AI Statistical Ensemble for Time Series Forecasting

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing time series forecasting technologies rely on either artificial intelligence (AI) or statistical methods, but lack a unified approach that integrates both AI and statistical techniques for automated time series forecasting.

Innovation Solution

The DI/SM-TS-RIV technology employs a multi-parameter algorithm that autonomously synthesizes a suite of AI and statistical techniques through automatic machine learning, enabling robust and precise time series predictions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If only AI methods are used for time series forecasting, then adaptability and pattern recognition improve, but robustness and interpretability deteriorate

Engineering Contradiction:
ImproveadaptabilityVSAvoidrobustness
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent combines AI methods (neural networks, machine learning) with statistical methods (ARIMA, exponential smoothing) into a unified forecasting system. This merging allows the system to leverage the adaptability and pattern recognition strengths of AI while incorporating the robustness and interpretability of statistical models, thereby resolving the contradiction between adaptability and robustness.

Inventive Principle:
Principle #5Merging (Combining)

2Reliability

If only statistical methods are used for time series forecasting, then robustness and interpretability improve, but adaptability and pattern recognition deteriorate

Engineering Contradiction:
ImproverobustnessVSAvoidadaptability
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent integrates statistical forecasting methods with AI-based approaches, creating a hybrid system where statistical models provide robust baseline predictions and AI components enhance adaptability to complex patterns. This combination ensures both robustness from statistical methods and adaptability from AI techniques.

Inventive Principle:
Principle #5Merging (Combining)

3Measurement precision

If manual configuration of forecasting models is used, then model precision improves, but automation and efficiency deteriorate

Engineering Contradiction:
Improveforecasting accuracyVSAvoidautomation
Core Design Contradiction:
Measurement precisionVSExtent of automation

Solution Approach 1:

The patent implements automated model selection and configuration systems that automatically choose appropriate forecasting methods, tune parameters, and generate predictions without manual intervention. The system self-configures by evaluating multiple models and selecting the optimal combination, thereby maintaining high forecasting accuracy while achieving full automation.

Inventive Principle:
Principle #25Self-service

4Measurement precision

If multiple forecasting methods are integrated, then forecasting accuracy improves, but system complexity deteriorates

Engineering Contradiction:
Improveforecasting accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent divides the forecasting system into modular components, each handling specific forecasting methods (AI-based modules, statistical modules). This segmentation allows multiple methods to be integrated while managing complexity through modular architecture, where each module can be independently configured and maintained.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent creates a universal forecasting platform that can handle multiple types of data and apply various forecasting methods through a common interface. This multi-functionality allows diverse forecasting techniques to be integrated without proportionally increasing system complexity, as the platform provides standardized mechanisms for model selection, parameter tuning, and result aggregation.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP4553724A1Time-series forecasting based on artificial intelligence and statistical methods ensemble
Publication Date: 2025.05.14 UAB RIVILE
  • EP4553724A1 patent drawingFigure 1
  • EP4553724A1 patent drawingFigure 2
  • EP4553724A1 patent drawingFigure 3

AI summary

The present invention (DI/SM-TS-RIV technology) stands at the forefront of artificial intelligence (AI) innovation, offering a groundbreaking approach to time series forecasting and management. Utilizing a sophisticated blend of AI and statistical methodologies, this technology introduces a multi-parameter algorithm that autonomously synthesizes a suite of AI and statistical techniques tailored for precise time series prediction, leveraging the power of automatic machine learning on existing datasets. As a pioneering solution in the market, this invention is unparalleled globally, marking a significant leap forward in predictive analytics. This technology is not only innovative but also versatile, designed for seamless integration with analytical platforms such as: MATLAB, R, Statistica, SPSS, Qubole, BigML, Statwing, Knime, RapidMiner, Pentaho, Domo, Sisense, Open Refine, Orange, Weka, Trifacta, Data Science Studio, DataPreparator, DataCracker, Tanagra, H2O and others. Its introduction promises to transform the landscape of robotic forecasting, providing an unmatched tool for researchers and industries seeking to harness the predictive power of AI with unprecedented efficiency and accuracy.