RESUMO
The molecularly imprinted polymer (MIP) is useful for measuring the amount of riboflavin (vitamin B2), in various samples using UV/Vis instruments. The practical optimization of the MIP synthesis conditions has a number of drawbacks, like the need to spend money, the need to spend time, the use of the compounds that cause contamination, needing laboratory equipment and tools. Using machine learning (ML) to predict the amount of riboflavin absorbance is a creative solution to overcome the problems and shortcomings of optimizing polymer synthesis conditions. In fact, by using the model without needing real work in the laboratory, the optimum laboratory conditions are determined, and as a result the maximized absorption of the riboflavin is obtained. In this paper, MIP was synthesized for selective extraction of the riboflavin, and UV/Vis spectrophotometry was used to quantitatively measure riboflavin absorbance. Various factors affect the performance of the polymer. The effect of six important factors, including the molar ratio of the template, the molar ratio of monomer, the molar ratio of cross-linker, loading time, stirring rate, and pH, were investigated. Then, using ensemble ML algorithms, like gradient boosting (GB), extra trees (ET), random forest (RF), and Ada boost (Ada) algorithms, an accurate model was created to predict the riboflavin absorption. Also, the mutual information feature selection method was used to determine the important features. The results of using feature selection method was shown that variables such as the molar ratio of the template, the molar ratio of the monomer, and the molar ratio of the cross-linker had a high effect on riboflavin absorbance. The GB and Ada boost algorithms performed better than ET and RF algorithms. After tuning the n-estimator hyper parameter (n-estimator = 300), the GB algorithm was shown an excellent performance in predicting the absorbance of riboflavin and the maximum R2-scoring of the model was obtained at 0.965995, the minimum of the mean absolute error (MAE), and mean square error (MSE) of the model respectively were obtained -0.003711 and -0.000078. Therefore, by using the proposed model, it is possible to predict riboflavin absorbance theoretically, and with high accuracy by changing the inputs of model, and using the model instead of working in the lab saves time, money, chemical compounds, and lab ware.
RESUMO
The molecularly imprinted polymers are artificial polymers that, during the synthesis, create specific sites for a definite purpose. These polymers due to their characteristics such as stability, easy of synthesis, reproducibility, reusability, high accuracy, and selectivity have many applications. However, the variety of the functional monomers, templates, solvents, and synthesis conditions like pH, temperature, the rate of stirring, and time, limit the selectivity of imprinting. The Practical optimization of the synthetic conditions has many drawbacks, including chemical compound usage, equipment requirements, and time costs. The use of machine learning (ML) for the prediction of the imprinting factor (IF), which indicates the quality of imprinting is a very interesting idea to overcome these problems. The ML has many advantages, for example a lack of human error, high accuracy, high repeatability, and prediction of a large amount of data in the minimum time. In this research, ML was used to predict the IF using non-linear regression algorithms, including classification and regression tree, support vector regression, and k-nearest neighbors, and ensemble algorithms, like gradient boosting (GB), random forest, and extra trees. The data sets were obtained practically in the laboratory, and inputs, included pH, the type of the template, the type of the monomer, solvent, the distribution coefficient of the MIP (KMIP), and the distribution coefficient of the non-imprinted polymer (KNIP). The mutual information feature selection method was used to select the important features affecting the IF. The results showed that the GB algorithm had the best performance in predicting the IF, and using this algorithm, the maximum R2 value (R2 = 0.871), and the minimum mean absolute error (MAE = - 0.982), and mean square error were obtained (MSE = - 2.303).
