Detecting Exoplanets with Anomaly Detection and Hotelling's Theory
CSEF · 2019 Mathematical Sciences
Overview
Objectives The objective is to determine which anomaly detection method is the most accurate at classifying lightcurves of stars, and which is the most suitable for use by astrophysicists. Methods A computer with Python 2.7 and multiple machine learning modules installed, and three base open-source machine learning programs were used. These were modified and tested for accuracy on labelled lightcurve data. I used cleaned lightcurve data from the Exoplanet Hunting in Deep Space challenge on Kaggle. Results Modified three programs to utilize anomaly detection strategies to classify lightcurves, and were tested on a labelled dataset. Many tests were run with different input parameters to determine the best performance of each strategy. Hotelling's Theory proved to be the best, having the highest accuracy of 74.7% when used with a threshold of 0.3, a dip threshold of 175, and a cutoff of 1500. Conclusions Hotelling's Theory ran significantly faster than the other two programs and was the most accurate showing that it was a effective method for detecting exoplanets. It's speed shows that it is suitable for use on large amounts of data, and these two features allow a significantly higher amount of exoplanets to be discovered in a shorter timeframe than the traditional manual method.
Summary statement
I found that Hotelling's Theory is a suitable method for detecting exoplanets automatically as shown by high testing accuracies and fast runtime.
Help received
Andrew Lookingbill advised me on how to set up LSTMs and RNNs, along with Conner Vercellino who advised me to begin the project focusing on the dataset.
Competition history
- CSEF 2019
Resources
Related projects
ISEF · 2025
Innovative Machine Learning Approaches for Exoplanet Detection and Anomaly Identification: Applying Computational Techniques to Uncover Hidden Patterns in TESS Data
ISEF · 2018
Challenging Limitations: Using Deep Learning, Time Series Analysis, and Statistical Methods for Noise Reduction to Develop an Innovative Approach to Exoplanet Candidate Detection Using Earth-Based Telescopes
ISEF · 2025
Optimizing Exoplanet Detection: A Comparison of Neural Networks and Traditional Machine Learning Models on Light Curve Data
ISEF · 2023
A Novel Method in Improving the Accuracy of Exoplanet Classification Using Machine Learning
CSEF · 2023
Developing a New Criterion and Novel Equation to Increase the Accuracy in Exoplanet Searches
AJAS · 2026
Exoplanet Detection Using Different Machine Learning Models
ISEF · 2021
A Mathematically Driven Physical Analysis into Exoplanet Detection Confirmation Surveys Using Bayesian Inferential Statistics, Machine Learning/Linear Algebra Sklearn and TensorFlow Techniques, the BATMAN Python Programming Library, 3D Printing Techniques, and a Custom-made Python Processing Pipeline for Image Analysis and Lightcurve Detection
ISEF · 2024
Beyond the Star: A Data-Driven Approach to Exoplanet Classification
Closest projects by meaning, across every fair and year in the corpus.
Browse more like this
Source: California Science & Engineering Fair public projects