predicting exoplanet using artificial intelligence

ok

1. A WORLD AWAY:HUNTING OF EXOPLANET WITH AI TEAM NAME:STELLAR SPARKS TEAM DETAIL SR.NO NAME 1 PAWAN KUMAR 2 HARSHAL PRAJAPARI

2. Problem statement PROBLEM Space missions like Kepler, K2, and TESS have collected huge amounts of data about stars. Currently, scientists check this data manually to find planets outside our solar system (exoplanets). manual work is : Very slow Error-prone Limits discovery SOLUTION Train Models – Use Random Forest + XGBoost (machine learning models) to classify signals as: Exoplanet candidate Not a planet (false positive) Evaluate – Check accuracy, precision, and recall to ensure the model works well.

3. OBJECTIVES Train ML model on NASA’s open-source datasets (Kepler/K2/TESS). Identify Confirmed, Candidate, or False Positive exoplanets. Provide a web interface for researchers and novices to: Upload new data Run analysis .

4. ADVANTAGES Automated, fast, and accurate detection User-friendly for both scientists & students Continuous improvement via active learning Future Enhancements Deploy for citizen science / public use Model accuracy: ~90%+ (example) Precision and Recall per class Visual examples of correctly classified planets Show before vs. after preprocessing or classification . RESULT & ACCURACY

5. DATASET USED Sources: Kepler Mission Data (MAST) K2 Mission Light CurvesTESS Exoplanet Data features include: Orbital periodTransit duration Planetary radius Signal-to-noise ratio Stellar parameters (brightness, distance, temperature)

6. OVERVIEW

7. FLOW-DIAGRAM Researcher Select model Parameter tunning Show prediction or report User Enter data-type Prediction FOR RESEACHER: FOR USER :