Search

International audience; Context. Serendipitous X-ray surveys have proven to be an efficient way to find rare objects, for example tidal disruption events, changing-look active galactic nuclei (AGN), binary quasars, ultraluminous X-ray sources (ULXs), and intermediate mass black holes. With the advent of very large X-ray surveys, an automated classification of X-ray sources becomes increasingly valuable.Aims. This work proposes a revisited naive Bayes classification of the X-ray sources in the Swift-XRT and XMM-Newton catalogs into four classes − AGN, stars, X-ray binaries (XRBs) and cataclysmic variables (CVs) − based on their spatial, spectral and timing properties and their multiwavelength counterparts. An outlier measure is used to identify objects of other natures. The classifier is optimized to maximize the classification performance of a chosen class (here XRBs) and it is adapted to data mining purposes.Methods. We augmented the X-ray catalogs with multiwavelength data, source class, and variability properties. We then built a reference sample of about 25000 X-ray sources of known nature. From this sample the distribution of each property was carefully estimated and taken as reference to assign probabilities of belonging to each class. The classification was then performed on the whole catalog, combining the information from each property.Results. Using the algorithm on the Swift reference sample we retrieved 99%, 98%, 92% and 34% of AGN, stars, XRBs, and CVs, respectively, and the false positive rates are 3%, 1%, 9% and 15%. Similar results are obtained on XMM sources. When applied to a carefully selected test sample, representing 55% of the X-ray catalog, the classification gives consistent results in terms of distributions of source properties. A substantial fraction of sources not belonging to any class is efficiently retrieved using the outlier measure, as well as AGN and stars with properties deviating from the bulk of their class. Our algorithm is then compared to a random forest method; the two showed similar performances but the algorithm presented in this paper improved insight into the grounds of each classification.Conclusions. This robust classification method can be tailored to include additional or different source classes and can be applied to other X-ray catalogs. The transparency of the classification compared to other methods makes it a useful tool in the search for homogeneous populations or rare source types, including multi-messenger events. Such a tool will be increasingly valuable with the development of surveys of unprecedented size, such as LSST, SKA and Athena, and the search for counterparts of multi-messenger events.

Studying compact objects (CO) such as white dwarfs (WDs), neutron stars (NSs) and stellar-mass black holes (BH) with masses typically ranging from 3−20 M is important for understanding the endpoints of stellar evolution and accretion/ejection processes, ubiquitous phenomena in the Universe. It provides further insights into the gas state and distribution in the early Universe. Studying the cosmic history of supermassive BH (with masses ranging from 106−10 M) growth could shed light on the formation and evolution of galaxies. It tells us more about their role in the reionization of the Universe and their impact on their (interstellar and/or intergalactic) environment. COs are also excellent laboratories to study matter in extreme conditions, such as matter at super-nuclear density in the NS interior or matter under the effect of strong gravity and magnetic fields in the vicinity of BHs and NSs, respectively. Due to accretion of matter that goes on in their vicinity, most COs were first detected in X-rays. For these reasons, all-sky X-ray surveys are most suitable for finding and studying such sources. The Swift observatory, inoperation since 2004, carries an X-ray telescope (XRT) and two other co-aligned instruments, the Burst Alert Telescope (BAT) and the UV/Optical Telescope (UVOT). This enables a multi-wavelength study of CO behaviour from optical to hard X-rays. Swift is dedicated to the study of gammaray bursts that appear randomly on the sky. This implies that Swift pointings have been performed all over the sky,covering1905 square degrees with the XRT, with many fields observed several times within a day,over a period from many days to weeks/months (or even years),allowing to probe variability on various time scales. In this work, I investigated the nature of the X-ray sources contained in the Swift-XRT catalog in order to isolate interesting objects possibly harbouring an accreting compact object (e.g. Xray binaries, active galaxy nuclei – AGN, ultra-luminous X-rays sources, tidal disruption events). To do so, I cross-correlated a subsample of the Swift X-ray Telescope Point Source (1SXPS) catalog, containing 98,762 sources with detections of best quality, with 16 external multi-wavelength catalogs which provide source type identification (active galactic nuclei, stars, X-ray binaries etc.) using the Topcat software. This enabled me to build a golden sample of known objects, divided into three main classes (AGNs, COs and stars) that represent the main types of objects observed in the X-ray sky. Within this subsample, I found that it consists of 4929 AGNs, 1125 stars and 231 COs. I studied their temporal, spectral and spatial properties in order to define selection criteria which would enable me to classify the rest of the 1SXPS sources. I found that COs are the most variable group, followed by AGNs and stars. Using spectral indicators from the 1SXPS catalog (power-law photon index ΓPL and hardness ratios), I found that the ΓPL-distributions of COs, AGNs and stars differ. For AGNs, the ΓPL-distribution is clustering around 1.71,which is consistent with typical values derived for these objects. For COs, the distribution of ΓPL-values is more widely spread, likely corresponding to the different spectral states that can be seen in X-ray binaries. Stars also display scattered ΓPL-distribution, but with generally high values, implying that the PL model is not physical and that thermal emission is more likely. I found that most stars can be isolated from the rest of AGNs and COs thanks to their low X-ray to optical or IR flux ratio. After establishing the selection criteria, I applied them to the golden sample of identified sources, in order to investigate the reliability of the defined selection scheme. 73% of the stars, 35% of the COs, and 95% of the AGNs were retrieved, with false classification probability of 8%, 78% and 15%, respectively. The low number of retrieved COs and their high probability of false identification is due both to the large fraction of AGN with respect to COs in the golden sample and the fact COs and AGNs share similar properties (ΓPL-distribution and X-ray to optical/IR distribution intervals coincide). Finally, the selection criteria were applied back to the rest of the 1SXPS source high quality sample, resulting in 78,918 AGN candidates (86%), 9294 star candidates (10%) and 3752 CO candidates (4%). I discuss the caveats of the method used and I propose possible improvements, in particular for helping decreasing the high probability of false identification rate for COs. I also discuss further interesting works that could be done from the obtained results.

