Categorization and Classification of Uber Reviews

This paper presents a technique for the categorization of reviews of the brand, Uber. This paper contains a classification algorithm which takes textual data as input. This algorithm takes many reviews and concepts (say, cost, safety, etc.). This algorithm is performed on online conversations happening on social media, taking into account the considered concepts. Further, the categorized reviews are classified according to the sentiment (that is, positive and negative). This paper also performs a comparison of different algorithms—Naive Bayes, KNN, decision tree and random forest. The results of comparison of these different models are then represented using certain identified parameters.