权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

知識学習機構を持つ高次企業診断システムの構築

构建具有知识学习机制的高水平企业诊断体系

基本信息

批准号：
05680328
负责人：
山口高平
金额：
$ 0.83万
依托单位：
Shizuoka University
依托单位国家：
日本
项目类别：
Grant-in-Aid for General Scientific Research (C)
财政年份：
1993
资助国家：
日本
起止时间：
1993 至无数据
项目状态：
已结题

项目摘要

従来の企業診断モデルは,定量的な財務諸表知識のみを利用しており,高度な企業診断モデルとは言えなかった.本研究では,定性的な経営対策知識と定量的な財務諸表知識を統合して利用する企業診断モデルを設計実装し,さらに,学習機構を利用してそのモデルを維持管理する,高次企業診断システムの枠組みを検討した.本企業診断モデルの処理法としては,数値ではなく区間を伝播させる方法(制約充足アルゴリズム)と最良優先探索技法に基づく方法を提案した.区間伝播法については,1992年にヒボネンによって汎用アルゴリズムが既に提案されているが,変数や制約数が多い(以下の実験では228個の財務変数と116個の財務方程式)と対処できないため,業界平均値から算出した財務変数初期値を利用して区間値の組合せ数を減少させる事により,そのような場合でも実行可能な区間伝播法を考案した.また,本区間伝播法と財務変数の正常区間値からの逸脱度等から決定される最良優先探索技法を統合した方法をProlog言語で実装すると共に,3種類の実際の経営診断データに適用した結果,50MIPS程度の計算機上で約20-30分程度で処理され,全財務変数の正常区間値からの逸脱度が,最良のケースで1/3程度減少する事を確認した.さらに,所与の財務諸表知識と経営対策知識の融合知識ベースに含まれる誤りを自動的に改訂するために,誤りを同定するための3種類の戦略,及び,32個の知識洗練化戦略を整理し,机上ではあるが,それらが有効に働くことを確認した.

In recent years, enterprises have been shut down, and quantitative information tables have been made to know how to make use of them to make use of them. In this study, qualitative and quantitative information management tables are used in this study. In this study, qualitative and quantitative knowledge tables are used in this study. In this study, qualitative and quantitative knowledge tables are used in this study. in this study, qualitative and quantitative knowledge tables are used in this study. in this study, qualitative and quantitative knowledge tables are used in this study. In this enterprise, it is best to explore the techniques and methods to explore the basic methods. The method of zonal broadcasting was used to calculate the number of subscribers in 1992. In 1992, we proposed to use both local and local governments, and the number of customers was more than one (the following models were 228, 228, 116, and 116). The average number of customers in the world calculated the number of customers in the initial stage, and the number of business units in the area was used to set up a number of business models. It is possible to make a case study of the broadcasting method. In this area, the method of broadcasting in this area, the method of broadcasting in the normal area, the degree of escape in the normal area, and so on, decided that it is best to explore the techniques of the system first. The method of Prolog has been installed in common, and the results of the three kinds of international equipment have been used. The average score is about 20-30 points on the 50MIPS computer, and the degree of escape is about 20-30 points on the 50MIPS computer. The best thing is to make sure that you have a minimum of 3 degrees. This is the integration of the knowledge and policy of the registration table, and the knowledge integration of the knowledge and policy of the customer service table, which is integrated with the knowledge of the system, and the error is the same as that of the system, and the 32 knowledge washers are sorted out and checked on the machine, and there are no customers to confirm the information.