A gauge for the measurement of radon and radon daughters concentration was tested in a radon chamber. Count rate distribution in time at the output of radiation detectors was measured and registered. The count rate distribution in time was then processed employing Principal Component Analysis (PCA) and the Root Mean Square Error (RMSE) of the count rate was investigated. It was found that PCA processing removes great part of count rate random fluctuations originating from radiation statistics, which is resulting in a decrease of the count rate random error and in random error of concentration. The RMSE of radon daughters concentration is about 3 times lower when "raw" results are PCA processed. Such decrease of error, without PCA signal processing, would require 9 times higher air flow through the air filter on which the radon daughters are deposited. In case of the measurements of the radon concentration the drop of the error is 2-3 times higher in case of long counting time.