User generated reviews, like those found on Yelp and Amazon, have become important reference material in casual decision making, like dining, shopping and entertainment. However, very large amounts of reviews make the review reading process time consuming. A text visualization can speed up the review reading process. In this thesis, we present the clustered layout word cloud -- a text visualization that quickens decision making based on user generated reviews. We used a natural language processing approach, called grammatical dependency parsing, to analyze user generated review content and create a semantic graph. A force-directed graph layout was applied to the graph to create the clustered layout word cloud. We conducted a two-task user study to compare the clustered layout word cloud to two alternative review reading techniques: random layout word cloud and normal block-text reviews. The results showed that the clustered layout word cloud offers faster task completion time and better user satisfaction than the other two alternative review reading techniques. [Permission email from J. Huang removed at his request. GMc March 11, 2014] / Master of Science

Online discussion forums and reflective writing are proven methods for enriching conceptual understanding and are hallmarks of engineering education. Plagiarism and many students’ apprehension to contribute to online journals can plague the effectiveness of these educational tools. Using elements of the engineering design cycle, I have created a blogging website that addresses these problems by restricting comment visibilities for users and includes a graphic visualization called a “word cloud” to supplement discussion. A prototype was tested with UTeach Engineering teachers for feedback on design and use. The critiques provided examples of classroom use, constructive design feedback, and ideas for its use as a formative assessment. The design could be used as a pedagogical tool for an investigation of formative assessments in engineering education, but further research for “word cloud” visualizations and journal data collection is needed to expand the current design. / text

This paper describes some current methods for customer's opinion mining and text processing. Application architecture is created along with database structure. Based on this architecture, new application is implemented. Emphasis is placed on automated downloading and processing of customers' reviews with opinion extraction and user-friendly presentation. Part of this paper is dedicated to optimizing the application for better efficiency and speed boost.

No description available.

No description available.

各種行動裝置服務不僅便利了生活，也讓資訊的傳遞更加即時且無界。2012年起，五家媒體藉助行動通訊軟體LINE在台灣的高人氣，每日固定推播四則新聞，總訂戶數至今突破450萬人，儼然成為一個新興的新聞傳播渠道。 本研究以情境理論的觀點，質化的使用者日誌記錄法和深度訪談法為主，理解使用者在不同情境下使用推播新聞的行為，並透過HTML5文字雲分析機，歸納出各家媒體推播內容的樣貌和趨勢。情境分析構面包含接收與閱讀情境下的「實體環境」、「社交環境」、「資訊環境」、「先前狀態」和「時間的影響」等；另再針對使用過程中面臨的「人機互動情境」綜合分析使用者的體驗。 這些情境因素不僅各自影響媒介使用的動機和行為，也會與新聞內容相互加強（或制約）吸引力，端賴哪方拉力（或推力）較強。根據訪談結果，「新聞內容」的影響力似乎又更為關鍵。受訪者普遍認為推播新聞的即時性和重要性偏低，新聞價值未能滿足期待，導致對LINE推播新聞的需要性和依賴性偏低。 未來業者可考慮設立不同的「主題式官方帳號」，提供針對性的內容；或者在目前的官方帳號中，增加一日新聞類型的多元性以滿足更多人的需求，就看媒體如何定位自家官方帳號的角色。此外，在發揮新聞標題創意的同時，仍應與內文保持一致，避免造成期待上的落差。不斷強化內容品質，並考量主要受眾的使用情境，可望改善體驗不佳的問題。 本研究以情境理論的觀點，開啟了「推播新聞」的學術研究基礎，並充實行動新聞領域的實證研究經驗，希望提供未來的研究者和新聞從業人員，一個了解推播新聞使用行為的先例。 / Mobile devices and services facilitate our lives, and also assist in delivering information more freely and immediately. Since 2012, five companies send four pieces of news daily through “LINE”, which is the most popular mobile instant messaging application in Taiwan. Nowadays, LINE becomes an emerging news channel, with over 450 million subscribers in total. This study is mainly based on the context theory, by means of diary method and in-depth interviews, to understand users’ behavior of push notifications news in different contexts, from news receiving aspect, reading aspect and human-computer interaction aspect. Besides, analyzing the news headlines by “HTML5 Text Analyzer” to explore the features of different media. Research found that users’ behavior and attitude influenced not only by contextual factors, but also by the value of content. Moreover, the effect of contextual factors and content would be reinforcing (or restraining) mutually, depends which one influence users’ behavior and attitude more crucially. According to respondents’ feedback, the value of push notification news, such as immediacy and importance, were far from their expectations. News editors may plan to set up more than one official account to reach target audience; or to increase the diversity of daily news by existing accounts, making effort to meet demand of more audience. In addition, the headline should be more concordant with the content of the news. Keep improving the quality of content, taking account of context factors while pushing news, may improve users’ experience.