Kiyota Hashimoto

An e-learning system for Japanese onomatopoeic expressions

The Japanese language have a great number of onomatopoeic expressions, which is one of the key characteristics of this language as well as other East-Asian languages like Korean and Indonesian. When a foreigner learns Japanese, thus, it is important to master them, but, due to its quite subjective nature, many feel great difficulty, and it is indeed the case with foreigners who wants to work in Japan. However they are often neglected in rapid Japanese teaching, and a supportive e-learning system for mastering Japanese onomatopoeic expressions is desirable. Based on linguistic reconsideration on Japanese onomatopoeic expressions, we are developing an online e-learning system of Japanese onomatopoeic expressions for foreign workers. Our system aims to present not only explanation of individual onomatopoeic expressions but various contextual information, and to offer a question-answer communication device that enables learners to have a quick, probable answer with relevant examples found in online data and instructors to explain the expression afterwards. Evaluation of our prototype system by foreign Japanese learners indicates that using this kind of e-learning system is suitable for complementing Japanese learning at school.

An Automated Method to Generate e-Learning Quizzes from Online Language Learner Writing

In this paper, the entries of Lang-8, which is a Social Networking Site SNS site for learning and practicing foreign languages, were analyzed and found to contain similar rates of errors for most error categories reported in previous research. These similarly rated errors were then processed using an algorithm to determine corrections suggested by a native speaker. Subject matter experts then evaluated the processed sentences to determine the quality in relation to use in tests or exams for language learners. The method describes the automatic generation of multiple choice and fill-in-the-blanks quizzes using the writings of language learners on public web based learning sites, in order to support learner reflection on corrections and practicing past errors to overcome problems.

An Investigation of Effectiveness Using Topic Information Order to Classify Tourists Reviews

The Sentiment analysis targeted customer reviews is an active research field in recent years. Customer reviews contain various topics that closely related to interest of reviewer and situation of utilization. Using this relation, it will be possible to show readers information that fits in interest of them. We collected customer reviews from a tourism information web site, estimate topics in the customer reviews by applying LDA and estimate that customer reviews was whether positive or negative.

Research trends with cross tabulation search engine

To help researchers in building a knowledge foundation of their research fields which could be a time-consuming process, the authors have developed a Cross Tabulation Search Engine (CTSE). Its purp ...

Mutimedia learner corpus of foreigner's basic presentation in English with evaluations

Presentation skill is essential both for academic and business purposes, but its training, particularly in Asian countries, still heavily depends on experiences. This paper reconsiders presentation skill and reports our multimedia learner corpus of basic presentation in English made by Japanese students. It consists of video, slides, prepared oral manuscripts, and holistic and element-based evaluations. With this evaluations, we discussed the correspondence between the holistic and element-based evaluations with a view to realizing semi-automatic evaluation of presentations.

Extraction of Hints and Advice from Hotel Reviews for Improving Small Hotel Management

There are various kinds of and huge amounts of hotel information both from providers and customers. Hotel information by hotels and travel agents are reliable. A much large number of reviews by general users are available, which might be less reliable compared to the official ones. However, those reviews are helpful, since we can hear their personal experience and opinion. This paper proposes a method to find hints and advice for improving small hotel management by extracting and analyzing the feature words of reviews. This paper focuses on the secondary major words and compares their occurrence probabilities in the business customers review with the family customers review. A visual interpretation is proposed by mapping the feature words on Word Net.

Prototypical Design of Learner Support Materials Based on the Analysis of Non-verbal Elements in Presentation

There is a growing need for a well-designed learner support system for presentation in English, particularly in non-English-speaking countries. We have developed a prototype of comprehensive learning support system of basic presentation that consists of several modules including digital contents of preliminary tutorials, an interactive aide for organizing a presentation and its corresponding slides, semi-automatic evaluation estimation, and a online review of recorded presentations. After trials, it has been clear that non-verbal aspects has to be extensively supported by such a system. In this study, we made an extensive observation and analysis of available professional and learner presentations, and extracted significant non-verbal elements in those presentations. Then, we designed learner support materials for non-verbal aspects, based on a non-verbal ontology we also designed. It is expected that the implementation of those materials will let learners to learn more effectively how to make and conduct a presentation.

A corpora-based detection of stylistic inconsistencies of text in the targeted subgenre

Any text, be it written by humans or generated by computers, must choose a specific style and be consistent. Textual style is based on the genre to which the text should belong. Each genre has its unique characteristics based on the style of narration, the topic, the target audience, and the author’s intentions. Thus, in order to write or produce better passages, a finer-grained style checker sensitive to differences among genres is to be developed. We propose a foundational method for this purpose: the flexible accumulation of data of the style features of atomic expressions based on any kind of contrastive sets of texts representing the good and bad examples for the targeted genre, an analysis of a given text using the style features, and a visualization to effectively help the author detect anomalies for the targeted genre. Any type of genre-sensitive style checker will implement our method or similar ones.

Feature Words of Moves in Scientific Abstracts

Extraction of structure from texts is a key issue of text mining. The rhetorical structure of move in scientific articles is useful for assisting in the reading and writing. In this paper, we classify move structure in the abstract of research articles with a small number of characteristic words that determine five moves of including background (B), purpose(P), method(M), result(R) and discussion(D). Eleven measures were introduced and used to select features of moves. Exhaustive parameter search were conducted to get the optimal combination of measure and the number of features. We applied support vector machine and evaluated 10 fold cross validations. The accuracies with optimal feature selections are 0.9022, 0.8322, 0.8442, 0.8820 and 0.8354 for B, P, M, R and D, respectively. They are 10% better than the baseline performance that use all keywords. This study surprisedly found that the negative feature words play central role for prediction performance improvement.

A Method to Classify Customer Reviews of Japanese Hotels by Support Vector Machine Using Estimation Sentence Patterns Information

Customer reviews are important sources of opinions of customers. The reviews contain some sentence pattern and some sort of patterns will be useful to extract the opinions of the customers easily. We propose a method to classify customer reviews using estimation expression patterns by support vector machine. We confirm an effectiveness of the proposed method by experimental results.