查詢結果分析
來源資料
相關文獻
- 藉單一米粒鮮度值檢驗一批稻米之新/舊混合狀態的統計分析
- 多重插補在民意調查中之應用及抽樣誤差之估計
- 利用混合分布模式求配兩自交系雜交試驗資料之方式估算數量性狀基因之效應(1)--方法與數例
- 因子分析中決定適當因子數--EM演算法與AIC、CAIC之應用
- The Valuation of Mortgage Insurance Contracts under Housing Price Cycles: Evidence from Housing Price Index
- A New Training Method for Speech Recognition
- NIG分配下之對稱性檢定
- 確定提撥制退休金之評價:馬可夫調控跳躍過程模型下股價指數之實證
- 基於多層級人形區塊HOGs之期望值最大化擁擠場景行人偵測
- 因素分析運用於股票與基金市場相關性探討
頁籤選單縮合
題名 | 藉單一米粒鮮度值檢驗一批稻米之新/舊混合狀態的統計分析=Statistical Analysis of the Freshness Measurements at Single-Grain Level when Examining the Blending Status of Fresh/ Old Grains in a Batch of Rice Grains |
---|---|
作者姓名(中文) | 林俊隆; 陳宗禮; | 書刊名 | 作物、環境與生物資訊 |
卷期 | 5:4 2008.12[民97.12] |
頁次 | 頁281-296 |
分類號 | 434.111 |
關鍵詞 | 單米一粒鮮度值; 新/舊米混合狀態; 常態混合分布的成分數; EM演算法; 循序的概度比測驗; 靴環重抽樣; Freshness measurements at single-grain level; Blending status of fresh/ old rice grains; Number of components of a normal mixture; EM algorithm; Sequential likelihood ratio test; Bootstrap re-sampling; |
語文 | 中文(Chinese) |
中文摘要 | 摘要 本研究之目的在建立分析酸鹼值或過氧 化酶活性檢測值等單一米粒鮮度檢驗值的統 計程序。首先,以過去研發稻米鮮度檢驗方 法時所獲得的單一米粒過氧化酶活性檢測值 的數據進行分析,結果顯示其機率分布概呈 常態分布。據此,我們以混合常態分布模式 為基礎,開發了一套統計分析程序與其SAS 巨集程式,以供檢驗一批稻米是否混有新鮮 度不同的米粒成分及各種新/舊米粒成分的 混入率,本研究同時循模擬試驗法評估這個 統計分析程序的效率。結果顯示,在舊米混 入率0~50% 及抽驗米粒數96~960 的範圍 內,這個統計分析程序冒犯第一型錯誤的經 驗機率,即將一批沒有混雜的純新米誤驗為 有新/舊米混合的機率,與預定的名義顯著水 準並無顯著的差異。其次,這個統計分析程 序的經驗檢驗力,即正確地驗出一批新/舊混 合米的機率,隨抽驗米粒數的增大而快速地 提升;當舊米混入率在10% 或以上時,只要 抽驗480 以上的米粒,檢驗力就可高達0.95 以上。這個統計分析程序對一批稻米內所包 含的米粒成分數與各成分的混入率之估計精 確度也呈現很好的漸進性質;但在舊米混入 率低至10%或以下的場合,估計效率就明顯 地偏低。 |
英文摘要 | ABSTRACT This study aims at developing a statistical procedure to analyze the freshness measurements at single-grain level, e.g. acidity or peroxidase activity, of a sample drawn from a batch of rice grains, thus its blending status of fresh/old grains can be determined. Data collected in previous studies revealed that normal model fitted the distribution of peroxidase activity measurements at single grain level very well. A statistical procedure based on normal mixture model and its corresponding SAS macro were developed to assess the number of components of different storage durations and estimate the mixing proportions of each components in a batch of rice grains. Simulation approach was then employed to validate this statistical procedure. Across a range of the mixing proportion of old grains from 0% to 50% and the number of rice grains examined from 96 to 960, the empirical type-one error rate, i.e., the probability of misclassifying a pure batch of fresh grains as a mixed batch offresh/old rice grains, was not significantly larger than the pre-set nominal significance level. The test power, i.e., the probability of correctly detecting a mixed batch of fresh/old rice grains, increased rapidly as the sample size getting larger. When the mixing proportion of old grains was 10% or higher, the test power could reach above 0.95 if 480 or more rice grains were examined. This statistical procedure also showed good asymptotic properties in estimating the number of components and the mixing proportions, but its efficiency was well below the desired level when the mixing proportion of old grains was as low as 10%. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。