Skip to content

Multi-label classification of star types using kNN and deploying the model to a web-app using Streamlit

Notifications You must be signed in to change notification settings

shuchita28/NASA-Star-Type-Classification

Repository files navigation

🌠: NASA-Star-Type-Classification

Project

To predict the Type of star based on data such as Temperature of surface, luminosity of star, color and other physical attributes.

Data

https://www.kaggle.com/datasets/brsdincer/star-type-classification

Variable definition

Predictor variables

Temperature : Average temperature of the surface of star measured in degree K L : Realtive Luminosity of the star with respect to the sun measured in L/Lo units R : Relative radius of the star with respect to the sun measured in R/Ro units AM: Absolute Mass of the star Color : General Color of Spectrum Spectral_Class : O,B,A,F,G,K,M / SMASS - https://en.wikipedia.org/wiki/Asteroid_spectral_types

Target variable

Type : Red Dwarf, Brown Dwarf, White Dwarf, Main Sequence , Super Giants, Hyper Giants

One hot encoded from categorical to numeric (0 to 5) as follows :

  • Red Dwarf - 0
  • Brown Dwarf - 1
  • White Dwarf - 2
  • Main Sequence - 3
  • Super Giants - 4
  • Hyper Giants - 5

Math and formulae :

Lo = 3.828 x 10^26 Watts (Avg Luminosity of Sun)

Ro = 6.9551 x 10^8 m (Avg Radius of Sun)

Deployment

Using streamlit

About

Multi-label classification of star types using kNN and deploying the model to a web-app using Streamlit

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published