A comparison of principal component analysis, reduced-rank regression, and partial least–squares in the identification of dietary patterns associated with cardiometabolic risk factors in Iranian overweight and obese women

Fatemeh Gholami,Ahmadreza Hajiheidari,Bahareh Barkhidarian,Neda Soveid,Mir Saeid Yekaninejad,Zahra Karimi,Niki Bahrampour,Seyed Ali Keshavarz,Gholamali Javdan,Khadijeh Mirzaei
DOI: https://doi.org/10.1186/s12874-024-02298-z
2024-09-28
BMC Medical Research Methodology
Abstract:According to epidemiological studies, unhealthy dietary patterns and lifestyle lead to rising obesity and cardiometabolic diseases in Iran. Hybrid techniques were used to identify a dietary pattern characterized by fiber, folic acid, and carotenoid intake due to their association with cardiometabolic risk factors such as anthropometric measurements, blood pressure, lipid profile, C-Reactive Protein (CRP), Plasminogen Activator Inhibitor (PAI), Homeostatic Model Assessment Index (HOMA Index), cardiometabolic index (CMI), and monocyte chemoattractant protein (MCP-1). So, the objective of the recent study is to compare the reduced-rank regression (RRR) and partial least–squares (PLS) approaches to principal component analysis (PCA) for estimating diet-cardiometabolic risk factor correlations in Iranian obese women.
health care sciences & services
What problem does this paper attempt to address?