18248760. PREDICTION METHOD, PREDICTION APPARATUS AND PROGRAM simplified abstract (Nippon Telegraph and Telephone Corporation)

PREDICTION METHOD, PREDICTION APPARATUS AND PROGRAM

Organization Name

Nippon Telegraph and Telephone Corporation

Inventor(s)

Hideaki Kin of Tokyo (JP)

Takeshi Kurashima of Tokyo (JP)

Hiroyuki Toda of Tokyo (JP)

PREDICTION METHOD, PREDICTION APPARATUS AND PROGRAM - A simplified explanation of the abstract

This abstract first appeared for US patent application 18248760 titled 'PREDICTION METHOD, PREDICTION APPARATUS AND PROGRAM

Simplified Explanation

The patent application describes a computer-based prediction method that optimizes parameters of two functions to calculate a prediction distribution of future observation values. Here are the key points:

• The method optimizes a parameter of a second function that outputs parameters of a first function from covariates.
• It also optimizes a parameter of a kernel function of a Gaussian process.
• The optimization is done using a series of past observation values and the corresponding covariates.
• The observation values are non-linearly transformed by the first function to follow the Gaussian process.
• The method then calculates a prediction distribution of observation values for a future period using the optimized parameters, the second function, the kernel function, and a series of covariates for that period.

Potential applications of this technology:

• Financial forecasting: Predicting stock prices, exchange rates, or other financial indicators.
• Weather prediction: Forecasting temperature, precipitation, or other weather variables.
• Disease outbreak prediction: Anticipating the spread of infectious diseases based on various factors.
• Demand forecasting: Estimating future demand for products or services.

Problems solved by this technology:

• Accurate prediction: The method optimizes parameters to improve the accuracy of predictions.
• Non-linear relationships: The first function allows for non-linear transformations of observation values.
• Incorporating covariates: The method considers covariates to enhance the prediction model.

Benefits of this technology:

• Improved decision-making: Accurate predictions can help make informed decisions.
• Real-time forecasting: The method can be used to continuously update predictions as new data becomes available.
• Flexibility: The non-linear transformations and incorporation of covariates allow for a wide range of prediction scenarios.

Original Abstract Submitted

A prediction method executed by a computer including a memory and a processor, the method includes: optimizing a parameter of a second function that outputs parameters of a first function from covariates, and optimizing a parameter of a kernel function of a Gaussian process, by using a series of observation values observed in a past and a series of the covariates observed simultaneously with the observation values, wherein values obtained by non-linearly transforming the observation values by the first function follow the Gaussian process; and calculating a prediction distribution of observation values in a period in future to be predicted by using the second function and the kernel function having parameters optimized in the optimizing, and a series of covariates in the period.