Rakuten Group has filed a patent for an information processing apparatus that predicts disease risk for a target user based on social relationships and user features. The apparatus acquires factual features of multiple users, creates a relationship graph based on these features, sets a target user, and predicts the risk of developing a disease for that user. GlobalData’s report on Rakuten Group gives a 360-degree view of the company including its patenting strategy. Buy the report here.
According to GlobalData’s company profile on Rakuten Group, dynamic premium pricing was a key innovation area identified from patents. Rakuten Group's grant share as of September 2023 was 51%. Grant share is based on the ratio of number of grants to total number of patents.
Predicting disease risk based on social relationships and user features
A recently filed patent (Publication Number: US20230317294A1) describes an information processing apparatus that can predict disease risk for a target user based on social relationships and user features. The apparatus includes an acquisition unit that collects factual features of multiple users, a creation unit that generates a relationship graph indicating social connections between users, a target user setting unit, and a prediction unit.
The prediction unit utilizes the relationship graph and the user features of the target user to forecast the risk of developing at least one disease. This prediction is made using a machine learning model that takes the user feature of the target user as input and outputs the disease risk.
To train the machine learning model, the apparatus includes a training unit that uses disease features obtained from the relationship graph. These disease features include information indicating whether each user has contracted a specific disease.
The creation unit constructs the relationship graph by representing each user as a node and connecting nodes with links based on their factual features. Users with the same factual features are connected with explicit links, while users without explicit links are connected with implicit links based on other connected pairs of user nodes. The creation unit also determines the closeness of connected pairs based on shared factual features.
The disease risk is represented by a numeric value ranging from 0 to 1 for each disease, with 1 indicating the highest likelihood.
In addition to the information processing apparatus, the patent also describes an information processing method and a non-transitory computer-readable medium storing a computer program for executing the processing steps. These include acquiring user features, creating a relationship graph, setting a target user, and predicting disease risk based on the relationship graph and the user feature of the target user.
Overall, this patent presents a system that utilizes social relationships and user features to predict disease risk for individuals. By leveraging machine learning and a relationship graph, the apparatus aims to provide valuable insights into potential health risks.