Hirotoshi Honma

Erratum and addendum: Erratum and addendum to A linear time algorithm for finding all hinge vertices of a permutation graph [Information Processing Letters 59 (2) (1996) 103-107]

This note points out and corrects an error in the algorithm proposed in [Ting-Yem Ho, Yue-Li Wang and Ming-Tsan Juan, A linear time algorithm for finding all hinge vertices of a permutation graph, Information Processing Letters 59 (2) (1996) 103-107].

Proposal of recommender system based on user evaluation and cosmetic ingredients

Recent years have witnessed web services drastically becoming popular in our daily lives, and many consumers take user reviews of products into account when planning purchases. The number of cosmetic review sites, users, and products posted have been increasing year by year. For example, when a user searches for skin lotions using the @cosme website, she consults with reviews of users with similar attributes to her own (age, skin quality, etc.) and searches for items that are compatible with the skin lotion that she uses on a daily basis. However, since different basic cosmetics may have different effects, it is difficult to find products that are compatible with a user, even using the review information from @cosme. In this study, we assume that the compatibility between a user and a basic cosmetic product depends on its cosmetic ingredients. Combining review information from Bihada-Mania website with that from @cosme, we extracted the effective cosmetic ingredients for each user attribute and developed a recommender system of basic cosmetics based on ingredients. We applied the IF-IPF method which applied the concept of TF-IDF method to extraction of effective ingredients of cosmetics. We have defined the scale “invalidated product number” to evaluate the effectiveness of our recommendation service. From the results of the two experiments, the invalidated product number is less than 5% for all user attributes. This indicates that our recommender system has certain reliability.

Automatic extraction of future references from news using morphosemantic patterns with application to future trend prediction

In everyday life people use past events and their own knowledge to predict future events. In such everyday predictions people use widely available resources (newspapers, Internet). This study focused on sentences referring to the future, such as the one below, as one of such resource.