Detection Biases and Methodological Impacts in Exoplanet Discovery

A lot more information about planets outside our solar system is available now due to exoplanet discoveries. However, the biases introduced by detection methods like Transit and Radial Velocity affect the observed characteristics of detected exoplanets. This work examined how various exoplanet discovery methods shape the observed characteristics of the detected exoplanet population. Advanced statistical analysis and machine learning techniques were used to preprocess the obtained exoplanet dataset and to assess biases in the detection methods. This study represents a unified statistical and machine learning-based framework to quantify and visualize detection biases across discovery methods based upon a comprehensive data set, as compared to previous works that focused on certain detection techniques or a single planetary characteristic. The dataset contained 23 planetary, stellar, and observational parameters obtained from the NASA Exoplanet Archive. The findings showed that the Transit method is biased towards the detection of short-period, low-eccentricity planets, whereas Radial Velocity is suitable for detecting massive planets around larger stars. High-mass and long-period planets are detected by imaging methods. Microlensing is the least biased but underused. Box, Violin and KDE plots show how the biases in detection methods lead to an incomplete picture of exoplanets. Improvement in detection methods is necessary to better understand exoplanets.