Rakuten Group. has filed a patent for a method using a one-class recommendation model to predict user-item interaction values. The method trains the model using only similar user-item pairs, incorporating loss terms to prevent a collapsed solution. This innovation aims to improve recommender systems’ accuracy and efficiency. GlobalData’s report on Rakuten Group gives a 360-degree view of the company including its patenting strategy. Buy the report here.
Smarter leaders trust GlobalData
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 January 2024 was 51%. Grant share is based on the ratio of number of grants to total number of patents.
One-class recommendation model for predicting user-item interactions
A recently filed patent (Publication Number: US20240037191A1) outlines a method for predicting user-item interaction values on an ecommerce platform using a one-class recommendation model. The method involves training the model with a training dataset consisting of only similar user-item pairs with known interactions, excluding dissimilar pairs. The model calculates a loss value using various loss terms, including an attractive loss term to minimize distance in the vector space, a pairwise distance loss term to maintain average distances, and an orthogonality loss term to reduce correlations between dimensions. The trained model is then used to predict interaction values for user-item pairs where no value is known, enabling recommendations for users or items on the platform.
Furthermore, the patent describes a system for predicting user-item interactions on an ecommerce platform, incorporating a one-class machine-learning recommendation model. The system includes a processor, memory units, and instructions for generating user-interaction scores based on the trained model. The method for training the model involves obtaining a training dataset of similar user-item pairs, applying the model to generate predicted interaction values, calculating loss using specific terms, adjusting parameters to minimize loss, and iterating through the process. This system allows for accurate predictions of user-item interactions, facilitating recommendations for users to shops or items on the platform based on predicted values.
To know more about GlobalData’s detailed insights on Rakuten Group, buy the report here.
Data Insights
From
The gold standard of business intelligence.
Blending expert knowledge with cutting-edge technology, GlobalData’s unrivalled proprietary data will enable you to decode what’s happening in your market. You can make better informed decisions and gain a future-proof advantage over your competitors.
GlobalData, the leading provider of industry intelligence, provided the underlying data, research, and analysis used to produce this article.
GlobalData Patent Analytics tracks bibliographic data, legal events data, point in time patent ownerships, and backward and forward citations from global patenting offices. Textual analysis and official patent classifications are used to group patents into key thematic areas and link them to specific companies across the world’s largest industries.
