A Method of Deducing a User's State of Mind from an Analysis of the Pictographic Characters Used in Mobile Phone Emails

As the ubiquitous environment is taking root, there are calls for services that deliver content appropriate for the individual user's personal interests and preferences. In fact, services are already being provided that deduce the interests or preferences of the user by analyzing his or her log data containing behavior history, the text of his or her emails, access history, etc., and provide content appropriate for his or her physical/mental/emotional state. However, it is difficult to deduce the ever-changing preferences of people who live in a complicated society. Such deduction is still at a research stage. In this paper, we focus on mobile phones, whose users are growing in number and which offer many sophisticated functions besides the ability to talk. We propose a method of deducing the state of mind of the user by analyzing the pictographic characters used in his or her emails. Pictographic characters are used to convey feelings that the user thinks are too complicated to express in text. Specifically, from among many pictographic characters, we have selected those that show facial expressions and are used to express feelings (states of mind). We have extracted the state-of-mind elements that are associated with each pictographic character, and assigned state-of-mind elements and corresponding vector values to each pictographic character. We have developed an algorithm to deduce the user's state of mind from an email by analyzing the pictographic characters used in that email. We have developed a prototype evaluation system, and evaluated the effectiveness of the proposed algorithm.