Proceedings of the 2nd International Conference on Mathematical Statistics and Economic Analysis, MSEA 2023, May 26–28, 2023, Nanjing, China

Research Article

Prediction and Analysis of the Future Development Trend of Wordle

Download309 downloads
  • @INPROCEEDINGS{10.4108/eai.26-5-2023.2334422,
        author={Yixiao  Wang and Wenhao  He and Yuxuan  Wang},
        title={Prediction and Analysis of the Future Development Trend of Wordle},
        proceedings={Proceedings of the 2nd International Conference on Mathematical Statistics and Economic Analysis, MSEA 2023, May 26--28, 2023, Nanjing, China},
        publisher={EAI},
        proceedings_a={MSEA},
        year={2023},
        month={7},
        keywords={wordle prediction model of epidemic diseases correlation analysis pls regression model},
        doi={10.4108/eai.26-5-2023.2334422}
    }
    
  • Yixiao Wang
    Wenhao He
    Yuxuan Wang
    Year: 2023
    Prediction and Analysis of the Future Development Trend of Wordle
    MSEA
    EAI
    DOI: 10.4108/eai.26-5-2023.2334422
Yixiao Wang1,*, Wenhao He1, Yuxuan Wang1
  • 1: China Agricultural University
*Contact email: wangyixiaoleo@qq.com

Abstract

Wordle is a popular puzzle game provided by New York Times. Exploring the factors that affect games makes sense for the future development of Wordle. We mainly study the future trend of the number of reported results and how to predict the percentage of scores. First, we set up the propagation model of Wordle with the reference of the model of epidemic diseases. Then we use particle swarm optimization algorithm to optimize the unknown parameters, we use the adjusted iterative model to simulate the changes of the number of Wordle reports. Moreover, in order to probe whether any attribute of the word will affect the percentage of difficult mode records, we selected several important attributes including parts of speech and frequency to label data and conducted correlation analysis on them. Finally we analyze the impact of the word attributes on the percentage in difficult mode. Due to the need to predict the percentage scores for a future date, we added the change of time into the prediction model and build the regression model with time variables containing contest number, the day of the week, if the day is weekend and word attributes variables containing part of speech and frequency as independent variables. Finally, we get the coefficient relationship between different tries and these variables, and regression prediction equations were constructed.