The reported scores are collected from Twettier only, so the discussion of the prediction of the change of time and the total number of reports is one-sided because it does not consider most users who do not share the game scores. For this problem, the SIRS model can be used to predict the future total number of reports. This model helps combine the process of real-life viral transmission and the changing pattern of the total number of reports curve, resulting in a prediction interval of [17645, 2469z4]. To investigate whether word attributes affect the percentage of reported scores in the complex mode, we use the feature importance scores ranked by the RReliefF algorithm, which builds a Gaussian regression model to rate the influence of word attributes on reported scores in the difficult mode. We found that two word attributes, frequency of letter use and frequency of word use, had the most significant effect on reported scores in the difficult mode, while the type of letters contained in the word had a less significant effect. we use an integrated prediction model and a Gaussian process regression prediction model; first, the data of the percentage of attempts are downscaled into three indicators by principal component analysis, and then the three parameters of word attributes (frequency of word use, cumulative frequency of letters, and the number of letter repetitions) and the three indicators of word difficulty are used as the training set to train the prediction model, Second, we use the prediction model for the example in the question of 2023 The prediction model to predict the word EERIE on March 1, 2023, as the example given in the question. Finally, the three indicators of word difficulty are derived. Because the distribution of the attempted percentage data is normally distributed, after mathematical backpropagation, the final seven broadcast percentages of the word EERIE were obtained: 0, 11, 36, 27, 17, 8, 1.

Haotian Lin, Xiaoyu Wei, Jiannan Lin, Wendan Liao. Research on Predicting Wordle Results Model. Academic Journal of Computing & Information Science (2023), Vol. 6, Issue 7: 64-71. https://doi.org/10.25236/AJCIS.2023.060710.

[1] http://www.wordcount.org/main.php

[2] C. Guo, J. Zhang and M. Yang, "Weighted Gaussian Process Regression for Single Image Super-resolution Based on Randomized Sample Clustering and Augmentation," 2021 IEEE International Conference on Artificial Intelligence and Computer Applications (ICAICA), Dalian, China, 2021, pp. 744-750.

[3] M. Ward, K. Malmsten, H. Salamy and C. H. Min, "Data Balanced Bagging Ensemble of Convolutional- LSTM Neural Networks for Time Series Data Classification with an Imbalanced Dataset," 2021 IEEE International Symposium on Circuits and Systems (ISCAS), Daegu, Korea, 2021, pp. 1-5.

[4] Jiang Y , Yang Y , Wu Q ,et al. Research on Predicting the Short-term Output of Photovoltaic (PV) Based on Extreme Learning Machine Model and Improved Similar Day[C]//2019 IEEE Innovative Smart Grid Technologies - Asia (ISGT Asia).IEEE, 2019.DOI:10.1109/ISGT-Asia.2019.8881620.

[5] Chen Xiaokang, Tu Xuan, Xu Weidong, Xie Runzhong. Tool Life Prediction Based on Bagging Integrated Gaussian Process Regression Model [J]. Manufacturing Technology and Machine Tool, 2020 (12): 110-115+121.