基于响应面法和机器学习算法分析不同条件下微藻生长影响的研究  

Analysis of the Effect of Microalgae Growth under Different Conditions Based on Response Surface Methodology and Machine Learning Algorithm

作　　者：高新颖 张静[1] 刘晶雪 冯亚男 张林林[1] 

机构地区：[1]山东科技大学安全与环境工程学院,山东 青岛

出　　处：《交叉科学快报》2024年第3期398-409,共12页Interdisciplinary Science Letters

摘　　要：目前随着人工智能的快速发展,机器学习已经广泛应用于各个领域,并对实验中获得的小数据进行模拟分析。本研究以微藻为基础,利用收集的9组文献中对微藻的研究结果为数据,使用四种算法,即BP神经网络、支持向量机、随机森林和径向基函数神经网络,使用MATLAB软件进行建模和分析,与原文献中的响应面分析法进行对比,得出以下主要结论:通过比较相对系数(R2)、平均绝对误差(MAE)、平均偏差误差(MBE)、均方根误差(RMSE)和均方误差(MSE),相比于传统的响应面分析法,机器学习算法表现出更好的预测效果,其中随机森林和径向基函数神经网络的相对系数最接近于1,预测效果最好,其次是BP神经网络和支持向量机。Currently, with the rapid development of artificial intelligence, machine learning has been widely applied in various fields and used to simulate and analyze small data obtained from experiments. This study based on microalgae and using the collected 9 groups of microalgae research results in the literature for the data, using four kinds of algorithms, namely, the BP neural network, support vector machine (SVM), random forests (RF) and radial basis function (RBF) neural network, using MATLAB software for modeling and analysis, comparing with the original documents by response surface analysis method. The following main conclusions: by comparing the relative coefficient (R2), the mean absolute error (MAE), mean bias error (MBE), root mean square error (RMSE) and mean square error (MSE), compared with the traditional response surface analysis method, machine learning algorithms show better prediction effect, random forests and the relative coefficient of radial basis function (RBF) neural network is the most close to 1, the prediction effect is best, followed by BP neural network and support vector machine.

关 键 词：机器学习 微藻 MATLAB 环境因子 响应面法 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]