Exoplanet Detection Using Different Machine Learning Models

Overview

In this study, several different machine learning models were trained on exoplanet data gathered by NASA’s Kepler Telescope. The study aimed to find if any models were significantly more accurate, and which are the most accurate. First, the data was downloaded from the source and modified until it was ready for a computer to understand. This includes steps such as dropping unwanted columns and numerizing text data. Each model was trained and tested on the data. In addition to testing several different types of models, hyperparameters were also used to find the best model possible. The models were then tested and statistics like their average accuracy, loss, and standard deviation were calculated. This was used to determine if the models were statistically significantly accurate using a threshold of 95% accuracy. This study found that all but three of the machine learning models performed significantly accurate. In addition, all of the models were significantly accurate when their best hyperparameters were used. Overall, this study concluded that machine learning is a powerful tool for exoplanet detection, and can be of excellent use in the future. The best types of models and hyperparameters for each model were also found in this study.

Competition history

• AJAS 2026 Category not listed

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Physics and Astronomy projects
• AJAS Physics and Astronomy projects
• Every AJAS 2026 project