Training dataset consists of a sample tuk rides completed. Nearly 10% of the fares are incorrect. The aim should be to train a model based on the given dataset and predict whether the fare of a new set of trips is correct or not.