Correction of meteorological effects on muon component of secondary cosmic rays significantly extends the usability of muon monitors. We propose a new data driven empirical method for correction of meteorological effects on muon component of secondary cosmic rays, based on multivariate analysis. Several multivariate algorithms implemented in Toolkit for Multivariate Data Analysis with ROOT framework are trained and then applied to correct muon count rate for barometric and temperature effects. The effect of corrections on periodic and aperiodic cosmic ray variations is analyzed and compared with integral correction method, as well as with neutron monitor data. The best results are achieved by the application of linear discriminant method, which increases sensitivity of our muon detector to cosmic ray variations beyond other commonly used methods. Plain Language Summary Primary cosmic rays are energetic particles that arrive at Earth from space. On their journey toward Earth they are affected by the solar wind (a stream of charged particles emanating from the sun), which has information about various solar processes embedded in it. In top layers of the atmosphere primary cosmic rays interact with nuclei of air molecules and produce large number of secondary particles that propagate toward Earth's surface. These secondary particles preserve information about variations of primary cosmic rays, which allows for the study of solar processes using Earth based detectors. One type of secondary particles that can be detected on the ground are muons. However, muons are affected by the conditions in the atmosphere, which can disturb the information about variations of primary cosmic rays. That is why it is important to model these atmospheric effects on cosmic ray muons as well as possible so they can be corrected for. In this study, we present a new method for modeling and correction of atmospheric effects on cosmic ray muons, that is based on multivariate analysis utilizing machine learning algorithms. This method increases sensitivity of our muon detector to cosmic ray variations beyond other commonly used methods. Key Points Correction of meteorological effects on muon component of secondary cosmic rays significantly extends the usability of muon monitors A new method for modeling of meteorological effects utilizing multivariate analysis and machine learning techniques is presented Correction efficiency of the best performing algorithm is greater than for other commonly used methods