Low-cost sensors (LCS) have emerged as promising tools for air quality measurement and are increasingly applied to indoor and outdoor monitoring situations. Calibration step ensures the reliability and accuracy of LCS measurements; however, current calibration methods face several challenges in addressing the complexities of real-world environmental conditions. This study proposes a unique and comprehensive experimental setup to investigate the suitability of different outdoor calibration models for multivariate regression tasks. By incorporating multiple pollutant variables and meteorological data, the calibration is enhanced in the presence of pollutant interference. An outdoor evaluation assesses the performance and effectiveness of various machine learning calibration models in calibrating the sensors. The resulting calibration pipeline applies to many sensor calibration situations.