Goal-seek analysis with spatial-temporal data

Patent metadata

Field	Value
Publication number	US-12572792-B2
Application number	US-202017066495-A
Country	US
Kind code	B2
Filing date	Oct 9, 2020
Priority date	Oct 9, 2020
Publication date	Mar 10, 2026
Grant date	Mar 10, 2026

Abstract

Performing a goal-seek analysis of spatial-temporal data by generating a hierarchical cluster according to spatial temporal data, determining a spatial-temporal location input for a target, determining spatial-temporal predictor values for the spatial-temporal location, and adjusting the hierarchical cluster according to and the spatial-temporal predictors.

First claim

What is claimed is:

1 . A computer implemented method for goal-seek analysis of spatial-temporal data, the method comprising:
receiving spatial-temporal data including data collected over time and having a location component from system sensors over a network; generating a spatial-temporal machine learning model comprising hierarchical layers from the spatial-temporal data by: generating a series of hierarchical layers using the spatial-temporal data; evaluating correlations between current system states and previous system states in terms of spatial-temporal variable values and sensor locations; determining the relative contributions of the spatial-temporal data from each location to the current state of each system location; removing any association in a current hierarchical layer between past location-based data and current location-based data when no correlation between the two is found; self-validating by a machine learning model an upward association of location-based data using drop-out testing of the correlations; self-validating by the machine learning model correlations between first layer clusters and successive layer clusters and correlations between associated location data and temporal data characteristics; and adding a new hierarchical layer including clusters highly correlated to clusters of a preceding layer above any layer wherein the self-validation of the layer indicates a high-level of correlation between the clusters of the layer and the associated clusters of a next layer up in the hierarchy; generating, by one or more computer processors, a hierarchical cluster structure for a target according to the spatial-temporal model: determining, by the one or more computer processors, a spatial-temporal location input for the target; determining, by the one or more computer processors, spatial-temporal predictor values for the spatial-temporal location; adjusting, by the one or more computer processors, the hierarchical cluster structure according to the spatial-temporal predictors; and providing, by the one or more computer processors over the network, a system model mapped with the spatial-temporal location input and the spatial-temporal predictor values.

2 . The computer implemented method according to claim 1 , wherein determining the spatial-temporal location input for the target includes conducting a goal-seek analysis using the hierarchical cluster structure.

3 . The computer implemented method according to claim 1 , wherein the hierarchical cluster structure comprises an adaptive layer.

4 . The computer implemented method according to claim 1 , wherein the hierarchical cluster structure comprises multiple adaptive layers.

5 . The computer implemented method according to claim 1 , further comprising adding, by the one or more computer processors, hierarchical cluster layers to the hierarchical cluster structure according to layer self-validation with the target.

6 . The computer implemented method according to claim 1 , wherein determining the spatial-temporal predictor values for the spatial-temporal location includes conducting goal-seek analysis using the hierarchical cluster structure.

7 . The computer implemented method according to claim 1 , further comprising:
determining, by the one or more computer processors, spatial-temporal inputs for the target by conducting a goal-seek analysis using the hierarchical cluster structure, wherein the hierarchical cluster structure comprises an adaptive layer; adding, by the one or more computer processors, hierarchical cluster structure layers according to layer self-validation with the target; and wherein determining spatial-temporal predictor values for the spatial-temporal location includes conducting goal-seek analysis using the hierarchical cluster structure.

8 . A computer program product for goal-seek analysis of spatial-temporal data, the computer program product comprising one or more computer readable storage devices and collectively stored program instructions on the one or more computer readable storage devices, the stored program instructions comprising:
program instructions to receive spatial-temporal data including data collected over time and having a location component from system sensors over a network; program instructions to generate a spatial-temporal machine learning model comprising hierarchical layers from the spatial-temporal data by: generating a series of hierarchical layers using the spatial-temporal data; evaluating correlations between current system states and previous system states in terms of spatial-temporal variable values and sensor locations; determining the relative contributions of the spatial-temporal data from each location to the current state of each system location; removing any association in a current hierarchical layer between past location-based data and current location-based data when no correlation between the two is found; self-validating by a machine learning model, an upward association of location-based data using drop-out testing of the correlations; self-validating by the machine learning model, correlations between first layer clusters and successive layer clusters and correlations between associated location data and temporal data characteristics; and adding a new hierarchical layer including clusters highly correlated to clusters of a preceding layer above any layer wherein the self-validation of the layer indicates a high-level of correlation between the clusters of the layer and the associated clusters of a next layer up in the hierarchy; program instructions to generate a hierarchical cluster structure for a target according to the spatial temporal model; program instructions to determine a spatial-temporal location input for the target; program instructions to determine spatial-temporal predictor values for the spatial-temporal location; program instructions to adjust the hierarchical cluster structure according to the spatial-temporal predictors; and program instructions to provide a system model mapped with the spatial-temporal location input and the spatial-temporal predictor values over the network.

9 . The computer program product according to claim 8 , wherein determining the spatial-temporal location input for the target includes:
conducting a goal-seek analysis using the hierarchical cluster structure.

10 . The computer program product according to claim 8 , wherein the hierarchical cluster structure comprises an adaptive layer.

11 . The computer program product according to claim 8 , wherein the hierarchical cluster structure comprises multiple adaptive layers.

12 . The computer program product according to claim 8 , the stored program instructions further comprising program instructions to add hierarchical cluster layers to the hierarchical cluster structure according to layer self-validation with the target.

13 . The computer program product according to claim 8 , wherein determining spatial-temporal predictor values for the spatial-temporal location includes:
conducting a goal-seek analysis using the hierarchical cluster structure.

14 . The computer program product according to claim 8 , the stored program instructions further comprising:
program instructions to determine spatial-temporal inputs for the target by conducting a goal-seek analysis using the hierarchical cluster structure, wherein the hierarchical cluster structure comprises an adaptive layer; program instructions to add hierarchical cluster layers to the hierarchical cluster structure according to layer self-validation with the target; and program instructions t

Assignees

• IBM

Inventors

• Wang Rui
• Xu Jing James
• Ma Xiao Ming
• Han Si Er
• Gao Lei

Classifications

• G06N3/082

  modifying the architecture, e.g. adding, deleting or silencing nodes or connections · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N3/08 Primary

  Learning methods · CPC title