| Item type |
SIG Technical Reports(1) |
| 公開日 |
2021-02-06 |
| タイトル |
|
|
タイトル |
Keyword extraction method using users' mouse behavior |
| タイトル |
|
|
言語 |
en |
|
タイトル |
Keyword extraction method using users' mouse behavior |
| 言語 |
|
|
言語 |
eng |
| 資源タイプ |
|
|
資源タイプ識別子 |
http://purl.org/coar/resource_type/c_18gh |
|
資源タイプ |
technical report |
| 著者所属 |
|
|
|
University of tsukuba |
| 著者所属 |
|
|
|
University of tsukuba |
| 著者所属(英) |
|
|
|
en |
|
|
University of tsukuba |
| 著者所属(英) |
|
|
|
en |
|
|
University of tsukuba |
| 著者名 |
Chunyang, He
Masao, Takaku
|
| 著者名(英) |
Chunyang, He
Masao, Takaku
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
Owing to the explosive growth of information, keywords play an essential role in summarizing information and helping search effectively. Existing keyword extraction approaches are mostly focused on the document side, instead of using reading side feedback. In this paper, we proposed a keyword extraction method that incorporates the mouse pointer behavior of the reader when browsing academic papers and conducted an experiment to verify the effectiveness of the proposed method. We developed a mouse tracker to record mouse trajectory, speed, and click behaviors when the participants browsed academic papers. Using a predefined weighting algorithm, a term-weighted ranking was created according to mouse features. We used the term frequency-inverse document frequency (TF-IDF) and TextRank methods as the baseline to compare the effectiveness, and evaluation was performed based on precision, recall, and F-score. The experimental results show that the proposed method outperforms the TextRank algorithm, but there are no significant differences between the proposed method and the TF-IDF algorithm. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
Owing to the explosive growth of information, keywords play an essential role in summarizing information and helping search effectively. Existing keyword extraction approaches are mostly focused on the document side, instead of using reading side feedback. In this paper, we proposed a keyword extraction method that incorporates the mouse pointer behavior of the reader when browsing academic papers and conducted an experiment to verify the effectiveness of the proposed method. We developed a mouse tracker to record mouse trajectory, speed, and click behaviors when the participants browsed academic papers. Using a predefined weighting algorithm, a term-weighted ranking was created according to mouse features. We used the term frequency-inverse document frequency (TF-IDF) and TextRank methods as the baseline to compare the effectiveness, and evaluation was performed based on precision, recall, and F-score. The experimental results show that the proposed method outperforms the TextRank algorithm, but there are no significant differences between the proposed method and the TF-IDF algorithm. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AN10114171 |
| 書誌情報 |
研究報告情報基礎とアクセス技術（IFAT）
巻 2021-IFAT-141,
号 2,
p. 1-6,
発行日 2021-02-06
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
2188-8884 |
| Notice |
|
|
|
SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc. |
| 出版者 |
|
|
言語 |
ja |
|
出版者 |
情報処理学会 |
IPSJ-IFAT21141002.pdf (697.4 kB)
