| Summary: | 碩士 === 國立屏東科技大學 === 食品科學系 === 87 === Principle Component Analysis （PCA）, a statistical method, has been used broadly to analyze data in which similar nature of data will be closely or roughly clustered together, while rest of data may be scattered around the graph. To obtain well results of clustering, PCA needs enough input data. Besides, it shows no quantitative information that cannot help to interpret the relationship between each datum.
In this paper, Grey Relational Analysis （GRA） was proposed to apply to assist PCA for data clustering, identification of food system, and analysis of the food quality. The applications have been extended to the adulteration problems of soybean sauce, sesame oil, and multiple attribute decision making that can be used for choosing the proper brand of market milk powders.
The experimental results show that GRA is capable of identifying and ordering those similar nature of input samples based on standard data that we have built in data-base. Therefore, GRA can be regarded as an efficient technology for food analysis.
|
|---|
