MC-ARIMA for Wordle Game

Our aim was to forecast the player count for the Wordle game. To achieve this, We selected and improved an ARIMA model using Markov Chain techniques. Then, We fitted a Gaussian Regression model to analyze the distribution of user attempt frequencies for specific words. Additionally, We categorized words based on their difficulties using K-means clustering and validated the clustering results with AdaBoost. The key contribution of this project is the development of a predictive model that not only estimates player numbers but also provides insights into player interactions and word difficulty levels.