Yuki Urabe

Bio: Yuki Urabe is an academic researcher from Hokkaido University. The author has contributed to research in topics: Emoticon & Visualization. The author has an hindex of 4, co-authored 17 publications receiving 46 citations.

Emoticon Recommendation for Japanese Computer-Mediated Communication

Abstract: This paper describes the development of an emoticon recommendation system based on emoticons numerically categorized by emotion. The emoticon recommendation system aims to help users express their feelings in computer-mediated communication by recommending emoticons appropriate to user input. In order to develop this system, the original emoticon database, a table of emoticons with the points expressed from each of 10 distinctive emotions, was developed. An evaluation experiment showed that 71.3% of user-selected emoticons were among the top 10 emoticons recommended by the proposed system. Moreover, we compared the proposed system to the current system used in iPhone by adopting a semantic differential (SD) scale of 1-7. The results showed that the proposed system scored higher than the current system by 1.05 points in ease of choice, 0.55 points in accuracy, and 0.55 points in specificity. We plan to make our proposed method open source, so that any developer can build in their own interfaces and enhance their own input methods using these emoticon recommendation systems.

Emoticon recommendation system for effective communication

Abstract: The existence of social media has made computer-mediated communication more widespread among users around the world. This paper describes the development of an emoticon recommendation system that allows users to express their feelings with their input. In order to develop this system, an innovative emoticon database consisting of a table of emoticons with points expressed from each of 10 distinctive emotions was constructed. An evaluation experiment showed that 71.3% of user-selected emoticons were among the top 10 emoticons recommended by the proposed system.

Emoticon Recommendation System to Richen Your Online Communication

Abstract: Japanese emoticons are widely used to express users' feelings and intentions in social media, blogs and instant messages. Japanese smartphone keypads have a feature that shows a list of emoticons, enabling users to insert emoticons simply by touching them. However, this list of emoticons contains more than 200, which is difficult to choose from, so a method to reorder the list and recommend appropriate emoticons to users is necessary. This paper proposes an emoticon recommendation method based on the emotive statements of users and their past selections of emoticons. The system is comprised of an affect analysis system and an original emoticon database: a table of 59 emoticons numerically categorized by 10 emotion types. The authors' experiments showed that 73.0% of chosen emoticons were among the top five recommended by the system, which is an improvement of 43.5% over the method used in current smartphones, which is based only on users' past emoticon selections.

Find right countenance for your input—Improving automatic emoticon recommendation system with distributed representations

Abstract: Emoticons are popularly used to express user’s feelings in social media, blogs, and instant messaging. However, the number of emoticons existing in emoticon dictionaries which users select from is large, thus, it is difficult for users to find the desired emoticon that matches the content of their messages. In this paper, we propose a method that supports users’ emoticon selection by reordering 167 unique emoticons in the emoticon dictionary by applying pre-trained models learned from large data in Japanese. We evaluated whether adapting a pre-trained model to our emoticon recommendation system achieves better results than just learning surface patterns of text and emoticon. We collected sets of sentences and emoticons in Japanese from the Internet and pre-trained models (i.e. Word2vec, ELMo, and BERT) that learned from large Japanese textual data and used deep learning techniques such as BiLSTM and fine-tuning for learning. We confirmed that fine-tuning our data with BERT achieved the best recommendation accuracy of 52.98%, recommending the correct emoticon within the top 25 (top 15%) of the emoticons. Moreover, we confirmed our intuition that widely used Wikipedia-based pre-trained models are not the best voice for the facemark recommendations.

Comparison of Emoticon Recommendation Methods to Improve Computer-Mediated Communication

Abstract: This paper describes the development of an emoticon recommendation system based on users’ emotional statements. In order to develop this system, an innovative emoticon database consisting of a table of emoticons with points expressed from each of 10 distinctive emotions was created. An evaluation experiment showed that our proposed system achieved an improvement of 28.1 points over a baseline system, which recommends emoticons based on users’ past emoticon selection. We also integrated the proposed and baseline systems, leading to a performance improvement of approximately 73.0 % in the same experiment. Evaluation of respondents’ perceptions of the three systems utilizing an SD scale and factor analysis is also described in this paper.

Oh that's what you meant!: reducing emoji misunderstanding

Abstract: Emoji provide a way to express nonverbal conversational cues in computer-mediated communication. However, people need to share the same understanding of what each emoji symbolises, otherwise communication can breakdown. We surveyed 436 people about their use of emoji and ran an interactive study using a two-dimensional emotion space to investigate (1) the variation in people's interpretation of emoji and (2) their interpretation of corresponding Android and iOS emoji. Our results show variations between people's ratings within and across platforms. We outline our solution to reduce misunderstandings that arise from different interpretations of emoji.

Beyond Just Text: Semantic Emoji Similarity Modeling to Support Expressive Communication

Abstract: Emoji, a set of pictographic Unicode characters, have seen strong uptake over the last couple of years. All common mobile platforms and many desktop systems now support emoji entry, and users have embraced their use. Yet, we currently know very little about what makes for good emoji entry. While soft keyboards for text entry are well optimized, based on language and touch models, no such information exists to guide the design of emoji keyboards. In this article, we investigate of the problem of emoji entry, starting with a study of the current state of the emoji keyboard implementation in Android. To enable moving forward to novel emoji keyboard designs, we then explore a model for emoji similarity that is able to inform such designs. This semantic model is based on data from 21 million collected tweets containing emoji. We compare this model against a solely description-based model of emoji in a crowdsourced study. Our model shows good perfor mance in capturing detailed relationships between emoji.

Automatic Construction of an Emoji Sentiment Lexicon

Abstract: Emojis have been frequently used to express users' sentiments, emotions, and feelings in text-based communication. To facilitate sentiment analysis of users' posts, an emoji sentiment lexicon with positive, neutral, and negative scores has been recently constructed using manually labeled tweets. However, the number of emojis listed in the lexicon is smaller than that of currently existing emojis, and expanding the lexicon manually requires time and effort to reconstruct the labeled dataset. This paper presents a simple and efficient method for automatically constructing an emoji sentiment lexicon with arbitrary sentiment categories. The proposed method extracts sentiment words from WordNet-Affect and calculates the cooccurrence frequency between the sentiment words and each emoji. Based on the ratio of the number of occurrences of each emoji among the sentiment categories, each emoji is assigned a multidimensional vector whose elements indicate the strength of the corresponding sentiment. In experiments conducted on a collection of tweets, we show a high correlation between the conventional lexicon and our lexicon for three sentiment categories. We also show the results for a new lexicon constructed with additional sentiment categories.

Role of Emoticons for Multidimensional Sentiment Analysis of Twitter

Abstract: Microblogging systems such as Twitter and Facebook have become popular. People can easily post their sentiments to the Internet in real time using such microblogging systems. Twitter is a text-based communication tool. Users cannot use non-verbal communication tools such as gestures and eye contact on Twitter. Users sometimes use emoticons as an alternative non-verbal communication tool to tweet delicate sentiments. In this paper, we propose a method of determining sentiment of a tweet based on the emoticon role. Specifically, we propose the following: (1) compilation of a sentiment lexicon and an emoticon lexicon; (2) emoticon roles can be classified into four types showing "Emphasis", "Assuagement", "Conversion", and "Addition", with roles determined based on a relation between sentiments of sentences and emoticons; and (3) the relation can be formalized using regression analysis in all roles excepting for "Addition".