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, particularly in handling non-overlapping and non-interleaved episodes, which requires effective frequency measures and computational efficiency.
Innovation Solution
A method and system utilizing automata to track and count occurrences of candidate episodes, incrementally increasing episode length, combining frequent episodes, and removing non-frequent sub-episodes to generate candidate episodes, with automata initialized to detect non-overlapping and non-interleaved occurrences, meeting a 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 efficiency deteriorates due to high computational burden
Solution Approach 1:
The patent segments the temporal data series into episodes of increasing length (k=1, 2, 3, ...) and processes them in successive passes. Each pass discovers frequent episodes of a specific length using automata, avoiding the need to process all possible episode lengths simultaneously. This segmentation reduces computational complexity while maintaining comprehensive pattern discovery.
Solution Approach 2:
The patent performs preliminary actions by first identifying frequent episodes of length k before proceeding to discover episodes of length k+1. The automata are initialized with frequent episodes from the previous pass, and candidate episodes for the current pass are generated based on these preliminary results. This incremental approach ensures that only promising candidates are evaluated, improving computational efficiency.
2Measurement precision
If all possible episode combinations are considered, then complete frequent episode discovery is achieved, but device complexity increases due to the need to handle non-overlapping and non-interleaved episodes
Solution Approach 1:
The patent introduces dynamic state tracking through automata that adapt their behavior based on the current episode length and frequency thresholds. The automata dynamically adjust their state transitions and counting mechanisms to handle episodes of varying lengths while maintaining consistent frequency measurement criteria for non-overlapping and non-interleaved episodes.
Solution Approach 2:
The patent uses automata as intermediary structures that mediate between the raw temporal data and the frequency counting process. These automata serve as intermediaries that systematically track episode occurrences, manage state transitions, and enforce the non-overlapping and non-interleaved constraints, thereby simplifying the overall algorithmic complexity.
3Measurement precision
If successive passes through temporal data series are performed to increment episode length, then frequent episodes of all lengths are discovered, but loss of time occurs due to multiple data series traversals
Solution Approach 1:
The patent creates copies of the temporal data series in the form of automata state representations during each pass. Instead of repeatedly traversing the original data series, the automata maintain copies of the data in their states, allowing efficient processing of episode patterns without repeated full data traversals. This copying mechanism reduces time loss while maintaining comprehensive episode length coverage.
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 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.