Miskoncepcije vezane za termodinamiku počinju s našim osjetilima. Naše potrebe kao ljudskog bića uzimamo kao referentnu točku u promatranju svijeta oko sebe, što više nije poželjno za studente fizike. Na taj način „toplije“ i „hladnije“ postaje „toplo“ i „hladno“. Analizom rezultata inicijalnog testa iz termodinamike i kinetičke teorije materije upoznali smo miskoncepcije kod studenata i teško savladiva područja gradiva. S tom informacijom, predavač može započeti promjenu u svom pristupu te osmisliti učinkovitija predavanja. Ova je generacija studenata postigla samo 24% mogućeg napretka, što nije zadovoljavajući rezultat. Njihov opći interes da fizikom objasne svijet oko sebe može se ocijeniti srednjom riješenošću pitanja iz grupe za moć opažanja, koja iznosi 60% za tu grupu na izlaznom ispitu. No korelirana pitanja imaju ispunjeno 40% od mogućeg napretka, pa možemo reći da su djelomično usvojili konstruktivnije i kritičnije razmišljanje, gledano kao skupina. Studentima je povećano shvaćanje ovisnosti između tlaka, volumena i agregatnog stanja sustava. U radu se često pojavljuje problem razumijevanja pojma unutarnje energije i faznih prijelaza općenito, no i razlikovanja svojstava homogenosti i izotropnosti, pa čak i Newtonovih zakona. Također studenti imaju problema s definicijom temperature, no najčešće spomenut problem nije vezan uz niti jedno područje teorije. To je tendencija studenta da idealizira realne situacije, i to ne u svrhu pojednostavljivanja problema. Velika većina vjeruje da stvarnost jest takva. Uz utjecaje drugih kolegija, miskoncepcije i načine razmišljanja koje studenti nose u sebi još od osnovne škole i nepotpuno opće znanje, teško je oformiti savršen pristup. Profesor jedino može pokušavati stvoriti najbolji mogući, i promatrati postignute rezultate. Naravno, nakon što ih dobije, ne smije zaboraviti uračunati omjer studenta i studentica kako bi realno ocijenio svoj uspjeh, pošto su razlike napretka između njih jako velike. Uz sve preporuke predavaču, koje se mogu eksplicitno i implicitno iščitati iz ovog rada, preporučuje se i veći broj koreliranih pitanja i pitanja koja se odnose na moć opažanja. Zaključci doneseni statističkom obradom trenutnog broja pitanja iz ovih skupina ne moraju biti vjerodostojni zbog malog broja uzoraka.