Welcome to Francis Academic Press

Academic Journal of Mathematical Sciences, 2023, 4(1); doi: 10.25236/AJMS.2023.040107.

Prediction modeling of nuclear bomb numbers based on random forests and LSTM


Yinghao Meng1, Erzhen Lou2, Yufei Zheng3, Bowen Liu4, Xun Zheng1

Corresponding Author:
Bowen Liu

1College of Mechanical and Electrical Engineering, Wenzhou University, Wenzhou, China

2College of Mathematics and Physics, Wenzhou University, Wenzhou, China

3School of Information Science and Engineering, University of Jinan, Jinan, China

4College of Big Data and Intelligent Engineering, Southwest Forestry University, Kunming, China


In this paper, we aim to identify the countries with the most frequent nuclear weapons activity in the last decade and to predict the number of nuclear bombs in the world. First we developed the TOPSIS evaluation model and used the number of nuclear tests and the change in nuclear weapons as two indicators for the model. We eventually concluded that North Korea has been the most active country in developing nuclear weapons in the last decade. We then built a random forest prediction model to predict the number of nuclear-armed states. An LSTM prediction model was then built to predict the number of nuclear weapons in the world and in each country. The predictions gave us: the number of nuclear-armed states in the next 100 years is 8. The number of nuclear weapons per country in 2123 is: China (360), France (258), India (183), Iraq (91), North Korea (11), Pakistan (200), Russia (200), UK (191) and USA (3319). Finally, we conclude with a few recommendations for peace in the human world based on our findings.


TOPSIS, Random Forest, LSTM, Nuclear Weapons

Cite This Paper

Yinghao Meng, Erzhen Lou, Yufei Zheng, Bowen Liu, Xun Zheng. Prediction modeling of nuclear bomb numbers based on random forests and LSTM. Academic Journal of Mathematical Sciences (2023) Vol. 4, Issue 1: 44-51. https://doi.org/10.25236/AJMS.2023.040107.


[1] Meng Xiaojing, Chen Xin, Chen Jiajing, Yang Honggang. Application of combined empowerment-TOPSIS in resilience assessment of urban areas under flooding [J/OL]. Journal of Safety and Environment, 2023: 1-9. 

[2] Fang X., Yang N. Construction of nonlinear error correction learning models based on LSTM networks [J/OL]. Statistics and Decision Making, 2022(24): 5-10.

[3] Sun Xiaoming, Yu Wujing, Ren Ruobing, Xiong Wang, Wang Yalan. A study of patent inventor name disambiguation based on decision tree algorithm [J/OL]. Science and Management, 2022: 1-20. 

[4] Shi Dawei, Shen Yang, Ma Chenchen, Dong Jingming, Yan Jiaren. Research on short-time intense precipitation forecasting in different regions of Jiangsu Province based on decision tree algorithm [J]. Meteorological Science, 2022, 42(05): 631-637.

[5] Li Zhen. Vulnerability evaluation and barrier degree analysis of marine economic system in Shandong Province based on entropy power TOPSIS model [J/OL]. Ocean Development and Management, 2023: 1-10. 

[6] Wang Qing. Research on the Prevention of Nuclear Terrorism Crimes [D]. Jilin University, 2011.