R Scripts for Full-Spectrum Analysis with Machine Learning Framework for Quantitative Assessment of SARS-CoV-2 Lateral Flow Immunoassays

This dataset contains R scripts for spectral data processing and machine learning model development used in the study "Integrated Full-Spectrum Analysis with Machine Learning Framework for Quantitative Assessment of Lateral Flow Immunoassays: Application to SARS-CoV-2 Detection."

The scripts enable the processing of visible-range spectral data (400-700 nm) from lateral flow immunoassays and the development of machine learning models for quantitative analysis. The repository includes:

1. Spectral preprocessing scripts for noise reduction (Savitzky-Golay filtering), standard normal variate (SNV) transformation, and principal component analysis.

2. Machine learning model development scripts for training and evaluating five algorithms: polynomial regression, partial least squares regression, support vector regression, random forest, and gradient boosting.

3. Dual-channel normalization methods implementing both differential (T-C) and ratio (T/C) spectral analyses.

These tools support the transformation of qualitative lateral flow tests into quantitative analytical tools by extracting meaningful patterns from spectral data of gold nanoparticle immunocomplexes. While the raw clinical data cannot be shared due to patient confidentiality restrictions, these scripts facilitate the reproducibility of the analytical methodology.

这个数据集包含用于研究《基于机器学习框架的全光谱整合分析在侧向流免疫测定定量评估中的应用：以SARS-CoV-2检测为例》的R脚本（R scripts），涵盖光谱数据处理与机器学习模型开发。

这些脚本可实现对侧向流免疫测定中可见光范围（400-700 nm）光谱数据的处理，以及用于定量分析的机器学习模型开发。该仓库包含以下内容：

1. 光谱预处理脚本，用于降噪（Savitzky-Golay滤波）、标准正态变量（SNV）变换及主成分分析；
2. 机器学习模型开发脚本，用于训练和评估五种算法：多项式回归、偏最小二乘回归、支持向量回归、随机森林及梯度提升；
3. 双通道归一化方法，可实现差分（T-C）与比率（T/C）光谱分析。

这些工具通过从金纳米颗粒免疫复合物的光谱数据中提取有意义的模式，支持将定性侧向流检测转化为定量分析工具。由于患者隐私保护限制，原始临床数据无法共享，但这些脚本可促进分析方法的可重复性。