Yukiko Kawai

Detecting age of page content

Web pages often contain objects created at different times. The information about the age of such objects may provide useful context for understanding page content and may serve many potential uses. In this paper, we describe a novel concept for detecting approximate creation dates of content elements in Web pages. Our approach is based on dynamically reconstructing page histories using data extracted from external sources - Web archives and efficiently searching inside them to detect insertion dates of content elements. We discuss various issues involving the proposed approach and demonstrate the example of an application that enhances browsing the Web by inserting annotations with temporal metadata into page content on user request.

A Novel Visualization Method for Distinction of Web News Sentiment

Recently, an increasing number of news websites have come to provide various featured services. However, effective analysis and presentation for distinction of viewpoints among different news sources are limited. We focus on the sentiment aspect of news reporters viewpoints and propose a system called the Sentiment Map for distinguishing the sentiment of news articles and visualizing it on a geographical map based on map zoom control. The proposed system provides more detailed sentiments than conventional sentiment analysis which only considers positive and negative emotions. When a user enters one or more query keywords, the sentiment map not only retrieves news articles related to the concerned topic, but also summarizes sentiment tendencies of Web news based on specific geographical scales. Sentiments can be automatically aggregated at different levels corresponding to the change of map scales. Furthermore, we take into account the aspect of time, and show the variation in sentiment over time. Experimental evaluations conducted by a total of 100 individuals show the sentiment extraction accuracy and the visualization effect of the proposed system are good.

What can history tell us?: towards different models of interaction with document histories

The current Web is a dynamic collection where little effort is made to version pages or to enable users to access historical data. As a consequence, they generally do not have sufficient temporal support when browsing the Web. However, we think that there are many benefits to be obtained from integrating documents with their histories. For example, a documents history can enable us to travel back through time to establish its trustworthiness. This paper discusses the possible types of interactions that users could have with document histories and it presents several examples of systems that we have implemented for utilizing this historical data. To support our view, we present the results of an online survey conducted with the objective of investigating user needs for temporal support on the Web. Although the results indicated quite low use of Web archives by users, they simultaneously emphasized their considerable interest in page histories.

Fair news reader: recommending news articles with different sentiments based on user preference

We have developed a news portal site called Fair News Reader (FNR) that recommends news articles with different sentiments for a user in each of the topics in which the user is interested. FNR can detect various sentiments of news articles, and determine the sentimetal preferences of a user based on the sentiments of previously read articles by the user. While there are many news portal sites on the Web, such as GoogleNews, Yahoo!, and MSN News, they can not recommend and present news articles based on the sentiments they are likely to create since they simply select articles based on whether they contain user-specified keywords. FNR collects and recommends news articles based on the topics in which the user is interested and the sentiments the articles are likely to create. Eight of the sentiments each article is likely to create are represented by an article vector with four elements. Each element corresponds to a measure consisting of two symmetrical sentiments. The sentiments of the articles previously read with respect to a topic are then extracted and represented as a user vector. Finally, based on a comparison between the user and article vectors in each topic, FNR recommends articles that have symmetric sentiments against the sentiments of read articles by the user for fair reading about the topic. Evaluation of FNR using two experiments showed that the user vectors can be determined by FNR based on the sentiments of the read articles about a topic and that it can provide a unique interface with categories containing the recommended articles.

Visualizing historical content of web pages

Recently, along with the rapid growth of the Web, the preservation efforts have also increased. As a consequence, large amounts of past Web data are stored in Web archives. This historical data can be used for better understanding of long-term page topics and characteristics. In this paper, we propose an interactive visualization system called Page History Explorer for exploring page histories. It allows for roughly portraying evolution of pages and summarizing their content over time. We use a temporal term cloud as a structure for visualizing prevailing and active terms appearing on pages in the past.

Using a sentiment map for visualizing credibility of news sites on the web

We have developed a visualizing news system that shows the trend of the news site on the Web for credibility. If users know the trend of the news site, users can evaluate the credibility of each news topic. This system detects and uses sentiments of each news article to resolve the trend of Web site. The trend of Web sites are extracted as average sentiments of the news articles that were written concerning a topic in each Web site. The sentiments of news articles are represented by four values calculated in four sentiment scales: Bright ⇔ Dark, Acceptance ⇔ Rejection, Relaxation ⇔ Strain, and Anger ⇔ Fear. The sentiment values of news articles are calculated using the sentiment dictionary that was constructed by our previously proposed method. If a user enters one or more topic keywords, our proposed system extracts the news articles that include the keywords from each predetermined news sites. Our system also calculates the sentiment values of the news articles and their average value in each sentiment from each news Web site. The system then generates a bar graph from the four average values in each news Web site and attaches all the bar graphs on the world map using Google Map API. We call the map a sentiment map in this paper. The sentiment map helps users intuitively and efficiently understand trends among multiple Web sites concerning a given topic. In this paper, we describe how to create a sentiment map and explain how we evaluated our proposed system through several experiments.

A browser for browsing the past web

We describe a browser for the past web. It can retrieve data from multiple past web resources and features a passive browsing style based on change detection and presentation. The browser shows past pages one by one along a time line. The parts that were changed between consecutive page versions are animated to reflect their deletion or insertion, thereby drawing the users attention to them. The browser enables automatic skipping of changeless periods and filtered browsing based on user specified query.

Sentiment Bias Detection in Support of News Credibility Judgment

Recently, an increasing number of online news websites have come to provide news browsing and retrieval services. For certain topics, certain news websites may hold sentiment bias, and therefore select and edit information according to their own standpoints before delivering news articles. Lacking conscious awareness of websites sentiment bias may result in blind obedience to the reported information. We focus on the sentiment aspect of news articles and develop a system which can detect and visualize sentiment tendencies of different websites. Given a topic, the system extracts relevant subtopics and presents sentiment difference between different subtopics. Once a subtopic is specified, sentiment difference between news websites is also provided. The background knowledge of sentiment difference between subtopics and between websites can assist users in judging the news credibility. In particular, the system analyzes four-dimension sentiment, which is more similar to human emotion than conventional positive-negative sentiment. Experimental evaluations show the accuracy of sentiment extraction and subtopic extraction is good, and our observation results show sentiment bias can be detected by the system.

My portal viewer: integration system based on user preferences for news web sites

We developed a novel web application called “My Portal Viewer (MPV)”, which automatically categorizes and integrates meta-data from many news pages based on the users preferences after gathering these news pages from various news sites. Our unique approach is based on two points: one is an automatic categorization of collected information based on users interests and knowledge, and the other is the look and feel of the MPV page, which is applied to the users favorite news portal page, and part of the original content is replaced by the integrated content. Whenever a user accesses the MPV page after browsing news pages, he/she can obtain the desired content efficiently because the MPV presents pages refreshed based on the users behavior through his/her favorite page layout, which reflects his/her interests and knowledge. In this paper, we describe the MPV framework, and methods that are based on the users preferences for replacing and categorizing content have been developed using an HTML table model and a vector matching model.

A3: framework for user adaptation using xslt

We propose a system called Adaptation Anywhere & Anytime(A3), which is a framework for making web sites/applications adaptable to users needs or interests, and we describe the implement of a web site on A3 by using XSLT. Web sites/applications built on A3 construct user ontologies for each user automatically and share them between sites/applications. Each site/application uses the user ontology to select an appropriate resource for the user and to present such resources in a suitable form. And A3 offers the method for constructing the adaptable web sites using XSLT. The author of web sites can easily make their sites adaptable by using XSLT.