2014 Future University-Hakodate 2014 System Information Science Practice Group Report Project Name Support for decision making by data analysis B Group Name Group (B) /Project No. 18-B /Project Leader 1012154 Tomoki Sugisawa /Group Leader 1012106 Tsubasa Ikawa /Group Member 1011213 Kengo Iwahashi 1012047 Keisuke Tanaka 1012062 Rintaro Yamada 1012106 Tsubasa Ikawa 1012131 Shoto Sukekawa 1012135 Kensuke Tanaka 1012156 Hiroki Nakashima 1012173 Misa Kobayashi Advisor Yasuhiro Katagiri Takashi Takenouchi Kiyohito Nagano 2014 1 14 Date of Submission January 14, 2015
B 2 2 4 B 1 R MeCab 4 2 4 HTML,, R,, : - i -
Abstract Abstract in English. The purpose is to develop course recommendation system with machine learning for 1st grade students in our University, who should select a course. The selection can be difficult or agonizing for some people. Then, we set out the development of course recommendation system to get the selection easier and more correct. Our goal in this project is to develop a logical course recommendation system with analysis of machine leaning of trends students in our university found by some kind of means. We organized two groups, Data gathering team and Data analysis team, for our goal. First, we studied R language, which would be utilized for data analysis. We learned method of text analysis through categorizing news articles with Mecab using a natural language analysis software as practice. At the same time, we discussed followings for development the system. Input and output data Framework of user interface A means of data gathering or analysis We decided that Data gathering team consider what king of data we need and get the dataset, and Data analysis group search and investigate the analysis methods. As a result of the discussion, we decided to have questionnaire for students from 2nd to 4th grade for the system development. In addition, we were gong to make HTML that show the information of courses to compensate the system ability giving advice about which course is suitable. Furthermore, we decide to take another data analysis method like factor analysis but not only machine learning to elaborate. Keyword data analysis, machine learning, R (programing language), support for making decision, Factor analysis : - ii -
1 1 1.1............................... 1 1.2...................................... 1 2 2 2.1............................ 2 2.1.1............. 2 2.2............................... 2 3 4 3.1................................ 4 3.1.1................................... 4 3.1.2.................................... 5 3.2................................. 5 3.3................................ 6 3.3.1................................ 6 3.3.2......................... 8 3.4................................. 10 3.4.1................................ 10 3.4.2............................... 12 3.4.3........................... 16 3.4.4............................... 16 3.4.5................................. 17 3.4.6........................ 19 3.4.7................................. 24 3.4.8................................ 30 4 37 4.1.................................. 37 4.2.................................... 37 5 39 40 - iii -
