Affiliation:
1. College of Information Science and Technology, Beijing University of Chemical Technology, Beijing, China
2. Intelligence and Information Engineering College, Tangshan University, Tangshan, China
Abstract
Background:
The small sample problem widely exists in the fields of the chemical in-dustry, chemistry, biology, medicine, and food industry. It has been a problem in process modeling and system optimization. The aim of this study is to focus on the problems of small sample size in modeling, the process parameters in the ultrasonic extraction of botanical medicinal materials can be obtained by optimizing the extraction rate model. However, difficulty in data acquisition results in problem of small sample size in modeling, which eventually reduces the accuracy of modeling prediction.
Methods:
A virtual sample generation method based on full factorial design (FFD) is proposed to solve the problem ofa small sample size. The experiments are first conducted according to the Box-Behnken design (BBD) to obtain small-size samples, and the response surface function is establis-hed accordingly. Then, virtual sample inputs are obtained by the FFD, and the corresponding virtual sample outputs are calculated by the response surface function. Furthermore, a screening method of virtual samples is proposed based on an extreme learning machine (ELM). The connection weights of ELM are used for further optimization and screening of the generated virtual samples.
Result:
The results show that virtual sample data can effectively expand the sample size. The preci-sion of the model trained on semi-synthetic samples (small-size experimental simples and virtual samples) is higher than the model trained merely on small-size experimental samples.
Conclusion:
The virtual sample generation and screening methods proposed in this paper can effec-tively solve the modeling problem of small samples. The reliable process parameters can be ob-tained by optimizing the model trained by the semi-synthetic samples.
Funder
National Natural Science Foundation of China
Publisher
Bentham Science Publishers Ltd.
Subject
Drug Discovery,General Medicine
Reference21 articles.
1. Zhao K.L.; Jin X.L.; Wang Y.Z.; Survey on few-shot learning. J Softw 2021,32(2),349-369
2. He P.; Sun F.; Hu X.F.; Lin Y.P.; Duan S.K.; Optimization of prediction system for sample laser cutting process parameters. Las J 2021,42(12),170-175
3. Sheng H.; Liu X.; Bai L.; Dong H.; Cheng Y.; Small sample state of health estimation based on weighted Gaussian process regression. J Energy Storage 2021,41(1)
4. Lv Y.Q.; Min W.Q.; Duan H.; Jiang S.Q.; Few-shot food recognition combining triplet convolutional neural network with relation network. Comput Sci 2020,47(1),136-143
5. Guo X.P.; Song Y.F.; Liu S.H.; Gao M.H.; Qi Y.; Shang X.Q.; Linking genotype to phenotype in multi-omics data of small sample. BMC. Geno 2021,22(1),1-11