Authors :
Vansh Teotia; Praveena Melepatte; Poojan Upadhyay; Saksham Chandna; Dr Aditi Panda
Volume/Issue :
Volume 8 - 2023, Issue 2 - February
Google Scholar :
https://bit.ly/3IIfn9N
Scribd :
https://bit.ly/3Kl31ss
DOI :
https://doi.org/10.5281/zenodo.7655926
Abstract :
The objective of this study was to use and
compare multi- ple classifying models that can be used
for classifying as- tronomical data and was tested upon
data obtained from Sloan Digital Sky Survey: Data
Release-16. Various Classi- fying models have been
trained and tested by dividing the data into two parts- 80
per cent of the data was then used for training purposes
and 20 per cent for testing. In order to achieve the task of
classifying the tabular data consist- ing of spectroscopic
and photometric parameters effectively, the study was not
just limited to usage of indiviual models. Stacking : the
combination of multiple Classifying mod- els has also
been implemented. Multiple stacking models were
created for the same .Stacking models have on mul- tiple
occasions proven to have higher evaluation metrics , thus
having significantly better performance than any individual classifier, proving that stacking is a better choice
to classify data. certain Individual models such as Bagging , Hard Voting etc have been found to have comparable
performance to that of Stacked Models. Box plots for indiviual classes were also plotted to compare and determine
the models that were capable in identifying a single class of
stellar objects. The models from this study could be used
as a reliable classification tool for a wide variety of astronomical purposes to accelerate the expansion of the sample
sizes of stars, galaxies, and quasars.
Keywords :
Classification, Classifiers, Stars, Galaxies, Quasars.
The objective of this study was to use and
compare multi- ple classifying models that can be used
for classifying as- tronomical data and was tested upon
data obtained from Sloan Digital Sky Survey: Data
Release-16. Various Classi- fying models have been
trained and tested by dividing the data into two parts- 80
per cent of the data was then used for training purposes
and 20 per cent for testing. In order to achieve the task of
classifying the tabular data consist- ing of spectroscopic
and photometric parameters effectively, the study was not
just limited to usage of indiviual models. Stacking : the
combination of multiple Classifying mod- els has also
been implemented. Multiple stacking models were
created for the same .Stacking models have on mul- tiple
occasions proven to have higher evaluation metrics , thus
having significantly better performance than any individual classifier, proving that stacking is a better choice
to classify data. certain Individual models such as Bagging , Hard Voting etc have been found to have comparable
performance to that of Stacked Models. Box plots for indiviual classes were also plotted to compare and determine
the models that were capable in identifying a single class of
stellar objects. The models from this study could be used
as a reliable classification tool for a wide variety of astronomical purposes to accelerate the expansion of the sample
sizes of stars, galaxies, and quasars.
Keywords :
Classification, Classifiers, Stars, Galaxies, Quasars.