PREDICTION OF HOTEL BOOKING CANCELLATION USING DEEP NEURAL NETWORK AND LOGISTIC REGRESSION ALGORITHM

Booking cancellation is a key aspect of hotel revenue management as it affects the room reservation system. has significant effect on which impact demand decisions in industry. In order to reduce effect, applies model addressing this problem with machine learning-based system developed. study, using data collection from Kaggle website name hotel-booking-demand dataset. The research objective was see performance deep neural network method two classification classes, namely cancel and not. Then optimized optimizers learning rate. And attribute most role determining level accuracy Logistic Regression algorithm. results obtained are Encoder-Decoder Layer by adamax optimizer higher than that Decoder-Encoder adadelta optimizer. After adding rate, for encoders decreased rate 0.001. top three ranks each after show smaller accuracy, but we don't know what optimal value is. By algorithm eliminating several attributes, influential state total_of_special_requests, where increases when removed because there 177 variations these attributes