- 學位論文
拉曼光譜法檢測農產品中農藥的準確性:系統評價和薈萃分析
Accuracy of Raman Spectroscopy for Detection of Pesticides in Agricultural Products: a Systematic Review and Meta-Analysis
摘要
本研究透過Pub Med、Web of Science、EMbase文獻資料庫及Google Scholar，進行標準化關鍵字MeSH搜尋，自1999年至2020年發表之英文文獻期刊，共尋找出11篇符合主題條件之文章納入薈萃分析中。其內容數據需涵蓋斜率 (Slope)、相關性 (Correlation)、及樣本數 (n)。通過Comprehensive Meta-Analysis (CMA)隨機效應模型進行拉曼光譜檢測農藥是否具有與GC/LC檢測方式之相關準確性，分別進行12筆及11筆資料分析，並透過Q檢定、量化不一致性質 (I-squared , I2)檢驗文獻中之異質性，12筆及11筆資料之Q值皆為p-value < 0.001，結果顯示出樣品間的異質性；其I2分別為95.8、82.2，亦顯示出高異質性，判定使用隨機模式進行薈萃分析。進行分析結果相關性 (Correlation)分別為0.995及0.993；95% C.I.分別為0.977~0.999及0.979~0.998。結果表明，拉曼光譜在農藥檢測上具有其相關準確性。進一步分析影響拉曼光譜準確性因子，分析結果顯示對於不同類農藥建議配置適合種類的拉曼光譜進行檢測以達到最佳振動靈敏度與精確度；另農藥在低濃度時可能會影響拉曼光譜檢測質穩定度，但多數農藥在拉曼檢測下是具有與GC/LC之相關準確性。
並列摘要
In this study, through PubMed, Web of Science, EMbase literature database, and Google Scholar, a standardized keyword Mesh search was conducted. From 1999 to 2019, English literature journals were found, and a total of 11 articles that met the subject conditions were found and included in the meta-analysis. The content data needs to cover slope, correlation, and the number of samples . The Comprehensive Meta-Analysis (CMA) random effect model was used to detect Raman spectroscopy to determine whether the pesticide has the accuracy-related to the GC/LC detection method, and 12 and 11 data analyses were performed respectively. The inconsistent nature was quantified through the Q test (I-squared, I2) test the heterogeneity in the literature. The Q value of this study is p value < 0.001. The results show the heterogeneity between samples; its I2 is 95.8, 82.2, also shows high heterogeneity. The meta-analysis results showed that the correlations (Correlation) were 0.995 and 0.993; 95% C.I. were 0.977-0.999 and 0.979-0.998, respectively. The results show that Raman spectroscopy has its relative accuracy in pesticide detection. Also, further analysis affects the accuracy factor of the Raman spectrum. The analysis results show that it is recommended to configure a suitable Raman spectrum for different types of pesticides to achieve the best vibration sensitivity and accuracy; besides, pesticides may affect the Raman spectrum at low concentrations. Detection quality stability, but most pesticides correlate with accuracy and GC/LC under Raman detection.