Predicting the outcomes of consideration of bills in the State Duma using a neural network mode

In this paper, we trained our machine learning and neural network models to predict the outcome of the bills’ consideration in the Russian State Duma. We used data collected from October 24, 1994 to December 1, 2022. A rubert-tiny model was used for data preprocessing, a random forest classifier, logistic regression and a neural network model of 3 linear layers were used for prediction. The models demonstrated qualitative results on real-life data: 94% accuracy was achieved by using attached documents’ texts as the models’ parameters and 87% accuracy by training on the data from the bill’s passport. Based on the text of the draft alone, the model’s accuracy accounted for 75.6%. The most important factor influencing the prediction result was the text of the Governmental conclusion. The second most important parameter influencing the results was the “Subject of the right of legislative initiative” with 31.5% of significance in the models’ prediction. Random forest algorithm performed best when working with combined text data and bill passport parameters while logistic regression and neural network showed promising results based on textual parameters alone. The probability of bill’s adoption was not significantly influenced by the financial justification text, the explanatory note text or the subject matter of the bill. The author draws conclusions about the practical applications of the trained models, as well as identifies further scientific problems in the field of mathematical analysis and prediction of lawmaking.

1. Ansolabehere S., Snyder Jr J.M., Stewart III C. Candidate positioning in US House elections. American journal of political science. 2001, Vol. 45, N 1, P. 136–159. DOI: https://doi.org/10.2307/2669364

2. Ba J.L., Kiros J.R., Hinton G.E. Layer normalization. arXiv preprint arXiv:1607.06450.2016.

3. Bari A., Brower W., Davidson C. Using artificial intelligence to predict legislative votes in the United States congress. 2021 IEEE 6 th International Conference on Big Data Analytics (ICBDA). 2021, P. 56–60.

4. Budhwar A., Kuboi T., Dekhtyar A., Khosmood F. Predicting the vote using legislative speech. Proceedings of the 19 th annual international conference on digital government research: governance in the data age. 2018, P. 1–10.

5. Cherryholmes C.H., Shapiro M.J. Representatives and roll calls: a computer simulation of voting in the eighty-eighth congress. Bobbs-Merrill, Indianapolis, 1969, 196 p.

6. Chiou F.Y., Rothenberg L.S. When pivotal politics meets partisan politics. American journal of political science. 2003, Vol. 47, N 3, P. 503–522. DOI: https://doi.org/10.2307/3186112

7. Clinton J., Jackman S., Rivers D. The statistical analysis of roll call data. American political science review. 2004, Vol. 98, N 2, P. 355–370. DOI: https://doi.org/10.1017/s0003055404001194

8. Devlin, J., Chang, M.W., Lee, K., Toutanova, K Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805. 2018.

9. Gerrish S.M., Blei D.M. Predicting legislative roll calls from text. Proceedings of the 28 th International Conference on Machine Learning. ICML 2011, 2011.

10. Gerrish S., Blei D. How they vote: Issue-adjusted models of legislative behavior. Advances in neural information processing systems. 2012, Vol. 25, P. 2753–2761.

11. Jackman S. Multidimensional analysis of roll call data via Bayesian simulation: Identification, estimation, inference, and model checking. Political analysis. 2001, Vol. 9, N 3, P. 227–241. DOI: https://doi.org/10.1093/polana/9.3.227

12. Karyagin M.E. The deputies voting strategies: what open data says about the work of the State Duma? Political science (RU). 2023, N 1, P. 303–321. DOI: http://www.doi.org/10.31249/poln/2023.01.13 (In Russ.)

13. Khashman Z., Khashman A. Anticipation of political party voting using artificial intelligence. Procedia computer science. 2016, Vol. 102, P. 611–616. DOI: https://doi.org/10.1016/j.procs.2016.09.450

14. Kim I.S., Londregan J., Ratkovic M. Estimating spatial preferences from votes and text. Political analysis. 2018, Vol. 26, N 2, P. 210–229. DOI: https://doi.org/10.1017/pan.2018.7

15. Korn J.W., Newman M.A. A deep learning model to predict congressional roll call votes from legislative texts. Machine learning and applications: an international journal (MLAIJ) 2020, Vol. 7, N 3–4.