1 1 2 : 1.1 1 1 : 1.2 : Group Report of 2014 SISP - 1 - Group Number 18-B
2 2.1 1 1 UI UI : 2.1.1 : 2.2 1. R R R 2. 3. 4. Group Report of 2014 SISP - 2 - Group Number 18-B
5. 6. 7. 8. 9. 10. 11. : Group Report of 2014 SISP - 3 - Group Number 18-B
3 3.1 R R R : 3.1.1 1 csv R 2013/5/17 11.8 3.1 9.4 71 2013/5/18 12.5 2.7 9.7 68 2013/5/19 11.2 2.3 9.5 72 2013/5/20 10.9 1.9 12 92 2014/5/17 11.4 7.4 8.1 60 3.1 1. 2. 3. 4. 5. 4 LDA( ) SVM( ) 2 : Group Report of 2014 SISP - 4 - Group Number 18-B
Latent Dirichlet Allocation( ) LDA R LDA 80% : Support Vector Machine( ) SVM R SVM 83% : 3.1.2 R R RMeCab R. [1] : 3.2 R 1. ( ) 2. 3. 4. 5. SNS : Group Report of 2014 SISP - 5 - Group Number 18-B
3.3 1. Web 2. 3. 4. : 3.3.1 Web [2] txt -0000.txt 1725 eco 275 edi 52 ele 90 env 22 fiv 140 IT inf 102 pol 454 reg 73 soc 87 spo 298 tec 59 uni 110 3.2 : Group Report of 2014 SISP - 6 - Group Number 18-B
0 1. RMeCab 2. 1 3. 7 3 4. IT 61 3 1 0 2 5 11 5 4 2 2 2 4 10 0 1 0 0 3 1 0 0 0 0 0 0 4 0 1 0 16 0 0 0 2 0 1 0 0 2 0 0 0 2 1 0 3 0 4 0 0 0 25 0 2 0 1 0 0 0 IT 4 1 0 0 0 25 0 0 0 1 0 0 3 1 20 1 5 1 83 6 4 7 1 5 1 0 0 0 1 0 0 1 1 0 0 0 6 0 1 0 1 0 2 5 8 0 1 0 6 0 0 0 2 0 3 3 1 79 0 1 1 0 0 0 0 1 0 2 2 1 7 0 1 3 0 0 1 0 2 0 1 0 1 22 3.3 64% 8% : LDA( ) LDA 1. RMeCab 2. 1 3. 7 3 4. LDA Group Report of 2014 SISP - 7 - Group Number 18-B
IT 53 0 0 2 0 6 4 3 2 1 1 1 1 9 0 2 0 1 0 0 0 0 2 0 1 0 16 0 0 0 14 0 0 0 0 0 0 0 0 3 0 0 0 4 1 0 0 0 1 1 0 0 40 0 2 0 1 0 1 0 IT 1 1 0 0 0 20 1 1 0 2 2 0 4 0 23 0 2 3 84 2 1 0 0 3 4 0 0 0 0 0 2 11 6 1 0 0 5 0 0 0 0 2 2 11 6 1 0 0 2 0 0 0 0 1 0 1 0 90 0 0 3 0 0 0 0 0 2 1 0 1 10 0 1 0 0 0 0 4 1 0 2 0 0 18 3.4 LDA 71% : 2 2 1725 22 7 15 : 3.3.2 3 79 0. 1. Group Report of 2014 SISP - 8 - Group Number 18-B
20 18 22 19 3.5 2. 3. A 4. 5. 6. 7. 8. 9. 10. PC SNS 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 1 35. Group Report of 2014 SISP - 9 - Group Number 18-B
36. 37. PC 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. Yes or No 10 10 15. 16. 19. 20. 21. 22. 28. 32. 34. 1 38. 10 4 : 3.4 3.4.1 Group Report of 2014 SISP - 10 - Group Number 18-B
2 4 1 5 1 30 240 120 2 4 95 45 70 70 280 7:3 1 120 280 80,40,60,60 280 95,45,70,70 Group Report of 2014 SISP - 11 - Group Number 18-B
1 : 3.4.2 ( ) Group Report of 2014 SISP - 12 - Group Number 18-B
TV ( ) Group Report of 2014 SISP - 13 - Group Number 18-B
2 5 web Group Report of 2014 SISP - 14 - Group Number 18-B
OS IT ( ) ( LAN ) ( ) web Group Report of 2014 SISP - 15 - Group Number 18-B
( ) : 3.4.3 web web 2 4 web google 1.4 5 5 1 5 1 5 1 : Group Report of 2014 SISP - 16 - Group Number 18-B
3.4.4 2 2 2 4 232 84 56 50 42 2 76 3 107 4 49 48 184 : 3.4.5 4 Group Report of 2014 SISP - 17 - Group Number 18-B
