\contentslabel
[\thecontentslabel]1.25cm \titlerule*[.5pc]. \thecontentspage\contentslabel[\thecontentslabel]1.25cm \thecontentspage[]\contentslabel[\thecontentslabel]1.25cm \titlerule*[.5pc]. \thecontentspage[]Thereisworktobedoneforthiscases,understandwhatisgoingonandinterpretcorrectlytheresults,butlastwegotsome.

CSOM

Well,asImentionedbeforeIdidsometestsusingtheESOMmethodbutsinceIwasntgettinganyresultsIthougthoftestingthismethond,asalwaysIstronglyrecommendtoreadcarefullyitsmanual,http://dame.dsf.unina.it/documents/SOFM_UserManual_DAME-MAN-NA-0014-Rel1.1.pdfandfullyundesrtandwhatisgoingonbehindthecurtains.Inthemeanwhile,thisismyownexplanation.ThismethodusesFITSfiles,doesnotsupportdatacubes,specifficallyusesaneighborhoodfunctioninordertopreservethetopologicalpropertiesoftheinputspace,itisatypeofartificialnetworkandismainlyunsupervisedlearningandproducesalowdimensionaldiscretizedrepresentationoftheinputspaceofthetrainingsamples.Iinthiscaseyoucanchoosethenumberofclusters/neuronsinthefirstlayer(neuralnetwork),thediameter,numberoflayers(intheneuralnetwork),learningrateandvarianceoneachlayer.Hereyouhavemoreinputparameterstocontrol.

ExpectedResults

Wellinthiscase,sinceonlyFITSimagesareallowed,whatweexpecttofindareareasindeitfyingthedifferentobjectsintheinterstellarmedium.Theimportantresultsinthiscase,aregotintheRunandTeststeps,intheTrainsteponlythenetworkconfigurationisoutputted.Whatweareinterestedonseeinaretheplottedclusters.

Tests

InthiscaseIdidsometestsontheCSOMworkspace,butnoneofthe,wheresuccesful,toomanyinputvariablestocontrolandtest.So,inthiscaseIwillleavethisparametersfreeforyoutotry.Idobelievethattismethodcouldbeveryusefulandifyoufindawaytoinputthedatacubeinadifferentconfigurationyouwillgetsomeinterestingresults,duetothefactthatinthismethodthepreservationofthetopologyisoneofthemainprinciples.