Time-Series Pattern Matching via Self-Join Computation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing time-series data in investment decisions are inefficient, requiring significant experience and computational power, and are not scalable for investors who cannot devote full-time attention, often leading to missed opportunities and impracticality for large data sets.
Innovation Solution
A system and method that access time-series data, calculate statistical parameters, and utilize external processing resources to compute a similarity self-join, transmitting data over a network to identify minimum subsequence distances and create a full similarity self-join, enabling efficient pattern recognition and alerting investors to favorable trends.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual review of time-series charts is used for investment decisions, then investment intuition can be applied, but significant time and computational power are required, and opportunities may be missed when not monitoring
Solution Approach 1:
The patent introduces an automated pattern recognition system that acts as an intermediary between the investor and time-series data. The system computes similarity self-joins and identifies patterns automatically, allowing investors to receive alerts without continuous monitoring, thus resolving the contradiction between maintaining pattern recognition accuracy and reducing time investment
Solution Approach 2:
The patent replaces the mechanical manual review process with an automated computational system that uses algorithms to identify patterns in time-series data. This substitution eliminates the need for manual chart analysis while maintaining or improving pattern detection capability through systematic computational methods
2Reliability
If full manual monitoring of time-series data is performed, then investment opportunities can be identified, but the approach scales poorly to investors who cannot devote full time
Solution Approach 1:
The system enables part-time investors to benefit from automated pattern recognition that serves itself without requiring continuous human attention. The automated alerting system notifies investors of identified patterns, allowing the system to effectively monitor and analyze data independently while investors receive information asynchronously
Solution Approach 2:
The patent transforms the monitoring approach from requiring continuous human observation to an automated system that can operate independently. By changing the operational parameters from manual review frequency to automated continuous analysis with selective alerting, the system becomes adaptable to investors with varying time commitments
3Measurement precision
If traditional computational methods are used for pattern recognition, then analysis can be performed, but it remains computationally intractable for large amounts of data
Solution Approach 1:
The patent divides the computational task of pattern recognition into manageable segments through the similarity self-join approach. By breaking down the analysis of large time-series datasets into smaller subsequence comparisons and using efficient distance calculations, the system reduces the overall computational burden while maintaining detection accuracy
Solution Approach 2:
The patent employs efficient distance calculation methods and similarity metrics that reduce computational complexity. By optimizing the mathematical parameters and algorithms used for comparing subsequences, the system achieves tractable computation even for large datasets without sacrificing pattern detection precision
Data Source
AI summary
A system includes a memory configured to store instructions and at least one processor configured to execute the instructions. The instructions include accessing time series data, calculating statistical parameters of the time series data, identifying a set of external processing resources, and conveying the time series data and the statistical parameters to the set of external processing resources. The instructions include instructing the set of external processing resources to compute a similarity self-join of the time series data for a window size having a specified length. The instructions include obtaining sets of minimum subsequence distances from the set of external processing resources over a communications network. The full similarity self-join indicates, for each reference subsequence of the specified length within the time series data, a minimum value of distances between the reference subsequence and other subsequences of the specified length within the time series data.