AI-based machine learning of knowledge and intuition possessed by bookstore staff to support in attracting customers

Nippon Shuppan Hanbai, Inc.,Fujitsu Limited

Tokyo, May 14, 2018

Nippon Shuppan Hanbai, Inc. (Nippan) and Fujitsu Limited announce the joint development of "SeleBoo" book selection service, which uses artificial intelligence to automatically select books that fit with the unique characteristics of a bookstore, such as the store concept and target customers. Nippan will make the service available to its affiliated bookstores across Japan, beginning in the summer of 2018.

SeleBoo, a word combination derived from "Select Book," is the industry's first AI-based book selection service to create book lists suited to a store's theme and customer base, from approximately 600,000 books on sale in Japan, through big data analysis using Fujitsu's AI. Based on these lists, Nippan will suggest books to its affiliated bookstores across Japan, helping retail locations attract customers through in-store promotions and events.

Through this service, Nippan and Fujitsu aim to invigorate the publishing industry market, and create new business models by matching books with products and services in other industries, such as cafes and general merchandise stores.

Background

The publishing industry is facing the pressure of dramatic change as reading rates fall among young people and ebooks become increasingly common. In light of this, Nippan, which connects publishers with bookstores, has joined with Fujitsu to develop SeleBoo, an AI book selection service, based on Nippan's data, including product information, sales results, and purchase records, in order to support a new strategy of bringing out the distinctiveness of bookstores.

Details of the SeleBoo Service

This service uses the Marketing AI Container^[1] , which is a big data analysis service using Fujitsu's AI, to analyze open data such as DBpedia^[2] and Lod4all^[3] in addition to Nippan's 3.5 million bibliographic items of information on books and magazines as well as sales information, from about 3,000 bookstores across Japan. The service then selects books to create a list based on the bookstore or retail store's theme and target customer base. In addition, by incorporating feedback from bookstore staff on the selection results, AI can apply machine learning to bookstore staff knowledge and intuition, improving its book selection capability.

The service will initially offer keyword book selection, which will select books related to a theme, key book selection, which will select books that are similar to a specified book, and location book selection, which will select books related to a specific area. Fujitsu and Nippan plan to further enhance the service going forward, developing functionality such as bookstore color book selection, which will select books that suit the characteristics of a bookstore, and cover book selection, which will select books based on their covers.

Diagram: The structure of SeleBoo

SeleBoo will be exhibited at the Nippan Partner Event^[4] on May 15, and Fujitsu Forum 2018^[5] , which will be held May 17-18.

• [1] Marketing AI Container

An analysis environment that can execute as-is the advanced analysis logic developed by Fujitsu on the basis of over 200 analysis projects aimed at specific data usage goals for each individual company. The container can provide an analysis environment in a flexible form that fits the customer’s needs.

• [2] DBpedia

A community site that extracts information from Wikipedia and publishes it as Linked Open Data (LOD).

• [3] Lod4all

An LOD usage platform that aggregates LOD published around the world, and enables unified searches.

• [4] Nippan Partner Event

An annual meeting, held for the purpose of deepening the participants’ understanding of Nippan’s management plans and policies, while building strong partnerships with affiliated bookstores and publishers.

• [5] Fujitsu Forum 2018

Fujitsu's largest event in Japan. The theme for this year is "Human Centric Innovation: Co-creation for Success." Held at the Tokyo International Forum (Chiyoda-ku, Tokyo).