Automated Classification of Variables and Milky Way Evolution

My project is the analysis of the VISTA Science Archive, in particular the results of the VVV survey, to identify the variable stars present in the sample. The resulting detection will involve a probability of the detected object belonging to specific categories of variable star. The method for detection and analysis will involve variability indices created by (Lopes et al.) and Spectral Energy Distribution (SED) fitting as part of an automated data analysis pipeline (potentially involving machine learning aspects to improve the quality of the identification process) to both detect variable behaviour and assign probabilities for each possible type of variable star that could fit the detected object.
The first step of the project is researching all possible types of variable stars and identify photometric or spectroscopic characteristics that could be used to distinguish between the different types.
The second step of the project will involve developing the analysis methods. This includes filtering the VVV database to remove any issues in the data and preparing it for analysis in the pipeline. Variability indices are then used to pick out the varying stars from the observed objects.
The third step uses SED fitting, photometric analysis (detecting the amplitude of light curve variations) and other methods such as colour and metallicity to pin down the stars position on the HR diagram.
More detail will be added later as the project proceeds.