# Modeling the Central Supermassive Black Holes Mass of Quasars via LSTM Approach

arXiv: https://arxiv.org/abs/2301.01459 · HTML: https://mlsanity.com/paper/2301.01459/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 5 of its 81 references that are in the map.
- First posted on arXiv 4 Jan 2023.

## Builds on (most cited first, 5 of 5)

- [Understanding LSTM -- a tutorial into Long Short-Term Memory Recurrent Neural Networks](https://mlsanity.com/paper/1909.09586/) (2019)
- [AI-assisted super-resolution cosmological simulations](https://mlsanity.com/paper/2010.06608/) (2020)
- [Depth-Gated LSTM](https://mlsanity.com/paper/1508.03790/) (2015)
- [QuasarNET: Human-level spectral classification and redshifting with Deep Neural Networks](https://mlsanity.com/paper/1808.09955/) (2018)
- [Deep Learning for Multi-Messenger Astrophysics: A Gateway for Discovery in the Big Data Era](https://mlsanity.com/paper/1902.00522/) (2019)

## Abstract (arXiv)

One of the fundamental questions about quasars is related to their central supermassive black holes. The reason for the existence of these black holes with such a huge mass is still unclear and various models have been proposed to explain them. However, there is still no comprehensive explanation that is accepted by the community. The only thing we are sure of is that these black holes were not created by the collapse of giant stars, nor by the accretion of matter around them. Moreover, another important question is the mass distribution of these black holes over time. Observations have shown that if we go back through redshift, we see black holes with more masses, and after passing the peak of star formation redshift, this procedure decreases. Nevertheless, the exact redshift of this peak is still controversial. In this paper, with the help of deep learning and the LSTM algorithm, we tried to find a suitable model for the mass of central black holes of quasars over time by considering QuasarNET data. Our model was built with these data reported from redshift 3 to 7 and for two redshift intervals 0 to 3 and 7 to 10, it predicted the mass of the quasar's central supermassive black holes. We have also tested our model for the specified intervals with observed data from central black holes and discussed the results.