A novel fast method for identifying the origin of Maojian using NIR spectroscopy with deep learning algorithms.

Maojian is one of China's traditional famous teas. There are many Maojian-producing areas in China. Because of different producing areas and production processes, different Maojian have different market prices. Many merchants will mix Maojian in different regions for profit, seriously disrupting the healthy tea market. Due to the similar appearance of Maojian produced in different regions, it is impossible to make a quick and objective distinction. It often requires experienced experts to identify them through multiple steps. Therefore, it is of great significance to develop a rapid and accurate method to identify different regions of Maojian to promote the standardization of the Maojian market and the development of detection technology. In this study, we propose a new method based on Near infra-red (NIR) with deep learning algorithms to distinguish different origins of Maojian. In this experiment, the NIR spectral data of Maojian from different origins are combined with the back propagation neural network (BPNN), improved AlexNet, and improved RepSet models for classification. Among them, improved RepSet has the highest accuracy of 99.30%, which is 8.67% and 0.70% higher than BPNN and improved AlexNet, respectively. The overall results show that it is feasible to use NIR and deep learning methods to quickly and accurately identify Maojian from different origins and prove an effective alternative method to discriminate different origins of Maojian.

Grouped-sampling technique to deal with unbalance in Raman spectral data modeling.

Due to limitations in disease prevalence and hospital specificity, spectral data are often collected with unbalanced sample size. To solve this problem, a new sampling method - grouped-sampling was proposed in this research, which is shown to be effective for unbalanced data. It avoids over-fitting of over-sampling and overcomes under-sampling utilization of under-sampling. In this study, we applied grouped-sampling to two unbalanced datasets where the sample proportions are 199:40 and 75:225. And then verified from two classic models: PCA-SVM (Principal Component Analysis-Support Vector Machine) and the deep learning algorithm GoogLeNet. The accuracy of these two datasets were 85.11% and 96.15% in PCA-SVM and 85.10% and 84.61% on GoogLeNet. Also, the F1-score were evaluated to measure the classification balance of sampling method, and result shows that F1-score of grouped-sampling is always the highest compared to over-sampling and under-sampling. In summary, compared to traditional sampling methods, grouped-sampling performs better on prediction for classes with smaller sample size, which means grouped-sampling can improve the balance of classification results and the potential of practical application. Therefore, we develop a group sampling method that distinguishes between under- and over-sampling, which greatly improves the accuracy and balance of predictions for unbalanced samples.