Author(s)

Yixuan Zeng ¹, Zihuang Gong ²

Affiliation(s)

¹ Guanggu Future School, Wuhan, 430074, Hubei, China
² Shanghai Experimental Foreign Language School, Shanghai, 201800, China

Corresponding Author

Yixuan Zeng

ABSTRACT

In the early hours of April 15, 1912, the RMS Titanic—the world's largest and most luxurious ocean liner at the deeply grieved time—struck an iceberg during its maiden voyage in the North Atlantic. Referred to as "unsinkable", the enormous ship sank completely within just 2 hours and 40 minutes. Due to a severe shortage of lifeboats and delayed rescue efforts, over 1,500 of the 2,224 passengers crew deceased, with only around 700 surviving, cementing the disaster as one of the most tragic maritime perils of the 20th century. This dissertation focuses on analyzing the key factors which influence passenger survival rates on the Titanic. By constructing a hard-voting ensemble model integrating logistic regression, random forest, and decision tree algorithms, we predict individual survival probabilities based on their unique characteristics. Our methodology systematically addresses missing data, utilizes innovative feature engineering to derive new variables, and applies a practicable model to the dataset, thereby uncovering the core determinants of survival and their underlying mechanisms. The integrated data processing and model construction demonstrate strong performance, achieving an accuracy of 89.21%. Expanding this model to maritime safety applications enables systematic collection and analysis of multi-source data, enhancing the precision of risk prediction significantly. In this context, the result not only improves disaster prevention but also contributes to the advancement of data-driven safety capabilities in modern navigation.

KEYWORDS

Statistical Inference; Arranging and Analysing Data; Titanic; Logistic Regression; Random Forest; Hard Voting

CITE THIS PAPER

Yixuan Zeng, Zihuang Gong. Statistical Inference and Multi-Model Comparative Study on the Titanic Survival Rate. Information and Knowledge Management (2026). Vol. 7, No.1, 56-66. DOI: http://dx.doi.org/10.23977/infkm.2026.070107.

REFERENCES

[1] Megan L. Risdal. (2017). Exploring Survival on the Titanic. Kaggle.com; Kaggle. https://www.kaggle.com/code/ mrisdal/exploring-survival- on-the-titanic.
[2] Hiteshp. (2018). Head Start for Data Scientist. Kaggle.com; Kaggle. https://www.kaggle.com/code/hiteshp/head-start-for-data-scientist#feature-engineering.
[3] Gregory, D. N., & Bader, K. (2018). Logistic and linear regression assumptions: violation recognition and control. In Proceedings of Midwest SAS User Group 2018 conference, Indiana, AA-091 (pp. 1-22).
[4] Shareef, A. Q., & Kurnaz, S. (2023). Deep Learning Based COVID-19 Detection via Hard Voting Ensemble Method: AQ Shareef, S. Kurnaz. Wireless Personal Communications, 1-12.

Sponsors, Associates, and Links

---

• Journal of Language Testing & Assessment
  
  
• Military and Armament Science
  
  
• Media and Communication Research
  
  
• Journal of Human Movement Science
  
  
• Art and Performance Letters
  
  
• Lecture Notes on History
  
  
• Lecture Notes on Language and Literature
  
  
• Philosophy Journal
  
  
• Science of Law Journal
  
  
• Journal of Political Science Research
  
  
• Journal of Sociology and Ethnology
  
  
• Advances in Broadcasting