18112944. ENHANCED TIME SERIES FORECASTING simplified abstract (Snowflake Inc.)

From WikiPatents
Jump to navigation Jump to search

ENHANCED TIME SERIES FORECASTING

Organization Name

Snowflake Inc.

Inventor(s)

Michel Adar of Campbell CA (US)

Boxin Jiang of Sunnyvale CA (US)

Qiming Jiang of Redmond WA (US)

John Reumann of Kirkland WA (US)

Boyu Wang of Menlo Park CA (US)

Jiaxun Wu of Sammamish WA (US)

ENHANCED TIME SERIES FORECASTING - A simplified explanation of the abstract

This abstract first appeared for US patent application 18112944 titled 'ENHANCED TIME SERIES FORECASTING

Simplified Explanation

Using an attributes model, this patent application proposes a time series forecasting model that determines a set of features based on time series data, including periodic components. The data can be divided into segments, and each segment can be assigned a weight based on its age. This results in a set of weighted segments of time series data.

The patent application also introduces a trend detection model that analyzes the weighted segments to determine trend data. By combining the set of features and the trend data, a time series forecast can be generated.

  • The patent application proposes a time series forecasting model based on an attributes model.
  • The model determines a set of features from time series data, including periodic components.
  • The time series data can be divided into segments, and each segment is assigned a weight based on its age.
  • The weighted segments of time series data are analyzed using a trend detection model to determine trend data.
  • The set of features and trend data are combined to generate a time series forecast.

Potential Applications

This technology has potential applications in various fields, including:

  • Financial forecasting: Predicting stock prices, market trends, and economic indicators.
  • Demand forecasting: Forecasting product demand, optimizing inventory management, and supply chain planning.
  • Energy forecasting: Predicting energy consumption, optimizing energy production, and grid management.
  • Weather forecasting: Forecasting weather patterns, improving climate models, and disaster preparedness planning.

Problems Solved

This technology addresses several problems in time series forecasting, such as:

  • Incorporating periodic components: By considering periodic components in the set of features, the model can capture recurring patterns in the data.
  • Weighting segments based on age: Assigning weights to segments based on their age helps prioritize recent data, which may be more relevant for forecasting.
  • Trend detection: The trend detection model allows for the identification of underlying trends in the data, enabling more accurate forecasts.

Benefits

The benefits of this technology include:

  • Improved accuracy: By considering both periodic components and trend data, the time series forecasts generated by this model are expected to be more accurate.
  • Flexibility: The model can be applied to various types of time series data and can adapt to different forecasting needs.
  • Efficiency: The use of weighted segments and trend detection helps optimize the forecasting process, reducing computational time and resources required.


Original Abstract Submitted

Using an attributes model of a time series forecasting model, determine a set of features based on time series data, the set of features including periodic components. The time series data may be divided into a set of segments. For each segment of the set of segments, a weight may be assigned using an age of the segment, resulting in a set of weighted segments of time series data. Using a trend detection model of the time series forecasting model, trend data from the set of weighted segments of time series data may be determined. A time series forecast may be generated by combining the set of features and the trend data.