Time-Series Interval Estimation via Median Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional approaches face difficulties in properly estimating the interval of time-series data, leading to compromised prediction accuracy due to either coarse or fine granularity, depending on the selection of intervals for imputation of missing values.
Innovation Solution
An estimating device that processes circuitry to estimate the median of intervals between adjacent data pieces as a uniform interval, allowing for the extraction of data pieces at these estimated intervals, thereby minimizing missing values and enhancing prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a narrow interval is selected among multiple intervals between data pieces, then future data can be predicted with finer granularity, but prediction accuracy is significantly lowered due to increased number of imputation values in past data
Solution Approach 1:
The patent changes the parameter selection criterion from arbitrary or manual interval selection to using the median of observed intervals as the basis for determining the uniform time interval. This parameter change optimizes the balance between granularity and accuracy by selecting a representative interval that minimizes imputation requirements while maintaining prediction detail.
2Reliability
If a wider interval is selected among multiple intervals between data pieces, then future data can be predicted with high accuracy since less missing values should be imputed, but prediction granularity becomes coarse
Solution Approach 1:
The patent changes the interval selection parameter from wide intervals to the median of observed intervals, which optimally balances accuracy and granularity. The median represents the typical observation interval, allowing predictions to maintain high accuracy while preserving appropriate levels of detail.
3Loss of information
If the interval of time-series data is unknown, then it cannot be known where data is missing, making proper estimation difficult
Solution Approach 1:
The system performs self-service by automatically estimating the uniform time interval from the input data itself. The processing circuitry calculates the median of intervals between adjacent data pieces, enabling the system to identify missing data locations and perform imputation without external intervention or complex manual configuration.
Data Source
AI summary
An estimating device includes processing circuitry configured to estimate a median of intervals between adjacent data pieces in input time-series data as a uniform interval for the time-series data, and extract data pieces at the estimated uniform intervals from the input time-series data.