Prediction of Wordle Game based on BP neural network optimization based on Grey Wolf algorithm

Abstract

In 2021, the word guessing game Wordle became an overnight hit around the world. It updated a different “inscription” every day, requiring players to guess a five-letter “inscription” within six times (more than six times deemed unsuccessful). This paper aims to use the BP neural network algorithm optimized by Grey Wolf algorithm to build a multi-input multi-output mathematical model through training and data analysis of the huge data set of Wordle game, and predict the number proportion distribution of future players on six guesses (1,2,3,4,5,6,X). This algorithm can reflect the improvement and enhancement of the prediction accuracy of BP neural network optimized by Grey Wolf algorithm compared with the traditional BP neural network, and show more powerful data processing ability, so as to extend the machine learning model to a wider range of prediction problems.