Time-Parameter Model Estimation for Non-Stationary Trends
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing models struggle to accurately capture data trends in non-stationary environments like semiconductor factories and chemical plants, failing to account for data insufficiency at specific points in time, leading to inaccurate predictions.
Innovation Solution
A method involving the estimation of multiple models over varying data periods and modeling time axis changes to enhance accuracy, allowing for precise analysis even at points with insufficient data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the data period is extended to improve model accuracy, then the model can capture more historical patterns, but the latest trends cannot be captured timely
Solution Approach 1:
The patent applies dynamics by making the model adaptive to changing conditions over time. It estimates models at multiple different time periods and selects or combines them based on current data characteristics, allowing the system to transition between capturing historical patterns and responding to latest trends dynamically.
Solution Approach 2:
The patent segments the time period into multiple different time periods for model estimation. Instead of using a single fixed data period, it divides the time range and estimates separate models for each segment, then uses a selection mechanism to choose the most appropriate model for the current situation.
2Loss of time
If the data period is shortened to capture latest trends, then timeliness is improved, but data deficiency occurs leading to inaccurate models
Solution Approach 1:
The patent creates a multi-functional system that can serve different purposes with different models. By estimating multiple models at different time periods and having a selection mechanism, the system can universally handle both data-scarce situations (using models from longer periods) and data-rich situations (using models from recent periods).
Solution Approach 2:
The system dynamically selects which model to use based on current conditions. The model selection is not fixed but adapts to the available data and current trends, allowing the system to optimize between accuracy and timeliness based on real-time requirements.
3Device complexity
If a single model is used to simplify the system, then device complexity is reduced, but the system cannot adapt to non-stationary environments
Solution Approach 1:
The patent implements a dynamic model selection system that adapts to non-stationary environments. Instead of using a single static model, it estimates multiple models at different time periods and selects the most appropriate one based on current data characteristics, enabling the system to adapt to changing environmental conditions.
Solution Approach 2:
The system incorporates feedback through model validation and selection mechanisms. It evaluates the performance of models estimated at different time periods and selects the one that best fits the current data, creating a feedback loop that continuously adapts to environmental changes.
Data Source
AI summary
According to an embodiment, an information processing device includes processors. The processors receive input of a plurality of pieces of input data obtained during K time periods. K is an integer equal to or greater than two. The processors estimate K first models. Each of the K first models receives input of input data and outputs output data. Each of the K first models is estimated for each period of the K time periods, using a plurality of pieces of input data obtained during the each period. The processors estimate a second model that indicates a relationship between first time parameters related to times of the K time periods, and the K first models. The processors estimate a first model corresponding to a specified second time parameter, based on the estimated second model.