Intelligent Classifier of Patterns for celestial bodies using a Two-dimensional approach

Jimenez Cruz Raul

Abstract

Human curiosity towards astronomy in recent decades has allowed the development of great technological advances, which has helped to deepen the knowledge of celestial bodies. Unfortunately, there are still certain inconsistencies in the terminology and classification, therefore this paper proposes the possibility of an intelligent classification based on the features of celestial bodies instead of calculating their electromagnetic fields. Physical criteria are proposed obtained from the data bank of the Digital Sky Survey which was modified in the Kaggle data repository for discrimination between different body classes. The classification obtained has some advantages, especially in the reduction of time and lower computational cost on KNN (K Neighbors Classifier), SVM, Naive Bayes Classifier, Decision Tree Classifier and Random Forest Classifier.