| 摘要 |
<p>The process comprises passing (102) a sample (E) through a processing line comprising microelectromechanical or nanoelectromechanical system type mass spectrometer electromechanical sensor for counting the number of successive detections of the adsorption of the component to obtain a signal representative of the molecular mass parameter, and estimating (104, 106, 108, 200) the molecular mass parameter using a signal processing device. The molecular mass parameter is defined on the basis of a parameter of temporal distribution of successive detections of the adsorption of the component. The process comprises passing (102) a sample (E) through a processing line comprising a microelectromechanical or nanoelectromechanical system type mass spectrometer electromechanical sensor for counting the number of successive detections of the adsorption of the component to obtain a signal representative of the molecular mass parameter, and estimating (104, 106, 108, 200) the molecular mass parameter using a signal processing device. The molecular mass parameter is defined on the basis of a parameter of temporal distribution of successive detections of the adsorption of the component by the electromechanical sensor. The estimating step is performed by Bayesian inference on the basis of direct analytical model of the signal according to the function of molecular mass parameter and a technical parameter of the processing line. The direct analytical model comprises a parameter representing the detection moments at which the sensor detects the adsorption of the component. The molecular mass parameter is defined from the parameter representing the detection moments. The parameter representing the detection moments is taken form a vector or a list of parameters for each detection of the component. The method further comprises detecting sites of adsorption of the components on the electromechanical sensor in which analytical model further comprises other parameter of temporal distribution of the adsorption sites and a deterministic function providing a frequency value. The molecular mass parameter and the processing line according to which the direct analytical model of the signal is defined have a probability dependency relationship. The estimating step, by approximating the probability to joint posterior of a parameter related to the molecular mass parameter and the technical parameter of the processing line conditionally to the signal obtained using a stochastic sampling algorithm, comprises providing the sampled values of the parameters using a sample loop of the parameters, and estimating the molecular mass parameter calculated from the sampled values. An independent claim is included for a device for estimating a molecular mass parameter in a sample.</p> |
