Novel near-infrared spectrum analysis tool: Synergy adaptive moving window model based on immune clone algorithm

Self-expansion full information optimization strategy: Convenient and efficient method for near infrared spectrum auto-analysis

An essential step in the application of near infrared spectroscopy technology is the spectrum preprocessing. A reasonable implementation of it ensures that the effective spectral information is correctly extracted and, also that the model's accuracy is increased. However, some analysts' research still uses the manual approach of trial and error, particularly those less skilled ones. Previous papers have provided preprocessing optimization algorithms for NIR, but there are still some problems that need to be resolved, such as the unwieldy sequence determination of preprocessing method or, the fluctuated optimization outcomes or, lack of sufficient statistical information. This research suggests a spectrum auto-analysis methodology named self-expansion full information optimization strategy, a new powerful open-source technique for concurrently addressing all of these above issues simultaneously. For the first time in the field of chemometrics, this algorithm offers a reliable and effective automatic near infrared auto-modelling method based on the statistical informatics. With the aid of its built-in modules, such as information generators, spectrum processors, etc., it is able to fully search the common preprocessing techniques, which is determined by Monte Carlo cross validation. Then the final ensemble calibration model is built by employing the optimized preprocessing schemes, along with the wavelength variables screening algorithm. The optimization strategy can offer the user objective useful statistics information created throughout the modeling process to further examine the model's effectiveness. The results demonstrate that the suggested method can easily and successfully extract spectrum information and develop calibration models by putting it to the test on two groups of actual near-infrared spectral data. Additionally, this optimization strategy can also be applied to other spectrum analysis areas, such Raman spectroscopy or infrared spectroscopy, by changing a few of its parameters, and has extraordinary application value.

A powerful tool for near-infrared spectroscopy: Synergy adaptive moving window algorithm based on the immune support vector machine

Traditional trial-and-error methods are time-consuming and inefficient, especially very unfriendly to inexperienced analysts, and are sometimes still used to select preprocessing methods or wavelength variables in near-infrared spectroscopy (NIR). To deal with this problem, a new optimization algorithm called synergy adaptive moving window algorithm based on the immune support vector machine (SA-MW-ISVM) is proposed in this paper. Following the principle of SA-MW-ISVM, the original problem of calibration model optimization is transformed into a mathematical optimization problem that can be processed by the proposed immune support vector machine regression algorithm. The main objective of this optimization problem is the calibration model performance; meanwhile, the constraint conditions include a reasonable spectral data value, spectral data preprocessing method, and calibration model parameters. A unique antibody structure and specific coding and decoding method are used to achieve collaborative optimization in NIR spectroscopy. The tests on four actual near-infrared datasets, including a group of gasoline and three groups of diesel fuels, have shown that the proposed SA-MW-ISVM algorithm can significantly improve the calibration performance and thus achieve accurate prediction results. In the case of gasoline, the SA-MW-ISVM algorithm can decrease the prediction error by 44.09% compared with the common benchmark partial least square (PLS). Meanwhile, in the case of diesel fuels, the SA-MW-ISVM algorithm can decrease the prediction error of cetane number, freezing temperature, and viscosity by 9.99%, 28.69%, and 43.85%, respectively, compared with the PLS. The powerful prediction performance of the SA-MW-ISVM algorithm makes it an ideal tool for modeling near-infrared spectral data or other related application fields.

System optimisation quantitative model of on-line NIR: a case of Glycyrrhiza uralensis Fisch extraction process

INTRODUCTION

The on-line analysis of active pharmaceutical ingredients (APIs) during the extraction process in herbal medicine is a challenge. Establishing a reliable and robust model is a critical procedure for the industrial application of on-line near-infrared (NIR) technology.

OBJECTIVE

To evaluate the advantages of on-line NIR model development using system optimisation strategy, Glycyrrhiza uralensis Fisch was used as a case. The content of liquiritin and glycyrrhizic acid was monitored during pilot scale extraction process of Glycyrrhiza uralensis Fisch in three batches.

METHODS

High-performance liquid chromatography (HPLC) was used as reference method for content determination of liquiritin and glycyrrhizic acid. The quantitative models of on-line NIR were developed by system optimisation of processing trajectory. For comparison, the models were simultaneously developed by stepwise optimisation. Moreover, the modelling parameters obtained through system optimisation and stepwise optimisation were reused in three batches. Root mean square error of prediction (RMSEP) and residual predictive deviation (RPD) were used to assess the model quality.

RESULTS

The average values of RMSEP and RPD of systematic model for liquiritin in three batches were 0.0361, 4.1525 (first batch), 0.0348, 4.7286 (second batch) and 0.0311, 4.9686 (third batch), respectively. In addition, the modelling parameters of systematic model for glycyrrhizic acid in three batches were same, and the average values of RMSEP and RPD were 0.0665 and 5.2751, respectively. The predictive performance and robustness of systematic models for the three batches were better than the comparison models.

CONCLUSION

The work demonstrated that system optimisation quantitative model of on-line NIR could be used to determine the contents of liquiritin and glycyrrhizic acid during Glycyrrhiza uralensis Fisch extraction process.