Abstract and Contents




Abstract and Contents

Models for Prediction

Origin of Model Selection Criteria

Criteria based on Trade-off of factors

Deriving Matchable Principal

Situation Decomposition

Whole situation and Partial situations

Case selection using Segment space

Matchability on Spread sheet form

Algorithm Overview

Segment selecting space without no-case

Searching on Sorted segments

Reject Saddle point

Simulation and Result

Powerful prediction using Matchable situations

Conclusions & Future work

作成者 :山川 宏