PDF web google PDF web google PDF web Group Report of 2014 SISP - 18 - Group Number 18-B
PDF web web web Google web Harbor View Site B PDF web web Harbor View Site : 3.4.6 4 Group Report of 2014 SISP - 19 - Group Number 18-B
Support for decision making by data analysis どのような入力を与えるのか 判別システム どのような手法で構成するのか 出力 どのように結果を表示するのか 関連情報 それぞれのコースの特徴を提示する この中の 関連情報 について提示する方法として システムを補助するページを作成することと した ページの概要 ページの大まかな情報の区分として コース紹介 講義案内 教員紹介 プロジェクト 学習 研究室案内 の 5 項目に分けて情報を掲載することにした また 学内で公開されている コース案内はもちろんのこと すでにコースに所属している学生にアンケートをとり どのような ことが学べるかについて 情報をワードクラウド化して表現することで見やすくまとめた 下記画 像 1 枚目では 知能システムコースの例を取り上げている このような コース紹介 を行った また それぞれのコースごとに受講する講義が異なるため どのような 講義 があるのか につ いて 2,3,4 年次と学科別にまとめ 講義案内とした 所属するコースを選択する上で決め手となる こととして どのような教員が在籍しているかということも考えた それぞれの講義を担当してい る 担当教員 はどのような研究を専攻しているのか ということについて 検索しやすいように 情報をまとめた プロジェクト学習 では 現在活動している 22 個のプロジェクトについてま とめた 1 年次と 2 年次は それぞれのプロジェクトがどのような活動をしているのかということ を 前期末のプロジェクトの中間報告発表会や後期末の成果発表会などで知ることができる しか し 現在の 3 年次は何を学びたいと思いプロジェクトに携わっているのか どのような方針で活動 をするのかなどを知るため情報は 3 年次のプロジェクト学習が始まる前の配属段階にしか公開が されていない それらを 1 年次に公開することで 自分が何に興味を持って何を勉強するべきなの か 少しでも自分について考えることができると考えた 3 年次に公開された 各教員の研究室で 行われた過去の研究テーマ等の紹介をすることで より自分の将来像を考えるようになると考えら れる 画像の 2 枚目のように コースごと 教員ごとに 過去に行われた研究テーマの一覧を Web ページにて公開をしている 図 3.1 HTML ページ Group Report of 2014 SISP - 20 - Group Number 18-B
HTML HTML Web (HOME.html) URL file:///c:/users/user/desktop/home/home.html URL PC HOME HOME.html 1 HOME.html Web HOME HOME.html PICTURE (gazo.jpg) PICTURE/gazo.jpg HTML HTML Web <HEAD></HEAD> Web <BODY></BODY> 2 Web Web <a href></a> <a href="url"> </a> URL <a href="\#"> </a> URL # <a href="javascript:void(0)" "window.open( URL, Web, width=400,height=400 );"> </a> JavaScript void(0) window.open() css HTML Web CSS CSS CSS <div></div> id class p <div id="p"> </div> CSS p CSS p {font-size:large; color:blue; line-height:1.5;} 1.5 CSS Web Group Report of 2014 SISP - 21 - Group Number 18-B
3 JavaScript d3.js d3-clud.js JQUERY d3.js JavaScript d3.js d3.js ccchart.js Chart.js 2 Chart.js JavaScript Group Report of 2014 SISP - 22 - Group Number 18-B
JavaScript 3.2 2 4 1 Web R Shiny Web Web URL URL Query Query system.html?course=1&type=1&p1=59&p2=7&p3=21&p4=13 URL? course 2 3 4 course=1 type 1 type Group Report of 2014 SISP - 23 - Group Number 18-B
3.3 Shiny 5,,,,,,,,,,,,,,,, type=1 p1 p2 p3 p4 p1 p2 p3 p4 p1=59&p2=7&p3=21&p4=13 59% 7% 21% 13% 6 value[5] (1 4), (1 4),,,, Query JavaScript HTML Chart.js <script> var p1 = Number(value[2]); var p2 = Number(value[3]); var p3 = Number(value[4]); var p4 = Number(value[5]); </script> : Group Report of 2014 SISP - 24 - Group Number 18-B
3.4.7 R R 1. ( ) 2. 6 5 3 2 R factanal Group Report of 2014 SISP - 25 - Group Number 18-B
1 31 31 33 59 63 52 2 72 77 60 50 51 68 3 34 27 47 34 33 20 4 44 44 54 43 35 48 5 58 57 54 45 61 63 3.6 6 3.4 6 ( ) plot 0.249 0.129 0.290 0.174 0.255 0.335 3.7 6 0.249 25% Factor1 Factor2 0.823 0.947 0.869 0.938 0.865 0.762 3.8 6 (Factor1) (Factor2) 2 Group Report of 2014 SISP - 26 - Group Number 18-B
