facts that contradict

the sequence, they are separated by a pair of

parentheses. And if there are any facts that

appear between a pair of the contradictory facts,

they are also separated by the same parentheses.

Example

TSI1 {f1 f2 f3 f4 f5}

TSI2 {f4 f2 f5}

CTSP of TSI1 and TSI2 is.. {f1 (f2 f3 f4) f5}

Contradiction State (CS) vs. Non-Contradiction State (NCS)

Example

{f1 f2 f3 f4} & {f1 f4 f2} <----- CS

{f1 (f3 f2) f4} & {f1 f2 f3} <--- NCS Re-ordering by the number of HEs' AEs (ROHA). Facts of a grouped NTSF are here referred to as

'elements'. Each element is highlighted one by

one, and for each highlighted element (HE) the

number of after elements (AEs), that is, the

number of elements which only appear after the

highlighted element, is determined. NTSF elements

are re-ordered in descending order of the

respective number of AEs.

Example

Population of TSIs

{f1 f4 f5}

{f1 f3 f5}

{f1 f2 f3 f5}

{f2 f3 f4 f5}

Obtained CTSP is ... {f1 (f4 f2 f3) f5}

1

Page 2 of 2

A group of NTSF (f4 f3 f2)

HE AF

f2 f3 f4

f3 f4...