机器学习支持的NIR光谱学. 第三部分:基于设计的超参数 (HyD) 的ANN-MLP优化,模型通用性和模型可转移性
Hussain Ali1, Prakash Muthudoss2,3,4, Chirag Chauhan5
1Christ (Deemed to Be University), Bangalore, 560029, Karnataka, India.
AAPS PharmSciTech
|December 7, 2023
概括
这项研究使用实验设计 (DoE) 和统计过程控制 (SPC) 优化了人工神经网络多层感知器 (ANN-MLP) 的超参数. 该方法有效地管理了模型漂移,并通过外部数据确保了可靠的预测.
科学领域:
- 数据科学数据科学数据科学
- 机器学习 机器学习
- 制药科学 制药科学
背景情况:
- 数据驱动模型面临由于变化的故障,导致模型漂移和不准确的预测.
- 监测和减轻模型漂移对于保持数据驱动应用程序的预测准确性至关重要.
- 像USFDA和ICH这样的监管机构强调基于风险的方法来管理制药变异.
研究的目的:
- 为了研究人工神经网络多层感知器 (ANN-MLP) 模型的超参数优化.
- 实施结合实验设计 (DOE),漂移分析和统计过程控制 (SPC) 的强大方法.
- 通过管理数据变化和模型生命周期,确保模型可靠性和准确预测.
主要方法:
- 使用实验设计 (DoE) 进行内部验证数据的预先选和优化,以定义设计和控制空间.
- 利用回归性能指标来选择最佳的超参数,优化建模时间和存储.
- 对外部验证数据进行了目标漂移分析,并使用SPC对平均绝对误差进行了趋势分析.
主要成果:
- 通过DoE识别了最佳的超参数,平衡性能与计算效率.
- 观察到外部数据的漂移,但证实它仍在内部数据的验证范围内.
- 建立了规范外和过程控制限制,证明了模型的可靠性和性能.
结论:
- DoE,漂移分析和SPC的综合方法使得强大的超参数优化和有效的模型生命周期管理成为可能.
- 该方法通过对内部和外部数据验证模型性能,确保准确和可靠的预测.
- 该研究为在现实应用中保持模型完整性和可靠性提供了宝贵的见解,与监管预期保持一致.
相关概念视频
01:21Mass Spectrometry: Complex Analysis
792
Mass spectrometry is an important technique for the identification of pure compounds. However, it has some limitations for the analysis of complex mixtures, often due to excessive fragmentation making the spectrum too complicated to decipher. Mass spectrometry can be combined with suitable separation methods in sequence, forming hyphenated methods, which are useful in the analysis of complex mixtures.
GC–MS is a powerful hyphenated method commonly used in forensics and environmental...
GC–MS is a powerful hyphenated method commonly used in forensics and environmental...
792
01:29Mechanistic Models: Compartment Models in Algorithms for Numerical Problem Solving
56
Mechanistic models play a crucial role in algorithms for numerical problem-solving, particularly in nonlinear mixed effects modeling (NMEM). These models aim to minimize specific objective functions by evaluating various parameter estimates, leading to the development of systematic algorithms. In some cases, linearization techniques approximate the model using linear equations.
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...
56
01:13Mass Analyzers: Overview
683
The mass analyzer is a crucial component of the mass spectrometer. In the ionization chamber, the vaporized sample is bombarded with a high-energy electron beam to generate a radical cation and further fragment into neutral molecules, radicals, and cations. A series of negatively charged accelerator plates accelerate the cations into the mass analyzer. The mass analyzer separates ions according to their mass-to-charge (m/z) ratios and then directs them to the detector. The common types of mass...
683
01:182D NMR: Overview of Heteronuclear Correlation Techniques
184
Heteronuclear correlation spectroscopy is an analytical technique that investigates the coupling between different types of nuclei, often a proton and an X-nucleus, such as carbon-13 or nitrogen-15. This method is commonly used in nuclear magnetic resonance (NMR) spectroscopy to gain insights into complex chemical compounds' structural and compositional aspects. A typical heteronuclear correlation spectrum displays X-nucleus chemical shifts on one axis and a proton spectrum on the other...
184
01:02¹H NMR: Interpreting Distorted and Overlapping Signals
1.0K
Spin systems where the difference in chemical shifts of the coupled nuclei is greater than ten times J are called first-order spin systems. These nuclei are weakly coupled, and their chemical shifts and coupling constant can generally be estimated from the well-separated signals in the spectrum.
As Δν decreases and the signals move closer, the doublets appear increasingly distorted. The intensities of the inner lines increase at the cost of those of the outer lines as the signals are...
As Δν decreases and the signals move closer, the doublets appear increasingly distorted. The intensities of the inner lines increase at the cost of those of the outer lines as the signals are...
1.0K