【來自SCI的前沿論文】———User-friendlyoptimization approach of fed-batch fermentation conditions for theproduction of iturin A using artificial neural networks and supportvector machine
作（著）者：陳福迪
導師：王斌
學術期刊或出版社名稱：ElectronicJournal of Biotechnology
期刊封面：
期刊簡介：ElectronicJournal ofBiotechnology期刊于1998年創辦，隸屬于Elsevier集團，收錄包括分子生物學、生物化學、地球環境科學以及計算應用科學等多個生物技術相關學科文章，2015年影響因子為1.403，為SCI-E和EI雙收錄。
DOI：http://dx.doi.org/10.1016/j.ejbt.2015.05.001
論文主要内容：
Background: In thefield of microbial fermentation technology, how to optimize thefermentation conditions is of great crucial for practicalapplications. Here, we use artificial neural networks (ANNs) andsupport vector machine (SVM) to offer a series of effectiveoptimization methods for the production of iturin A. Theconcentration levels of asparagine (Asn), glutamic acid (Glu) andproline (Pro) (mg/L) were set as independent variables, while theiturin A titer (U/mL) was set as dependent variable. Generalregression neural network (GRNN), multilayer feed-forward neuralnetworks (MLFNs) and the SVM were developed. Comparisons were madeamong different ANNs and the SVM.
Results:The GRNN has the lowest RMS error (457.88) and the shortest trainingtime (1 s), with a steady fluctuation during repeated experiments.
The MLFNs havecomparatively higher RMS errors and longer training times, which havea significant fluctuation with the change of nodes.
In terms of the SVM,it also has a relatively low RMS error (466.13), with a shorttraining time (1 s).
Conclusion:According to the modeling results, the GRNN is considered as the mostsuitable ANN model for the design of the fed-batch fermentationconditions for the production of iturin A because of its highrobustness and precision, and the SVM is also considered as a verysuitable alternative model. Under the tolerance of 30%, theprediction accuracies of the GRNN and SVM are both 100% respectivelyin repeated experiments.
作者簡介：陳福迪，伟德 官网海洋生物學2014級研究生，主要研究方向為生态毒理學建模，碩士期間曾獲得2015和2016年研究生國家獎學金，參與項目《基于IPv6的水産品物聯網試驗系統的設計與應用》，畢業論文題目為《有機磷酸酯對雙齒圍沙蠶幼體急性毒性預測模型的研究》。
作者生活照
