UCLA

Adults are often told that it may be too late for them to learn a new language, and, sometimes, they may even be intimidated out of continuing their learning journey. While it may be a daunting task, it is not an impossible one. In this thesis, using Duolingo’s dataset of about 13 million learning traces, four different machine learning models are fitted to predict whether the probability of recall of a word is greater than or equal to 0.5. Of the four, logistic regression fared the best with an accuracy score of 93%. It also identified certain word features that contribute to improving the chances of recalling a word.