Dresden 2017 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
DY: Fachverband Dynamik und Statistische Physik
DY 46: Modelling and Data Analysis
DY 46.3: Vortrag
Donnerstag, 23. März 2017, 15:30–15:45, HÜL 186
MDL based multidimensional step detection in piecewise constant signals — •Marius Bauer1,2, Gerald Hinze1, Klaus Müllen1,3, and Thomas Basché1,2 — 1Insitut für Physikalische Chemie, Universität Mainz, Duesbergweg 10-14, 55128 Mainz, Germany — 2Graduate School Materials Science in Mainz, Staudingerweg 9, 55128 Mainz, Germany — 3Max-Planck-Insitut für Polymerforschung, Ackermannweg 10, 55128 Mainz, Germany
Recent experimental advances allow tracking single molecule features not easily addressable by ensemble techniques. As a general drawback, low signal intensities compared to ensemble experiments often accompany such experiments, requiring more demanding data evaluation.
Here we present an approach to analyze time dependent single molecule fluorescence data employing a parameter free algorithm based on the minimum description length (MDL) concept. Following single molecule fluorescence, features such as rotational or energy transfer dynamics can be traced. Common models for these processes predict stepwise fluorescence intensity changes, where the identification of steps in the noisy data remains the challenging part of the data analysis.
We introduce a new method for the identification of steps and signal levels in noisy data for both recurring and nonrecurring signal levels. The method has been tested by Monte Carlo simulation to establish its reliability. As an example we have investigated the rotational dynamics of several organic dye molecules by polarization dependent single molecule measurements, yielding time scales and geometry of the reorientation process.