16. Kraft P., Jain H., Rush A.M. An embedding model for predicting roll-call votes. Proceedings of the 2016 conference on empirical methods in natural language processing. 2016, P. 2066–2070.

17. Ladha K.K. Coalitions in congressional voting. Public choice. 1994, Vol. 78, P. 43–63. DOI: https://doi.org/10.1007/bf01053365

18. Matthews D.R., Stimson J.A. Yeas and nays: Normal decision-making in the US House of Representatives. New York: Wiley-Interscience, 1975, 190 p.

19. McFadden D. Econometric models of probabilistic choice. In: Manski C., McFadden D. (eds). Structural analysis of discrete data with econometric applications. Cambridge: MIT Press, 1981, P. 198–272.

20. Montufar, G.F., Pascanu, R., Cho, K., & Bengio, Y. On the number of linear regions of deep neural networks. Advances in neural information processing systems. 2014, Vol. 4, P. 2924–2932.

21. Nay J.J. Predicting and understanding law-making with word vectors and an ensemble model. PloS one. 2017, Vol. 12, N 5, P. e0176999. DOI: https://doi.org/10.1371/journal.pone.0176999

22. Panov P.V., Sulimov K.A. Ideological «harmony» in the State Duma: emergence, content, boundaries. Political science (RU). 2023, N 1, P. 61–91. DOI: http://www.doi.org/10.31249/poln/2023.01.03 (In Russ.)

23. Pomiguev I.A. The Council of the State Duma: real veto player or a technical executive? Polis. Political studies. 2016, N 2, P. 171–183. DOI: https://doi.org/10.17976/jpps/2016.02.12 (In Russ.)

24. Pomiguev I.A., Alekseev D.V. Resetting bills: discontinuity as a political technology for blocking policy decision. Polis. Political studies. 2021, N 4, P. 176–191. DOI: https://doi.org/10.17976/jpps/2021.04.13 (In Russ.)

25. Poole K.T., Rosenthal H. A spatial model for legislative roll call analysis. American journal of political science. 1985, P. 357–384. DOI: https://doi.org/10.2307/2111172

26. Ruby, U., Yendapalli, V. (Binary cross entropy with deep learning technique for image classification. International journal of advanced trends in computer science and engineering. 2020, Vol. 9, N 10. DOI: https://doi.org/10.30534/ijatcse/2020/175942020

27. Sharma S., Sharma S., Athaiya A. Activation functions in neural networks. International journal of engineering applied sciences and technology. 2017, Vol. 6, N 12, P. 310–316. DOI: https://doi.org/10.33564/ijeast.2020.v04i12.054

28. Snyder Jr J.M., Groseclose T. Estimating party influence in congressional roll-call voting. American journal of political science. 2000, Vol. 44, N 2, P. 193–211. DOI: https://doi.org/10.2307/2669305

29. Srivastava N. Improving neural networks with dropout. Toronto, 2013, 26 p. Mode of access: https://www.cs.toronto.edu/~nitish/msc_thesis.pdf (accessed: 28.05.2024)

30. Wang, E., Liu, D., Silva, J., Carin, L., Dunson, D. Joint analysis of time-evolving binary matrices and associated documents. Advances in neural information processing systems. 2010, Vol. 23.

31. Wang, E., Salazar, E., Dunson, D., Carin, L. Spatio-temporal modeling of legislation and votes. Bayesian analysis. 2013, Vol. 8, N 1, P. 233–268. DOI: https://doi.org/10.1214/13-ba810

32. Yang, Y., Lin, X., Lin, G., Huang, Z., Jiang, C., Wei, Z Joint representation learning of legislator and legislation for roll call prediction. Proceedings of the Twenty-Ninth International Conference on International Joint Conferences on Artificial Intelligence. 2021, P. 1424–1430. DOI: https://doi.org/10.24963/ijcai.2020/198

33. Yano T., Smith N.A., Wilkerson J. Textual predictors of bill survival in congressional committees. Proceedings of the 2012 conference of the North American chapter of the Association for computational linguistics: human language technologies. 2012, P. 793–802.