A Study of Knowledge Model Development using Fuzzy Model

利用模糊模型开发知识模型的研究

• 批准号: 17500135
• 负责人: RYOKE Mina
• 金额: $ 2.28万
• 依托单位: University of Tsukuba
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2005
• 资助国家: 日本
• 起止时间: 2005 至 2007
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-17500135/
• 关键词: Fuzzy Model; Knowledge Model; Systems Methodology; Risk Management; Consumer Behavior; Ddemand Prediction; 惣菜部門; 領域拡張アルゴリズム; ベイズ推測

The purpose of this study is to propose a knowledge model development in order to predict demand and to manage a section. The typical case of the situation is the fresh food section in a supermarket which is under complex and dynamically changing circumstance. The section is focused. The knowledge model is developed by combining ensemble model to treat the fluctuation on evaluation and context model to treat the context of complex situation. A systems methodology "I-System" for knowledge integration and creation is applied to develop the knowledge model in order to manage the demand prediction in the fresh food section. The concrete approach is to divide the problem into three domains consisting of the manager's knowledge, the consumer behavior and the demand prediction and to clarify them. In the domain of the manager's knowledge, two interview surveys are carried out in order to collect the manager's knowledge on the section management and decision making items and so forth. The obtained knowledge are illustrated using KJ method, text mining techniques and Bayesian Theorem etc. In the domain of the customer behavior, a questionnaire survey is carried out via Internet in order to collect their thinking on the fresh food section. The obtained data is analyzed to find the important factors when the customers select flesh foods. Then, comparison analysis between demand side and supply side thinking is done. The results of two domains are integrated to build the demand prediction model which is different from the conventional statistical model. At the same time, the system is also developed. The system supports the manager's decision making to consider the chance loss and disposal loss.

本研究的目的是提出一种知识模型开发，以预测需求并管理某个部分。这种情况的典型案例是超市的生鲜食品区，其环境复杂且动态变化。该部分是重点。知识模型是通过结合处理评估波动的集成模型和处理复杂情况上下文的上下文模型来开发的。应用用于知识集成和创建的系统方法“I-System”来开发知识模型，以管理生鲜食品部分的需求预测。具体做法是将问题分为管理者知识、消费者行为和需求预测三个领域并加以阐明。在管理者知识领域，通过两次访谈调查，收集管理者对部门管理、决策事项等方面的知识掌握情况。利用KJ方法、文本挖掘技术和贝叶斯定理等对所获得的知识进行说明。在顾客行为领域，通过互联网进行问卷调查，收集顾客对生鲜食品的看法。对获得的数据进行分析，找出顾客选择肉类食品时的重要因素。然后，对需求方和供给方的思路进行了比较分析。整合两个领域的结果来构建不同于传统统计模型的需求预测模型。与此同时，系统也得到了发展。该系统支持管理者的决策时考虑机会损失和处置损失。

项目成果

Analysis of Demand-side Requirement and Supply-side Management on Fresh Side Dish Section

生菜板块需求侧需求与供给侧管理分析

• 发表时间: 2006
• 期刊: Proc. of Information Processing and Managenent of Uncertainty in Knowledge-Based Systems
• 作者: Mina Ryoke;Kiichiro Hori;Yukihiro Yamashita;Yoshiteru Nakamori
• 通讯作者: Yoshiteru Nakamori

Knowledge Discovery of Interview Survey on Fresh Food Management

生鲜管理访谈调查的知识发现

• 发表时间: 2005
• 作者: M. Ryoke;Y. Yamashita;K. Hori and Y. Nakamori
• 通讯作者: K. Hori and Y. Nakamori

Quantification of Relationships between Evaluation Targets and Adjectives through Fuzzy Correspondence Analysis and OWA Operator

通过模糊对应分析和OWA算子量化评价目标与形容词之间的关系

• 发表时间: 2007
• 作者: Ryoke;M.・Nagase;K.・Nakamori;Y.
• 通讯作者: Y.

A Context-dependent Knowledge Model for Evaluation of Regional Environment

区域环境评价的上下文相关知识模型

• 发表时间: 2005
• 期刊: Environmental Modeling&Software 20
• 作者: S.;Kawano;V;N.;Huynh;M.;Ryoke;Y.;Nakamori
• 通讯作者: Nakamori

A Study on Fresh Food Management System

生鲜食品管理体系研究

• 发表时间: 2007
• 作者: Y. Yamashita;M. Ryoke and Y. Nakamori
• 通讯作者: M. Ryoke and Y. Nakamori