Time Series Trend Detection Engine for Large Datasets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in identifying and presenting significant trends within large datasets in a timely and meaningful manner, particularly in fields like healthcare, finance, and network security, where early detection of trends can prevent issues or facilitate prompt responses.
Innovation Solution
A computing device equipped with a trend detection engine, a trend evaluation engine, and a graphical user interface engine analyzes time series data to determine trends, prioritize them based on stability and significance, and alert users through a GUI, using statistical methods to identify trends and calculate trend scores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data processing capacity is increased to analyze larger datasets, then the ability to detect trends improves, but the time required to process and present meaningful data increases
Solution Approach 1:
The patent segments the large dataset into multiple time series, each representing a specific parameter or aspect of the data. This segmentation allows the system to process and analyze trends in smaller, manageable units rather than attempting to analyze the entire large dataset as a single entity, thereby reducing processing time while maintaining detection accuracy.
Solution Approach 2:
The system extracts only the most significant trends from the processed time series data based on calculated significance values. Instead of presenting all processed data, the system selectively extracts and presents only those trends that meet predetermined significance thresholds, reducing the time required to present meaningful information to users.
2Reliability
If all data points are processed to ensure comprehensive trend detection, then detection completeness improves, but computational complexity increases
Solution Approach 1:
The patent changes the parameter of significance by calculating a significance value for each time series based on multiple factors including trend strength, stability, and recency. This parameter transformation allows the system to prioritize and process time series based on their actual importance rather than treating all data points equally, reducing computational complexity while maintaining detection completeness for significant trends.
Solution Approach 2:
The system performs partial processing by focusing computational resources on time series that show potential significant trends rather than uniformly processing all data points. By applying excessive action only where needed (i.e., deeper analysis of promising time series), the system achieves comprehensive detection of meaningful trends while avoiding unnecessary computational complexity in areas where no significant patterns exist.
3Measurement precision
If trend significance thresholds are lowered to detect more trends, then detection sensitivity improves, but the number of false positives increases
Solution Approach 1:
The system incorporates feedback mechanisms by continuously monitoring trend significance values and adjusting detection parameters based on observed performance. The significance calculation itself provides feedback by incorporating multiple factors (trend strength, stability, recency) that collectively reduce false positives while maintaining sensitivity. This multi-factor feedback approach allows the system to detect subtle trends without excessive false alarms.
Solution Approach 2:
The patent creates a composite significance metric by combining multiple individual measures (trend strength, stability, recency) into a single composite significance value. This composite approach is analogous to creating composite materials - each component contributes specific properties that, when combined, create a more robust and accurate detection mechanism that reduces false positives while maintaining sensitivity.
Data Source
AI summary
Examples disclosed herein relate, among other things, to a method. The method may obtain a time series comprising a plurality of data points associated with a sub-segment of a segment, obtaining a plurality of weights associated with a plurality of data point pairs from the plurality of data points, and based on the plurality of weights and the plurality of data point pairs, determine whether the time series comprises a trend. Based on a determination that the time series comprises a trend, the method may calculate a trend score for the trend based on at least one characteristic of at least one of the segment and the sub-segment, and provide the trend for display.