Application and Research of Prediction Model Based on ARIMA-LSTM and Multi-Sequence LSTM Algorithm - Taking Wordle as an Example

Zhengze Li; Hui Zhang

doi:10.54097/hset.v60i.10333

Authors

Zhengze Li
Hui Zhang

DOI:

https://doi.org/10.54097/hset.v60i.10333

Keywords:

ARIMA-LSTM, wavelet denoising, canonical correlation analysis (CCA), Wordle puzzle

Abstract

The article aims to explore the relationship between the changing trends in the number of Wordle players and the distribution of players' attempts at the game based on word and time attributes. This paper employs wavelet analysis and ARIMA-LSTM model to predict the number of players, and uses canonical correlation analysis and multi-sequence LSTM model to respectively determine the correlation and classification results of word attributes and player attempt distribution. The results indicate that on March 1st, 2023, the number of Wordle players is 16,352. If the solution word on March 1st is 'EERIE', the distribution of player attempts will be [0,8,19,25,23,17,8].

Downloads

Download data is not yet available.

References

Ling Xiaojing, Zhuang Yaming, Sun Liling. Research on SEIR Online Opinion Dissemination Model with Saturated Contact Rate[J]. Journal of Intelligence, 2015, 34(3): 150-155.

Zhang Jidong, Jiang Liping. A study on the dynamic evolution model of mobile social network opinion dissemination incorporating user group behavior[J]. Journal of Modern Information,2021, 41(5): 159-166+177.

Peng Zhihang, Bao Changjun, Zhao Yang, et al. ARIMA multiplicative seasonal model and its application in predicting the incidence of infectious diseases[J/OL]. Journal of Applied Statistics and Management, 2008(2): 362-368.

SHAM N M, KRISHNARAJAH I, SHITAN M, et al. Time series modelon hand, foot and mouth disease in Sarawak, Malaysia[J/OL]. Asian Pacific Journal of Tropical Disease, 2014, 4(6): 469-472.

Ouyang Zisheng, Lu Ming, Zhou Xuewei. Study on Risk Spillover and Early Warning in China's Financial Sector Based on TVP-VAR-LSTM Model[J]. Journal of Statistics and Information, 2022, 37(10): 53-64.

Li, H. J., &Yang, J. M., &Deng, T. T., &Zong, J. L. (2014). Research on Turbine vibration prediction based on wavelet analysis. Computer engineering and applications, 12, 263-265+270.

Lou, L., &Liu, L., &Liu, X. J., &Shi, S. Z. (2021). ARIMA-LSTM hybrid model based on wavelet denoising and its prediction of a stock price index. Journal of Changchun University of Science and Technology, 02, 119-123.

Lai Xiaoying, Qian Jun. The study of ARIMA-LSTM-XGBoost weighted combination model in predicting the incidence trend of pulmonary tuberculosis [J]. Modern Preventive Medicine, 2021,48(01):5-9.

Yu Mingjie, Guo Peng. Research on the relationship between input and output of regional innovation system based on canonical correlation analysis [J]. Science of Science and Management of S.& T, 2012,33(06):85-91.

Qu Zongxi, Sha Yongzhong, Li Yutong. Research on ensemble prediction of major infectious diseases based on grey wolf optimization and multi-machine learning - taking COVID-19 epidemic as an example[J]. Data Analysis and Knowledge Discovery, 2022,6(08):122-133.