Expanding the Horizons of Principal Component Analysis: Versatile Applications from Environmental Monitoring to Chemometrics

Principal Component Analysis (PCA) is a popular statistical method for large dataset analysis and has established itself in many fields, ranging from environmental modeling to spectroscopy. In this work, we bring to light its use in monitoring NO2 air pollution from satellite data during the lockdown periods of COVID-19, incorporating the strengths of Weighted PCA and Rescaled PCA to achieve improved predictability. Additionally, PCA has been applied in resonant ultrasound spectroscopy to optimize measurement points, especially in samples with complex geometries, demonstrating its effectiveness in reducing data points of collection while maintaining accuracy. Further, PCA's application coupled with classification methods like LDA and SVM has effectively determined the geographic origin of Indonesian coconuts, demonstrating its effectiveness in enhancing classification accuracy and chemometric analysis. The versatility of PCA is also evidenced in its use in clustering high-dimensional data through Adaptive Local PCA, which employs a neural network-based approach with adaptive learning rates to enhance clustering quality in dynamic data environments. These examples show the flexibility and utility of PCA in big data analysis across different fields, and a greater application of PCA in gaseous pollutant analysis and other complex data issues is needed.