3.5 6 p p 1.2 13 biplot 3.6 6 ( ) ( ) ( ) ( ) Group Report of 2014 SISP - 27 - Group Number 18-B
19 19 1. 2. web 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 19 10 19 10 10 1. 2.web 4. 6. 7. 8. 9. Group Report of 2014 SISP - 28 - Group Number 18-B
3.7 19 11. 14. 18. 10 3 p 2 p 0.034 3 p 0.479 4 p 0.972 2 Group Report of 2014 SISP - 29 - Group Number 18-B
Factor1 Factor2 q1 0.473 q2 q4 0.547 q6 0.223-0.347 q7 0.777 0.113 q8 0.151-0.252 q9 0.654 q11 q14 0.154 0.466 q18 0.233 0.644 3.9 3.8 biplot biplot biplot 4 4 Group Report of 2014 SISP - 30 - Group Number 18-B
3.9 biplot 4 : Group Report of 2014 SISP - 31 - Group Number 18-B
3.4.8 csv R R 1. 2. 3. 4. class q1, q2,, q10 5. 1 2 3 4 q1 q2 q3 q4 q5 q6 q7 q8 q9 q10 class 4 2 4 5 4 2 2 3 4 4 1 5 3 4 2 5 3 2 5 5 5 1 5 1 1 3 1 3 1 4 5 5 1 3 2 4 3 1 4 1 1 2 5 2 3 4 4 2 4 4 2 1 2 4 2 1 3 1 5 1 2 1 4 1 1 2 1 5 1 5 1 5 1 3 1 1 3 4 1 4 1 4 4 3 2 4 4 3 4 4 3 3 2 1 2 3 2 3 3 3 4 2 5 2 4 2 4 2 4 4 2 4 4 3 4 5 4 3 2 4 4 2 2 3 2 2 4 1 4 2 3 4 3.10 R Group Report of 2014 SISP - 32 - Group Number 18-B
3 LDA( ) SVM( ) 1. 8 2 2. 3. 4. 1 3 5 5 LDA 42% SVM 84% 40% SVM R SVM 1 0 Group Report of 2014 SISP - 33 - Group Number 18-B
2 4 1. 4 2. 3. 30 5 1 2 3 4 5 3.11 5 2 1 2 3 4 5 3.12 4 1. 2. 4 3. 1 Group Report of 2014 SISP - 34 - Group Number 18-B
Ian Frank 4 4 3 4 4 4 4 1 4 V. Riabov 2 3.13 1. 2. 5 3. 4. Factor1 Factor2 Factor3 Factor4 Factor5 q1 0.41 0.05 0.00 0.01 0.02 q2-0.11 0.99 0.03 0.05 0.00 q3 0.17 0.04 0.20 0.96 0.06 q4-0.07-0.13 0.03 0.09 0.50 q5 0.44-0.01 0.33 0.24 0.48 q6 0.09 0.36 0.03-0.13 0.67 q7 0.17 0.02 0.95 0.23 0.07 q8 0.16-0.10-0.13 0.09-0.13 q9 0.58-0.07 0.03 0.05 0.02 q10 0.65-0.16 0.26 0.21-0.10 3.14 Group Report of 2014 SISP - 35 - Group Number 18-B
4 web web R shiny UI UI shiny shiny UI URL web HTML web 1. UI SVM 2. SVM 3. SVM 4. a b c 5. URL UI Group Report of 2014 SISP - 36 - Group Number 18-B
web web UI web R shiny runapp : Group Report of 2014 SISP - 37 - Group Number 18-B
Support for decision making by data analysis 第 4 章 成果 4.1 完成したシステム 完成したシステムは次のようなものとなった R の Shiny というパッケージで Web ブラウザ 図 4.1 完成したシステム 上でシステムを動作させている そして質問事項に回答したのちに HTML ページに飛ぶとシス テムの判別した情報がクエリとして引き渡されて HTML ページにて結果が表示される仕組みで ある 文責: 井川翼 4.2 システムの評価 1 年生 20 人にシステムを実際に使用してもらい 評価をもらったところ 以下のような結果と なった 1. 参考になった 50% 2. 少し参考になった 20% 3. どちらでもない 20% 4. あまり参考にならなかった 10% Group Report of 2014 SISP - 38 - Group Number 18-B
: Group Report of 2014 SISP - 39 - Group Number 18-B
5 1 : Group Report of 2014 SISP - 40 - Group Number 18-B
[1]. R., 2008. [2], YOMIURI ONLINE, http://www.yomiuri.co.jp/. [3], R-Tips, http://cse.naro.affrc.go.jp/takezawa/r-tips/r.html. [4] TAKESHI ARABIKI, R, http://www.slideshare.net/abicky/ r-22325351. [5], R Random Forest, http://www.slideshare.net/ hamadakoichi/introduction-torandomforest-tokyor. [6] grayspace, R, http://d.hatena.ne.jp/ grayspace/20140503/1399106654. [7] kj-ki, R nnet, http://d.hatena.ne.jp/kj-ki/20120124/p1. Group Report of 2014 SISP - 41 - Group Number 18-B