Using API with Logistic Regression Model to Predict Hotel Reservation Cancellation by Detecting the Cancellation Factors

The aim of establishing hotels is to provide a service activity to its customers with the aim of making a profit. So, for that the cancellations are a key perspective of inn income administration since their effectiveness on room reservation systems. Cancelling the reservation eliminates the outcome. Many expected factors affect this problem. By knowing these factors, the hotel management can make a suitable cancellation policy. This project aims to create an API that can provide a function to predict if a reservation is most likely to cancel or not. That API can integrate with the hotel management systems to evaluate each reservation process with the same parameters. To do this, the study starts by defining the factors using Chi test, correlation to find the effective variables, and coefficient of the variables in the linear regression. And the results that have been found for the factors are: is repeated guest, previous cancellations, previous bookings not cancelled, required car parking spaces, and deposit type. For API function, the intercept and coefficients have been used from the logistics regression model to create a scoring function. Scoring function can be calculated by the sum of the factors multiplied by their coefficients in addition to the intercept. This score is to be evaluated as a probability later using the logistic function.