Hierarchical Neural Network for IT Ticket Assignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing IT service management systems face challenges in efficiently assigning service incident tickets to the optimal queue and sub-queue, leading to potential misassignment and service level degradation due to large numbers of queues and imbalanced data distribution.
Innovation Solution
A service management system utilizing a hierarchical deep neural network architecture, specifically a Coarse Fine Network (CFNet), that predicts both coarse and fine classes in a single model, using summaries and descriptions of service tickets to optimize ticket assignment and handle imbalanced classification issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional IT service management systems are used to assign tickets to queues, then the system structure remains simple, but the assignment accuracy deteriorates due to large numbers of queues and imbalanced data distribution
Solution Approach 1:
The patent segments the ticket assignment system into hierarchical levels (queue level and sub-queue level) and implements a two-stage neural network model (coarse network for queue prediction, fine network for sub-queue prediction). This segmentation allows the system to handle large numbers of queues by breaking down the complex classification problem into manageable stages, thereby improving assignment accuracy without overwhelming system complexity
Solution Approach 2:
The patent introduces a hierarchical dimension to the ticket assignment system by creating a two-level structure (queues and sub-queues). This dimensional change transforms the flat classification problem into a hierarchical one, allowing the neural network to first predict the broader queue category and then the specific sub-queue, thereby improving assignment precision while managing complexity through structured organization
2Adaptability or versatility
If more queues and sub-queues are created to handle diverse service issues, then service specialization improves, but data imbalance worsens making accurate prediction difficult
Solution Approach 1:
The patent segments the data distribution problem by implementing hierarchical classification where the coarse network handles high-level queue categorization with abundant data, and the fine network handles specific sub-queue classification. This segmentation allows the system to maintain reliable predictions for common queues while still providing specialized routing for rare issues, thereby maintaining prediction reliability across diverse service coverage
Solution Approach 2:
The patent applies local quality by allowing different parts of the neural network (coarse and fine networks) to have different levels of complexity and data requirements. The coarse network handles general patterns with high reliability, while the fine network handles specific cases locally. This allows the system to achieve high service coverage through specialized sub-queues while maintaining overall prediction reliability through the robust coarse-level classification
3Measurement precision
If manual review processes are implemented to verify ticket assignments, then assignment accuracy improves, but processing time increases
Solution Approach 1:
The patent implements preliminary action by using the neural network model to pre-process and predict ticket assignments before manual review. The system generates predicted queue and sub-queue assignments with confidence scores in advance, allowing manual reviewers to quickly verify or correct predictions rather than performing complete analysis from scratch. This preliminary automated classification significantly reduces processing time while maintaining high accuracy through targeted manual verification of uncertain cases
Data Source
AI summary
Embodiments assign an information technology service ticket to a queue and a sub-queue for optimized servicing. Embodiments extract from the service ticket a summary of the service ticket and a description of the service ticket. Embodiments provide as input to a trained neural network model the summary and description, the trained neural network model including a coarse network and a fine network. Embodiments predict the queue using the coarse network and predict the sub-queue using the fine network. Embodiments determine an uncertainty loss for the neural network model and when the uncertainty loss is below a threshold, assign the service ticket to the predicted queue and sub-queue.