Time-Series Data Pre-Processing Pipeline for Predictive Model Generation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing technologies face challenges in efficiently processing and analyzing large volumes of time-series data, particularly in handling faulty and incomplete data, which hinders the generation of accurate predictive models.

Innovation Solution

An electronic device is designed to build a pre-processing pipeline capable of handling single or multiple types of time-series data, applying appropriate pre-processing modules to segment and process data based on specific time intervals, and using existing trained predictive models when insufficient data is available for inference.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If pre-processing pipeline is applied to time-series data, then data quality and reliability are improved, but processing time and computational complexity increase

Engineering Contradiction:

Improvedata qualityVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The pre-processing pipeline is divided into multiple independent modules (e.g., data cleaning, transformation, feature extraction) that can be selectively applied to different segments of time-series data. This allows the system to process only the necessary portions of data through appropriate preprocessing, reducing overall processing time while maintaining data quality where needed.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary actions by pre-defining and caching pre-processing pipeline configurations and results. When new time-series data arrives, the system can leverage previously computed pre-processing templates and patterns, avoiding redundant computations and significantly reducing processing time for similar data structures.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If predictive model is generated with limited data, then model generation speed increases, but prediction accuracy decreases

Engineering Contradiction:

Improvemodel generation speedVSAvoidprediction accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system generates synthetic data copies that replicate the statistical characteristics and patterns of limited real data. These synthetic copies are created by copying and transforming existing data patterns, allowing the predictive model to be trained on a larger effective dataset without requiring additional real-world data, thus maintaining both speed and accuracy.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The system applies parameter changes by transforming and augmenting the limited available data through various transformations (e.g., time shifts, scaling, normalization). This creates multiple versions of the same data with different parameter characteristics, effectively increasing the training data volume while preserving the underlying patterns needed for accurate predictions.

Inventive Principle:

Principle #35Parameter changes

3Adaptability or versatility

If multiple types of time-series data are processed, then system versatility increases, but device complexity increases

Engineering Contradiction:

Improvedata type handling capabilityVSAvoidsystem complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The pre-processing pipeline is designed as a universal framework that can handle multiple types of time-series data through a single unified architecture. The system uses generic data processing components and algorithms that adapt to different data types (e.g., sensor data, financial data, industrial data) without requiring separate specialized processing systems, thus reducing overall complexity while increasing versatility.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system introduces intermediary layers and abstraction mechanisms that mediate between the diverse input data types and the predictive model. These intermediaries standardize and normalize different data formats and characteristics into a unified representation, allowing the same downstream processing and prediction mechanisms to handle multiple data types without increasing complexity.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20250117696A1Electronic device and method of electronic device generating predictive model based on classification of patterns of time-series data to which pre-processing pipeline has been applied

Publication Date: 2025.04.10 KOREA ELECTRONICS TECH INST

AI summary

A method of generating a predictive model is proposed. The method may include receiving single time-series data or multiple types of time-series data collected in a specific domain, drawing time-series data corresponding to the same domain and time interval, among the received single time-series data or multiple types of time-series data. The method may also include pre-processing the drawn time-series data by applying a pre-processing pipeline built by applying at least one pre-processing module to the drawn time-series data, and generating a pattern classification model for classifying patterns of the pre-processed time-series data based on the clustering of the pre-processed time-series data. The method may further include generating a predictive model for predicting feature information of the drawn time-series data based on a cluster that is generated as the results of the clustering of the pre-processed time-series data, and storing the pattern classification model and the predictive model.