This project is a cooperation between Mathematical Statistics and the department of Biology at Lund University, and the research will apply classification algorithms on a new exiting application area. A long-term study of the a Great reed warbler population in Sweden is ongoing where among other parameters, the song is recorded. The properties in the song are still an unexplored field, where the aim for future studies is to understand the role of the song in an ecological and evolutionary context. This work is also motivated by recent developments within time-frequency analysis of non-stationary processes and the increased use of time-frequency analysis in many application areas. We focus on using multitapers, which are low-rank approximations of time-frequency kernels. The low-rank multitaper spectrogram estimators are much more computationally efficient compared to many other time-frequency techniques. The multitapers can also be designed to fulfill properties of data from a specific application. The project will focus on the construction of a reliable program tool for analysis and classification of syllables of the song.
Students with basic eligibility for third-cycle studies are those who have completed a second-cycle degree- have completed courses of at least 240 credits, of which at least 60 credits are from second-cycle courses, or- have acquired largely equivalent knowledge in some other way, in Sweden or abroad.Special eligibility In addition to the basic qualifications for postgraduate studies, 60 ECST credits are required in mathematical subjects, of which at least 30 ECTS credits in mathematical statistics. A degree project comprising 30 ECTS credits is also required.