Computer system and method for predicting intervention effect

The computer system uses machine learning models with GANs to adjust for confounding and selection biases, enabling accurate prediction of intervention effects across multiple scenarios.

JP7590280B2Active Publication Date: 2024-11-26HITACHI LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
JP2021105786
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2021-06-25
Publication Date
2024-11-26
Estimated Expiration
2041-06-25

Smart Images

  • Figure 0007590280000005
    Figure 0007590280000005
  • Figure 0007590280000006
    Figure 0007590280000006
  • Figure 0007590280000007
    Figure 0007590280000007
Patent Text Reader

Abstract

To predict the effects of several interventions on a person.SOLUTION: A computer system manages: a first model that generates a feature by mapping a vector generated by machine learning and composed of values of multiple factors representing a state of a person onto a feature space; and a second model that outputs predictive values of the effect of multiple interventions on a person from the feature. The first model maps multiple pieces of learning data used in machine learning onto the feature space so that differences in the distribution of the multiple pieces of learning data in the feature space are reduced. The computer system accepts input data including values of the multiple factors, inputs the input data to the first model to generate a feature of the input data, and inputs the feature of the input data to the second model to calculate a predictive value of the effect of multiple interventions.SELECTED DRAWING: Figure 4
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Interposition effect estimation system, interposition effect estimation method, and program used in interposition effect estimation system

    JP2018005707A

  • Causal relationship for machine learning system

    JP2019194849A

  • Utilizing a deep generative model with task embedding for personalized targeting of digital content through multiple channels across client devices

    US20200245009A1