SIMBAD references

Over the past few decades, a remarkable surge in exoplanet discoveries has led to a substantial increase in available data. As of this writing, over 5000 exoplanets have been confirmed, with thousands more expected from current and future observatories. This wealth of data calls for a need to efficiently and thoroughly process complex exoplanetary data sets in order to identify significant trends and gain insights into the characteristics of exoplanets. In this paper, we demonstrate the advantage of leveraging machine learning to understand exoplanet occurrence and formation based on patterns observed within astronomical data. We use a technique called association analysis to identify frequent co-occurring planet subtypes in thousands of exoplanetary systems and generate association rules. We then explore the association rules derived from our algorithm to discover and understand the significant patterns found in the data. For instance, one of our rules shows that a planetary system with a stellar spectral type of “M” and containing a temperate gas giant (≥200 and <400 K) tends to also include a cold gas giant (<200 K). Our findings include rules that predict the presence of cold and long-period gas giants in systems with G and M dwarf stars, despite observational biases against detecting such planets. Additionally, we uncovered patterns suggesting the potential presence of temperate gaseous planets, which are of interest to astrobiology. Overall, this project establishes a valuable and extendable framework for future research, with its predictive power expected to grow along with the continually increasing abundance of exoplanet data.