Decoding the Origins of the Universe with Statistics

Introduction

Enter the age of the 1950s, where astronomers were reaching for the stars with a new tool: the radio telescope. With these powerful instruments, they probed the heavens, searching for answers to the mysteries of the universe. But as they paired the signals received with visible examinations of the heavens, they encountered perplexing anomalies. Some of the smaller point-source objects didn't have a visual match. Dubbed "quasi-stellar radio sources" or "quasars," these enigmatic signals seemed to come from one place, like a star, yet nothing was seen. It took years of study to realize that these distant specks, which seemed to indicate stars, were in fact created by particles accelerated at velocities approaching the speed of light, leaving scientists awestruck by their sheer power and beauty. They were pointlike like stars, and not fuzzy like nebulae or galaxies. Here’s one the Hubble telescope saw:

The universe is a wondrous and enigmatic place where stars, galaxies, and quasars coexist in a breathtaking display of cosmic beauty. At the heart of the quasar lies a supermassive black hole, an insatiable monster that devours everything in its path. As it gobbles up matter such as dust, gas, and even entire stars, it unleashes enormous amounts of energy, resulting in luminosities that surpass entire galaxies. The sheer magnitude of this spectacle is beyond comprehension, with some quasars reaching luminosities thousands of times greater than that of the Milky Way.

However, amidst this awe-inspiring splendor, it can be hard to distinguish between stars, galaxies, and quasars, each one a captivating enigma that invites us to explore the mysteries of the universe, as quasars provides some clues as to the end of the Big Bang's reionization.

How to study these objects in a quantitative way?

Modern astronomy is concerned with the study and characterization of distant objects such as stars, galazies, or quasars. Objects can often be very quickly characterized through measurements of their optical spectrum. A spectrum is a measure of the photon flux as a function of wavelength.

The above spectrum is that of the star Vega, the brightest star in the northern constellation Lyra. Its surface is at about 9600 degrees Kelvin, and its spectrum is roughly that of a 9600K black-body, with absorption due to molecules in its cooler atmosphere. The deepest of these absorption spikes are due to the energy levels of Hydrogen. From examination of high-resolution spectra like this one, one can learn a lot about the physical processes at work in a distant astronomical source.

Here are a few more examples (low resolution) of stars and quasars from awesome SDSS website.

The Statistical Problem

The statistical problem is simple. Given two spectrum, how to predict which is a star, galaxy, or quasar? We need quite a lot of data of such spectrum, and preprocess the data into interpretable quantities, and use them as features to classify the objects. This is a classification problem.

Are you excited enough, just like me?

Data has power to probe beyond your own existence.

Imagine how minuscule and grandeur is your existence.

Do you even realize it?

References

I have learnt these from online resources. All these information are taken from multiple sources, and I am grateful to them to share such information in an organized manner. I have connected different sources of information from a statistical point of view. I have shared that in the comments, since LinkedIn algorithm doesn't work well with link in the article.