Sampling Error Upper Limit Calculation for Time-Series Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for diagnosing time-series data using waveform similarity struggle to balance diagnostic accuracy and processing time due to the lack of a clear method for calculating the sampling error upper limit, leading to inefficiencies in generating appropriate sample subsequences.
Innovation Solution
An information processing apparatus that includes a data acquisition unit, a sampling error upper limit calculation unit, and a sample subsequence generation unit, which calculates and uses a sampling error upper limit to integrate similar learning subsequences into sample subsequences, optimizing the balance between diagnostic accuracy and processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the sampling error upper limit is set high to reduce processing time, then productivity improves, but measurement precision deteriorates
Solution Approach 1:
The patent dynamically adjusts the sampling error upper limit parameter based on the characteristics of the learning data. By calculating the upper limit from statistical properties (standard deviation) of the learning data rather than using a fixed value, the system adapts the parameter to achieve optimal balance between processing speed and diagnostic accuracy for different data sets.
Solution Approach 2:
The patent performs preliminary calculation of the sampling error upper limit using learning data before actual diagnosis. This preliminary action establishes appropriate parameters in advance, allowing the main diagnosis process to run efficiently without real-time parameter adjustment, thus improving processing speed while maintaining accuracy.
2Measurement precision
If the sampling error upper limit is set low to improve diagnostic accuracy, then measurement precision improves, but productivity deteriorates
Solution Approach 1:
The patent changes the sampling error upper limit parameter from a fixed low value to a dynamically calculated value based on learning data characteristics. This allows the system to use the minimum necessary threshold for accuracy while avoiding excessive conservatism that would increase processing time.
Solution Approach 2:
The patent introduces dynamics to the previously static sampling error upper limit by making it adaptable to different learning data sets. The upper limit is recalculated based on the specific characteristics of each learning data set, allowing optimal performance across different scenarios rather than being constrained by a fixed low value.
3Measurement precision
If all learning subsequences are used for diagnosis to maintain accuracy, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent extracts and integrates similar subsequences from the learning data to create a condensed representation. Instead of processing all original subsequences, the system extracts representative sample subsequences that capture the essential patterns, significantly reducing calculation complexity while maintaining diagnostic accuracy.
Solution Approach 2:
The patent merges similar subsequences by integrating them into representative sample subsequences. Multiple similar learning subsequences are combined into single representative samples, reducing the total number of items to be processed during diagnosis while preserving the diagnostic information contained in the original set.
Data Source
AI summary
An information processing apparatus includes a data acquisition unit to acquire input data that is time-series data; a sampling error upper limit calculation unit to calculate, when similar learning subsequences selected from among a plurality of learning subsequences extracted from learning data that is time-series data are integrated to generate a sample subsequence. The information processing apparatus further includes a sampling error upper limit using data taken from the input data, the sampling error upper limit being an upper limit on dissimilarity between the learning subsequences to be integrated; and a sample subsequence generation unit to generate the sample subsequence from the learning data using the sampling error upper limit.