Temporal Data Mining Automata for Frequent Episode Discovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for temporal data mining struggle to efficiently discover frequent episodes in time-stamped data series that incorporate time durations, as they often require extensive computational resources and fail to effectively handle non-overlapping and non-interleaved episodes.
Innovation Solution
A method and system that utilize automata to iteratively find frequent principal episodes by incrementing episode length, combining episodes, removing non-frequent sub-episodes, and tracking occurrences, with the ability to distinguish between non-overlapping and non-interleaved episodes, to produce a set of frequent principal episodes meeting a specified threshold frequency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to discover frequent episodes in temporal data series, then comprehensive pattern discovery is achieved, but computational resources are excessively consumed and processing efficiency deteriorates
Solution Approach 1:
The patent segments the temporal data series into discrete events with start and end times, and further segments the pattern discovery process into iterative passes that incrementally build episodes of increasing length. This segmentation allows the system to process data in manageable units rather than attempting to analyze the entire temporal series at once, reducing computational burden while maintaining discovery accuracy.
Solution Approach 2:
The patent performs preliminary actions by pre-processing temporal data into standardized event representations with explicit time durations, and by pre-defining the iterative episode discovery framework. These preliminary structures enable more efficient processing during the actual pattern discovery phase, as the system doesn't need to perform complex temporal analysis from scratch during execution.
2Adaptability or versatility
If traditional methods process temporal data with time durations, then episodes with dwelling times can be discovered, but the complexity of handling non-overlapping and non-interleaved episodes increases computational burden
Solution Approach 1:
The patent employs dynamic programming techniques to efficiently track and count occurrences of episodes with varying time durations. The algorithm dynamically adjusts its processing based on the specific episode patterns being discovered, optimizing performance for different episode types (non-overlapping, non-interleaved) without requiring separate complex processing pipelines for each type.
Solution Approach 2:
The patent introduces intermediate data structures and processing layers that mediate between the raw temporal data and the final pattern discovery results. These intermediaries include event representations with explicit duration attributes and iterative episode candidates that systematically explore pattern space, simplifying the overall algorithmic complexity while maintaining comprehensive handling of different episode types.
3Productivity
If iterative passes are used to increment episode length and combine frequent episodes, then computational burden is reduced, but the number of passes through data increases processing time
Solution Approach 1:
The patent maintains continuity of useful action by performing multiple beneficial operations within each iterative pass through the data. Rather than requiring separate passes for different processing tasks, the algorithm simultaneously identifies frequent episodes, combines them to form longer episodes, and validates patterns all within the same data traversal, maximizing the utility of each pass while minimizing total processing time.
Data Source
AI summary
A system for temporal data mining includes a computer readable medium having an application configured to receive at an input module a temporal data series having events with start times and end times, a set of allowed dwelling times and a threshold frequency. The system is further configured to identify, using a candidate identification and tracking module, one or more occurrences in the temporal data series of a candidate episode and increment a count for each identified occurrence. The system is also configured to produce at an output module an output for those episodes whose count of occurrences results in a frequency exceeding the threshold frequency.