Machine Learning Travel Intent Prediction with Optimized Ticket Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Travel organizations struggle to discern business travelers from leisure travelers based on changing booking practices, as more business travelers book directly through service providers or online travel agencies, limiting data processing capabilities.
Innovation Solution
A method involving multiple machine learning models trained on ticket attribute information from different sources, with hyperparameter tuning to optimize data elements, and iterative training to predict travel intent, using algorithms like decision trees and random forests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If traditional booking data categorization is used to distinguish business travelers from leisure travelers, then the classification process is simple, but the accuracy deteriorates as booking practices change and more business travelers book directly
Solution Approach 1:
The patent replaces the mechanical/manual classification system with an automated machine learning model that uses natural language processing and pattern recognition to classify travelers, eliminating the need for manual categorization while maintaining high accuracy
Solution Approach 2:
The patent transforms the classification approach by changing from simple booking source categorization to analyzing multiple ticket attributes and behavioral parameters, using hyperparameter tuning to optimize the model for accurate classification
2Measurement precision
If all available ticket attribute data elements are used for training the machine learning model, then the predictive accuracy may improve, but the computational resources and training time increase significantly
Solution Approach 1:
The patent extracts and selects only the most relevant and informative data elements from the complete dataset through feature selection and hyperparameter tuning, removing redundant or less useful features to reduce computational burden while maintaining predictive accuracy
Solution Approach 2:
The patent uses a subset of the most critical data elements rather than processing all available data, applying partial action to achieve sufficient accuracy with reduced computational resources
3Adaptability or versatility
If the machine learning model is trained with a large number of data elements, then the model complexity increases, but the training efficiency and deployment speed decrease
Solution Approach 1:
The patent segments the data elements into different categories and prioritizes the most important features for training, organizing the data structure to improve processing efficiency while maintaining model versatility
Solution Approach 2:
The patent trains the model with a carefully selected subset of the most impactful data elements rather than all available features, achieving adequate model capability with faster training and deployment
Data Source
AI summary
A method and system for performing a prediction of a travel intent of a trip are disclosed. The method includes receiving sets of ticket attribute information, identifying a set of data elements, parsing at least one data element, identifying a pattern among a select portion of the parsed at least one data element and setting the respective portion as a data element, performing hyper parameter tuning to reduce a number of data elements to be included in a training dataset among the data elements, and iteratively training a machine learning model to the training dataset and evaluating accuracy of output provided by the trained machine learning model with respect to a reference threshold predicting whether a trip is a business type or a leisure type based on ticket attribute information associated with the trip.