In the ‘Description’ variable, even though most of the sentences are written in English, there are still some values are written in Chinese or Malay. Separating into different language bases and do natural language processing. Another distinct method may be considered to apply to predict the adoption speed, such as using XGBoost for classifying, which follows a sequential method and compare the results to the best Random Forest model, which applies random parallel method. Since to classify 5 labels with a supervised learning method might not obtain an ideal accuracy score, another prediction could be further supported by combining the ‘AdoptionSpeed’ variable into ‘Being adopted’ and ‘Not being adopted’, and then try the same process might increase the accuracy.