LEARNING SYSTEM OF HIERARCHIC NEURAL NETWORK

Title:

LEARNING SYSTEM OF HIERARCHIC NEURAL NETWORK

Document Type and Number:

Japanese Patent JPH0636061

Kind Code:

A

Abstract:

PURPOSE: To shorten the learning period of time by regarding, for example, time-series data as an object and reducing learning data concerning a prediction system which utilizes the hierarchic neural network and predicts an output result corresponding to unknown input data.

CONSTITUTION: The neural network consisting of an input layer, 1 or more intermediate layers, and an output layer is equipped with a variation rate calculating means 1 which calculates the variation rate of the data and constitutes so as to let the hierarchic neural network 2 learn the learning data including the calculated variation rate.

Inventors:

NOZAKI CHIKANORI
MURAI JUNKO
NAKATANI YOSHIHIRO

Application Number:

JP19364292A

Publication Date:

February 10, 1994

Filing Date:

July 21, 1992

Export Citation:

Click for automatic bibliography generation Help

Assignee:

FUJITSU LTD

International Classes:

G06F15/18; G06G7/60; G06N3/04; G06N3/08; G06N99/00; (IPC1-7): G06G7/60; G06F15/18

Attorney, Agent or Firm:

Yoshiyuki Osuge (1 outside)

Next Patent: INVERSE PROBLEM ANALYTIC NETWORK