Active learning-based optimal strategy for Wordle

This article describes a simulation model that utilizes a word bank to guess a puzzle answer and adjust the model in real-time based on results analysis. The model successfully predicted the percentage of the word "ERIRE" for each guess attempt. The presented algorithm classified words into three difficulty levels and analyzed the differences in each indicator. The article established a discriminant function to determine that guessing the word "EEIRE" was of medium difficulty.