Automated Sequence Clustering for Automation Discovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying significant business processes for automation are manual, inefficient, and biased, often missing important flows and requiring extensive resources, with existing systems struggling to group sporadic sequences with gaps and requiring high-level event logs for every action.
Innovation Solution
A system that collects and analyzes low-level user actions to automatically identify case IDs, using unsupervised machine learning to connect dispersed actions across different times and create longer, significant automation routines without the need for high-level event logs or customer intervention, employing a memory-state-machine to construct sequences and cluster actions based on distinct data characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual methods are used to identify business processes for automation, then the process can be performed with existing tools and knowledge, but the process is time-consuming, expensive, and prone to bias and missed opportunities
Solution Approach 1:
The patent replaces manual mechanical analysis of business processes with an automated machine learning system. The system automatically extracts actions from logs, identifies sequences, detects case IDs, and groups related actions without human intervention, thereby eliminating the time-consuming and biased manual process while maintaining or improving identification accuracy.
Solution Approach 2:
The system performs self-service by automatically analyzing its own output to identify patterns and business processes. The machine learning model continuously learns from the data, automatically adjusts its analysis criteria, and generates automation recommendations without requiring external manual guidance or intervention, significantly reducing the time and resources needed for process identification.
2Reliability
If existing systems use high-level event logs for every action, then the data structure is simplified and easier to process, but the systems cannot group sporadic sequences with gaps in time or context
Solution Approach 1:
The patent segments the data analysis into distinct functional components: action extraction, sequence identification, case ID detection, and grouping. By processing data in these segmented stages rather than requiring a pre-structured high-level log format, the system achieves reliable sequence grouping while maintaining flexibility in data collection and processing complexity.
Solution Approach 2:
The system introduces an intermediary processing layer that transforms raw action logs into structured sequences. This intermediary layer includes the machine learning model that acts as a mediator between the raw data and the final business process identification, enabling the system to handle sporadic sequences with gaps in time or context without requiring complex pre-structured data.
3Ease of operation
If unsupervised machine learning is used to automatically identify case IDs, then customer intervention and labeling are eliminated, but the system must handle unlabeled data and extract patterns without guidance
Solution Approach 1:
The unsupervised machine learning system performs self-service by automatically learning and identifying case ID patterns from unlabeled data. The system extracts meaningful patterns and groups sequences without requiring customer intervention, labeling, or external guidance, thereby simplifying operation while the underlying complexity of pattern recognition is handled internally by the learning algorithm.
Solution Approach 2:
The system changes the approach from supervised learning with labeled data to unsupervised learning with unlabeled data. By transforming the learning paradigm and using algorithms that automatically detect patterns and cluster similar actions, the system eliminates the need for manual labeling while effectively handling the complexity of pattern recognition through automated parameter adjustment and feature extraction.
Data Source
AI summary
A method and system for analyzing and connecting computer-based actions into sentences may include for a series of computer-based actions, determining the case ID for the action for each action where an identifier or case ID can be determined, creating sequences of subsets of the series of computer-based actions using the case ID, and merging sequences having computer-based actions having the same case ID. A set of case IDs may be extracted from the actions using a clustering algorithm based on features of potential case IDs such as gaps in appearance of potential case IDs in a sequence of actions and consecutive appearances of potential case IDs in a sequence of actions. The extracted case IDs may be used when creating sequences.