MODELING QUASAR SUPERMASSIVE BLACK HOLE MASS THROUGH LONG SHORT-TERM MEMORY ALGORITHMS

Abstract

Quasars are among the most luminous and energetic objects in the universe, powered by accretion processes occurring around supermassive black holes (SMBHs) located at the centers of active galactic nuclei (AGN). Accurate estimation of SMBH mass is fundamental for understanding galaxy evolution, quasar formation, accretion physics, and the coevolution of black holes with their host galaxies. Traditional black hole mass estimation techniques, including reverberation mapping and virial scaling relations, provide valuable measurements but often require extensive observational resources and long-term monitoring campaigns. The rapid growth of astronomical datasets has motivated the adoption of machine learning and deep learning approaches capable of extracting complex nonlinear relationships from large-scale observations. Recent studies have demonstrated that Long Short-Term Memory (LSTM) neural networks can effectively model temporal and sequential astrophysical data for predicting quasar black hole masses. This study investigates the application of LSTM algorithms for modeling the masses of quasar supermassive black holes. The proposed framework utilizes observational quasar datasets containing redshift measurements, luminosity parameters, spectral properties, and variability information. Data preprocessing techniques, including normalization, feature engineering, and sequence generation, are employed to prepare the datasets for deep learning analysis. The LSTM architecture is selected because of its ability to capture long-term dependencies and temporal patterns present in astronomical observations. By learning complex relationships between observational parameters and black hole masses, the model seeks to improve prediction accuracy and provide scalable alternatives to conventional estimation methods. The study evaluates model performance using standard metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), coefficient of determination (R²), and prediction accuracy. Results from previous investigations indicate that LSTM-based models can achieve high predictive performance while effectively handling nonlinear astrophysical relationships. The analysis further explores the astrophysical significance of predicted mass distributions and their implications for understanding quasar evolution across cosmic time. The findings suggest that LSTM algorithms offer a promising computational framework for black hole mass estimation, particularly when large observational datasets are available. While challenges related to interpretability, data quality, and model generalization remain, advances in deep learning and astronomy are expected to enhance predictive capabilities. The integration of artificial intelligence with observational astrophysics provides new opportunities for studying SMBHs and advancing our understanding of the dynamic universe