Resources

Final database (Work package 4).

The files include:

“flux_density_data_GRATOS_final”: our final GRATOS data sample of 139 galaxies (102 ULIRGs of Farrah et al. 2009 + 37 PG quasars of Symeonidis et al. 2016)

“GRATOS_results_images_and_descriptions”: Result images, with their descriptions in text files:

1. Similarity graph
2. Leiden algorithm graph
3. Shi & Malik clustering algorithm graph (5 clusters)
4. Mean Signals
5. Fork diagram of 11.2 μm PAH emission & Si absorption (Shi & Malik algorithm)

Data description (txt files):

The data are presented in .txt files, extracted from the CASSIS website (Combined Atlas of Sources with Spitzer IRS Spectra) available here. Each text file contains 3 columns with the values of:

1) Observed wavelength (in microns – μm) of each data point,

2) Observed flux density (in Jansky – Jy) at the corresponding observed wavelength,

3) Flux density error values (in Jansky – Jy) of the corresponding flux density value.

The data were extracted by downloading the Low-Resolution spectra (in fits format) of the query no. 105: Spectroscopic Study of Distant ULIRGs II by Houck et al, 2004.

Data Analysis Methodology:

These values were redshift-adjusted, in order to retrieve the rest wavelength flux and error values, and then resampled with interpolated flux values at specific rest wavelength bins, which were in turn used for the extraction of the physical features studied in this work (such as the PAH emission & Silicate absorption features).

The resampled fluxes were then used to compare pairs of spectra in order to generate the similarity graph, on which we implement several Clustering Algorithms and performed our Graph Theoretical Analysis.

Finally, we study the correlation between the observed physical features and identified clusters on our graphs (some of which are presented in the “Results” page). This correlation provides the key to understanding the temporal evolution of the power sources of ULIRGs, throughout different stages of the merger event.