Examination of discrimination of multivariate data by quality engineering discrimination scale and application to quality control

2021 Fiscal Year Final Research Report

• Project/Area Number: 19K04891
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 25010:Social systems engineering-related
• Research Institution: Nihon University
• Principal Investigator: YANO Koya 日本大学, 生産工学部, 教授 (30514153)
• Co-Investigator(Kenkyū-buntansha): 中島 尚登 東京慈恵会医科大学, 医学部, 准教授 (90207788)
• Project Period (FY): 2019-04-01 – 2022-03-31
• Keywords: 品質工学 / 多変量解析 / パターン認識 / 品質管理

Outline of Final Research Achievements

In recent years, big data has been used, and identification, discrimination, classification, and prediction have been performed using a large amount of data such as images and audio information. However, the conventional statistical analysis method requires more sample data than the item information in the data. Therefore, there are often problems that the number of data required for analysis is not available or that there is too much item information to handle the analysis. In addition, although efficiency and accuracy are required in pattern recognition, appropriate analysis is often not possible due to the large amount of item information and the imbalance in the number of sample data. Therefore, by using the transcription function of quality engineering, reducing a large number of item information to two items and analyzing with two types of statistics, we were able to perform accurate identification, discrimination, and classification.

Free Research Field

品質工学

Academic Significance and Societal Importance of the Research Achievements

大量のデータの解析には従来は多変量解析という手法が適用されてきたが、データ数や情報量に制約がつくケースが多かった。ここでは品質工学で用いる転写という概念を適用し、多数の情報を全て2に圧縮し、精度を落とさずに2つの統計量で1つの尺度にまとめることを可能とし、識別、判別、分類等を可能とした。
対象分野は無数にあるが、一例として医学データや品質管理上の成分分析で実施を行い、効率の良い的確な類型分類を可能とした。数万以上のデータも2変量に圧縮し、最終的に1つの尺度で議論が可能なために、数値管理も容易になることが期待できる。