Research on Outcome Prediction Model Based on LSTM Neural Network Model: An Example of Wordle Game

Abstract

Wordle is popular worldwide, and predicting the results of the game has attracted many people's attention. Based on this, this article uses the LSTM neural network model, information entropy theory to predict and analyze the number of gamers and word difficulty of Wordle's daily words. The results indicate that: (1) The number of players in the Wordle game shows a daily decreasing trend; (2) The information entropy of words with two or more duplicate letters is lower than words without duplicate letters; (3) There is no significant correlation between the information entropy and the pass rate in the Wordle difficult mode.