Streak Discovery in Time-Series Data Using Linear Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing streak identification techniques in time-series data face limitations such as loss of information due to binary conversion, impracticality in predicting streak continuity, and inability to handle overlapping streaks, which are crucial for accurate decision-making in dynamic environments like investment analysis and sports performance evaluation.
Innovation Solution
A method and system that convert time-series data into a linear graph, allowing for node merging and identification of streaks based on value increases, enabling the detection, ranking, and prediction of streaks in both binary and numerical data formats, including overlapping streaks, using statistical and probabilistic models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If time-series data is converted to binary data using threshold values, then streak identification becomes simpler, but information loss occurs
Solution Approach 1:
The patent segments the time-series data processing into multiple stages: first converting to binary format for simplified streak detection, then using the identified streak positions to analyze the original numerical data. This segmentation allows the system to benefit from both simplified processing and complete information retention.
Solution Approach 2:
The patent introduces streak positions and streak values as intermediary elements that bridge the binary conversion process and the final analysis. These intermediaries carry essential information from the simplified binary representation back to the original numerical data context, preventing information loss while maintaining processing simplicity.
2Device complexity
If existing streak identification algorithms are used, then processing is simpler, but overlapping streaks cannot be detected
Solution Approach 1:
The patent implements a dynamic streak identification process where streak values are updated iteratively. When overlapping streaks are detected, the algorithm dynamically adjusts streak values to reflect the overlap, allowing the system to adapt to complex patterns while maintaining reasonable computational complexity.
Solution Approach 2:
The patent merges multiple streak detection results by combining streak values at overlapping positions. This merging process allows the system to detect and represent overlapping streaks comprehensively, enhancing versatility while building upon simpler base algorithms.
3Ease of operation
If binary conversion is applied to time-series data, then streak detection is easier, but numerical data patterns are lost
Solution Approach 1:
The patent performs preliminary streak detection on binary-converted data to identify potential streak positions and patterns. These preliminary results are then used as guides for analyzing the original numerical data, ensuring that numerical patterns are captured while benefiting from the ease of binary detection.
Solution Approach 2:
The system uses feedback from the binary conversion process to refine the analysis of numerical data. Streak positions identified in the binary representation provide feedback that guides the examination of numerical patterns, ensuring that information is preserved rather than lost.
4Device complexity
If disjointed streaks are assumed, then analysis is simpler, but practical overlapping streak cases are missed
Solution Approach 1:
The patent transitions from a static assumption of disjointed streaks to a dynamic model that allows streaks to overlap. The streak values are dynamically updated to reflect overlapping relationships, improving reliability while managing complexity through iterative processing.
Solution Approach 2:
The patent changes the parameter representation from simple streak presence/absence to streak values that can represent overlapping conditions. This parameter change allows the system to capture practical overlapping cases while building upon simpler analytical frameworks.
Data Source
AI summary
The disclosed embodiment relates to identifying performance regions in time-series data. An exemplary method comprises identifying, with a computing device, one or more streaks in the time-series data based on at least one streak parameter, ranking, with a computing device, the identified streaks based on at least one characteristic of the identified streaks, and predicting, with a computing device, a future occurrence of at least one streak based on the characteristics of the identified streaks. The steps of identifying and ranking may be carried out using at least one of a linear graph method, a statistical based approach, a curve-line intersection method, and a hypothesis-based method, and the step of predicting the future occurrence of at least one streak may comprise predicting at least one of how long a current streak will continue, when a current streak will end, and when a new streak will begin. The disclosed embodiment also relates to a system and computer-readable code that can be used to implement the exemplary methods.