Project Proposal - Predicting Star Type¶

Motivation:¶

Problem¶

Patterns have emerged among traits of stars that can help determine at what stage of evolution the star is in.
link

However, determining at what stage a star is in its evolution requires a lot of complex calculations and must take into account a large variety of factors.

Solution¶

Because of the current scope of astronomical research, there is a lot of data containing star classification and specs of the stars themselves (e.g. luminosity, temperature, radius, etc). By using this data, we can create a program that can use known star features and predict what stage of stellar evolution the star is in.

Impact¶

If successful, this work may yield a classifier which predicts the stage of stellar evolution/star type of a star based on its physical properties. This would be helpful to astronomers and researchers to quickly be able to assign stellar classifications to newly discovered stars or stars that are going to soon enter a different evolutionary stage.

Further Reading¶

  • https://astro.unl.edu/naap/hr/hr_background3.html#:~:text=The%20main%20spectral%20classes%20in,M%E2%80%9D%20stars%20on%20the%20right.
  • https://astronomy.swin.edu.au/cosmos/h/hertzsprung-russell+diagram

Dataset¶

Detail¶

A Kaggle Star Dataset will be used to observe the following features for each star:

  • temperature
  • luminosity
  • radius
  • absolute magnitude
  • star type
  • star color
  • spectral class

The luminosity and radius of each star was calculated with respect to the average luminosity and radius of the sun.

  • Average luminosity of sun: Lo = 3.828 x 10^26 Watts
  • Average radius of sun: Ro = 6.9551 x 10^8 m

In the dataset there are 6 star types, represented by numbers 0-5.

  • Brown Dwarf -> Star Type = 0
  • Red Dwarf -> Star Type = 1
  • White Dwarf-> Star Type = 2
  • Main Sequence -> Star Type = 3
  • Supergiant -> Star Type = 4
  • Hypergiant -> Star Type = 5
Temperature (K) Luminosity(L/Lo) Radius(R/Ro) Absolute magnitude(Mv) Star type Star color Spectral Class
3068 0.0024 0.17 16.12 0 Red M
3042 0.0005 0.1542 16.6 0 Red M
2600 0.0003 0.102 18.7 0 Red M
2800 0.0002 0.16 16.65 0 Red M
1939 0.000138 0.103 20.06 0 Red M

Method:¶

This problem will utilize a classification model of machine learning. Temperature, luminosity, radius, and absolute magnitude data will be used to output the predicted star type (0-5).

Potential Problems¶

Based on my current knowledge of machine learning, I'm unsure if I'd be able to incorporate Star color or Spectral Class into the classifier since they are strings and not ints/floats like the other features. This could impact the accuracy of the predictions as we would be missing two features of the stars.