Word difficulty classification prediction based on ARIMA and DBSCAN

Abstract

Wordle is an online word-guessing game, where players are given a five-letter word each day and are asked to guess the word within six tries. Since the launch of the game, it has been loved by a wide range of players. To classify the experiment data and point the ‘EERIE’ at an appropriate label, the article uses unsupervised learning (with DBSCAN and ARIMA), cluster “sum(x)” and “rank(y)” and obtain the values (5.18185152, 8.83924243). Meanwhile, according to the image ‘The word positions by DBSCAN’, we define the ‘EERIE’ in moderately easy.