Modeling and quantitative structure-retention relationship (QSRR) studying of the constituents of Citrus. sinensis CV. Thamson extracted by gas chromatography-mass spectrometry using genetic algorithm-multiple linear regression

Document Type : Original Article

Author

• Saeed Nekoei

Shahrood

Abstract

Quantitative structure-retention relationship (QSRR) study to predict the kovats index (KI) of Citrus. sinensis CV. Thamson was performed using multiple linear regression (MLR). After extracting the essential oil and injecting it into the GC-MS device, its various compounds were identified. Then, in order to model and predict the quantum index (KI) values ​​of the compounds, first the structure of the compounds, the drawing and the appropriate group of descriptors were calculated. Then the step-wise selection method (SW) and genetic algorithm (GA) were used to obtain the best descriptors that were most related to the KI of the desired compounds. Multiple linear regression was constructed for linear modeling. Statistical data show that both SW-MLR and GA-MLR methods provide acceptable predictions. 

Keywords

• Quantitative structure-retention relationship
• genetic algorithm
• multiple linear regression

---