Method and apparatus for predicting time-series data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing time-series analysis methods have low prediction performance when dealing with invalid data, such as power consumption data of 0, which affects the accuracy of predictions in applications like air conditioner energy forecasting.
Innovation Solution
A time-series data prediction method that substitutes invalid data with valid data, using models like Prophet, ARIMA, RNN, or LSTM, and restores prediction data to its original invalid form based on data periods or probability distributions, improving prediction accuracy by comparing output data with reference data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If invalid data (e.g., power consumption of 0) is directly input to the time-series model, then the processing is simple, but the prediction accuracy deteriorates
Solution Approach 1:
The patent applies preliminary action by substituting invalid data with valid data before inputting to the time-series model. The controller identifies invalid data (e.g., power consumption of 0) and replaces it with valid data from the same time period of the previous year or through interpolation, ensuring the model receives clean input data for accurate predictions without dealing with invalid values during the prediction process itself
2Measurement precision
If invalid data is substituted with valid data before prediction, then the prediction accuracy is improved, but the data processing complexity increases
Solution Approach 1:
The patent applies parameter changes by transforming the state of invalid data parameters into valid data parameters. Specifically, it changes the power consumption value from 0 (invalid) to a meaningful value based on historical data from the same period last year, or uses interpolation to generate valid parameter values, thereby making the data suitable for accurate time-series prediction while maintaining systematic processing
3Reliability
If prediction data corresponding to invalid input data is output as is, then the processing is straightforward, but the reliability of prediction results deteriorates
Solution Approach 1:
The patent applies mechanics substitution by replacing the direct output of prediction data with a restored version. Instead of outputting prediction data that corresponds to substituted invalid input data, the system restores these prediction values using historical data from the same period last year or through interpolation, thereby substituting unreliable prediction outputs with more reliable restored values
Data Source
AI summary
A time-series data prediction method is provided. The method includes storing a time-series model, inputting time-series data to the time-series model, and outputting a plurality of pieces of prediction data which is a preset number of pieces of prediction data or more based on the time-series model. The inputting of the time-series data to the time-series model includes substituting invalid data included in the time-series data with preset valid data, and inputting the time-series data including the valid data to the time-series model.