An Aviation Industry Recommender System(AIRS) using K-nearest Neighbour and Cosine Similarity

The Aviation Industry has flourished recently with a big boom in tourism.However, tourists are confused about selecting the best airline amongst available options. The existing systems just compare the price and analyze past data related to it and predict accordingly. We focus on calculative analysis of multivariate data derived through survey for recommendation. Our approach is to develop a system which will recommend an airline which is supreme amongst everyone.The broad criteria for selection is based on the quality of services, network & passengers per aircraft and factor of safety of the aircraft. With the help of k-nearest neighbor as the main algorithm, assisting with directional features provided by cosine similarity model makes our system robust in predicting output. The architecture also possesses calculations using pivot matrix along with a random number generating method recommending the best airline for given input. Label encoding technique is implemented in preprocessing to resolve and rectify error of conversion of deferring input data types. This paper covers fundamental application of KNN as well as cosine similarity in an intuitive way.