Additional Funding Sources
The project described was supported by the Pacific Northwest Louis Stokes Alliance for Minority Participation through the National Science Foundation under Award No. HRD-1410465.
Abstract
Reading habits of young children directly influence their academic success and maturing process as members of society. Although its immense importance, finding the right book for children is often challenging. On top of the overwhelming amount of books available that vary in levels of difficulty to read and understand, each reader is different and desires something unique from their book-reading experience. To identify an appealing book, parents, educators, and especially younger readers themselves can turn towards recommenders (RS) as their primary goal is to suggest items based on users’ interests. One caveat of traditional/existing RS is that they rely heavily on historical data (e.g., ratings, likes, previous purchases, and other users' data). This type of data is rarely available for a protected user group, such as young children. In this paper, we introduce BRITE, Book Recommendations for Emergent Readers, a simple, yet effective, RS that bypasses the need for historical data by eliciting from young readers their preferences about a main character and book setting they wish to read about through image selection while directly considering book traits often preferred by users in the readers’ age group. We aim for book recommendations produced by BRITE to be more tailored to young readers to motivate them to read and foster good reading habits while remaining mindful of their personal information.
Tales of a Young Reader: Personalizing Book Recommendations for Children Based on Visual Preference Elicitation and Age-Favored Book Traits
Reading habits of young children directly influence their academic success and maturing process as members of society. Although its immense importance, finding the right book for children is often challenging. On top of the overwhelming amount of books available that vary in levels of difficulty to read and understand, each reader is different and desires something unique from their book-reading experience. To identify an appealing book, parents, educators, and especially younger readers themselves can turn towards recommenders (RS) as their primary goal is to suggest items based on users’ interests. One caveat of traditional/existing RS is that they rely heavily on historical data (e.g., ratings, likes, previous purchases, and other users' data). This type of data is rarely available for a protected user group, such as young children. In this paper, we introduce BRITE, Book Recommendations for Emergent Readers, a simple, yet effective, RS that bypasses the need for historical data by eliciting from young readers their preferences about a main character and book setting they wish to read about through image selection while directly considering book traits often preferred by users in the readers’ age group. We aim for book recommendations produced by BRITE to be more tailored to young readers to motivate them to read and foster good reading habits while remaining mindful of their personal information.