In today’s world, people take high amounts of responsibilities. These high responsibilities lead to stress and wear out. To relax everybody thinks of going on a holiday or vacation. Reservations are made for hotels at some destination. It is not necessary that once the reservation is made, the person who made the reservation will show up at the day of reservation. They might have to cancel it due to unforeseen circumstances. These reservation cancellations are a huge problem faced by the hotel owners or managers as they are at a loss because they cannot take another booking on the same day.

Catering to all the problems stated above, we have developed a model which can predict whether the made reservation will be cancelled. This model can be used by hotel owners and managers for predicting the booking cancellations. This model does not ask for personalised data such as name, age, gender, religion, address, etc. You can use this web application anytime you receive a reservation, and the model will the output for the reservation cancellation.This model should be used for reducing losses by the hotel owners and should not be the only factor for making important decisions.

Technical Architecture

Project Flow:

Project Flow

• User interacts with the UI to enter the input.

• Entered input is analyzed by the model which is integrated.

• Once model analyses the input the prediction is showcased on the UI

• Define problem / Problem understanding

• Specify the business problem

• Business Requirements

• Literature Survey

• Social or Business Impact

• Data Collection 

• Collect the dataset

• Data Preparation

• Exploratory Data Analysis 

• Descriptive statistical 

• Checking Unique Values

• Visual Analysis 

• Data Pre-processing

• Splitting features and target 

• Balancing the data

• Splitting into training and validation data

• Model Building 

• Creating a function for evaluation

• Training and testing the Models using multiple algorithms 

• Performance Testing & Hyperparameter Tuning 

• Testing model with multiple evaluation metrics 

• Comparing model accuracy for different number of features.

• Comparing model accuracy before & after applying hyperparameter tuning 

• Checking the Predictions

• Model Deployment 

• Save the best model 

• Integrate with Web Framework 

Create project folder which contains files as shown below: