(SNS) Twitter 2014 12 LINE Facebook 3 SNS Twitter [1] [2] 500 Web Web API Twitter,, - i -

2015 Future University Hakodate 2015 System Information Science Practice Group Report Twitter Project Name Twitter Localization Group Name Analysis Group /Project No. 19- /Project Leader 1013039 Hirohito Maruyama /Group Leader 1013169 Naohira Akasaka /Group Member 1013003 Sho Ishibashi 1013007 Seiji Ogawa 1013010 Tatsuya Kawamukai 1013095 Haruka Sugawara 1013014 Tsuyosi Denpouya 1012207 Masato Murao Advisor Kengo Terasawa Takashi Takenouchi Kiyohito Nagano Yasuhiro Katagiri 2016 1 20 Date of Submission Janualy 20, 2016

(SNS) Twitter 2014 12 LINE Facebook 3 SNS Twitter [1] [2] 500 Web Web API Twitter,, - i -

Abstract The number of social networking service (SNS) users has increased in the past several years. In Japan, Twitter is the third-largest SNS next to LINE and Facebook in December, 2014. There are many Twitter applications, including personality test applications like appli-maker[1] and tweet-profiling[2]. However, the existing personality test applications have little flexibility. They cannot analyze tweets to the full extent, and they can deal with only 500 tweets for each account. The purpose of the analysis group is to provide a new personality analysis application that settles the above problems. We proposed a personality analysis web application, which classifies the personalities of Twitter users, and determines the recommended tourist spots in Hakodate by using machine learning. The analysis results of the proposed application are future analysis, which is the user s future personality, and group analysis, which is the personality of a group of randomly chosen mutual followers. In addition, we developed a tourist spot recommendation system in Hakodate. The recommendation system utilizes the result of tweet analysis and machine learning methods. In the first semester, we split into Web team, API team, and Egogram team to develop the personality analysis application. In the second semester, we continued to develop the application, and added a recommendation system of tourist spots in Hakodate. Keyword Twitter, Personality analysis, Machine learning - ii -

1 1 2 3 3 4 3.1.......................................... 4 3.2.......................................... 4 4 5 4.1................................ 5 4.2.......................... 5 4.3............................ 6 4.4............................ 9 4.4.1 API..................................... 9 4.4.2................................. 10 4.4.3 Web.................................... 11 4.5.................... 12 4.5.1.................................... 12 4.5.2.................................... 13 4.6.................................. 14 4.6.1................................... 14 4.6.2............................. 14 4.6.3.............................. 16 4.7.......................... 16 4.7.1 API +........................... 16 4.7.2 Web.................................... 21 4.8.................. 26 4.8.1.................................... 26 4.8.2.................................... 27 4.9................................ 29 5 31 5.1....................................... 31 5.2........................................ 33 5.3....................................... 34 5.4....................................... 35 5.5....................................... 36 5.6....................................... 37 5.7....................................... 38 - iii -

6 40 7 42 8 43 44 - iv -

1 Web (BBS) (Weblog) (SNS) [3] SNS Web [4] SNS 2006 Twitter 2008 4 23 Twitter Japan [5] 2014 12 LINE Facebook 3 [6] Twitter 2 8400 2000 [7] 140 Twitter 80%[7] Twitter Twitter Twitter Facebook SNS 2 1 1 5700 [8] 2 10 20 Twitter Twitter 2 1, Twitter 2 2, Web Web Web Group Report of 2015 SISP - 1 - Group Number 19-

Web Group Report of 2015 SISP - 2 - Group Number 19-

2 Twitter 140 Twitter Twitter ( ) [2] [1] Twitter Twitter 500 500 Twitter Twitter Group Report of 2015 SISP - 3 - Group Number 19-

3 3.1 Web Web API 3 3.2 Web TwitterAPI 3 Web Group Report of 2015 SISP - 4 - Group Number 19-

4 4.1 Twitter Twilog Twitter Twitter Analytics TweetDeck Krile STARRYEYES RtRT Plume Twitter 140 Twitter Twitter Twitter 2 2 4.2 4 Group Report of 2015 SISP - 5 - Group Number 19-

Web Twitter 3 3200 3 3 4.3 4 Group Report of 2015 SISP - 6 - Group Number 19-

Web Web 4 Twitter Web Twitter 1 TwitterAPI 3200 4.1 Web Web Web Web TwitterAPI API 3 4.1 Twitter 2 Web 1 3200 30 Group Report of 2015 SISP - 7 - Group Number 19-

4.2 Web 15 100 4.3 Web Web Web API+ 2 4.2 4.3 Group Report of 2015 SISP - 8 - Group Number 19-

4.4 4.4.1 API API Web TwitterAPI TwistOAuth[9] Web Web TwitterAPI 200 API 200 200 ID ID 200 3200 5 Twitter Twitter Web Twiter TwitterAPI Web Web eclipse PHP HTML Group Report of 2015 SISP - 9 - Group Number 19-

API : 4.4.2 M CP(Critical Parent) : NP(Nurturing Parent) : A(Adult) : FC(Free Child) : AC(Adapted Child) : 5 [11] CP NP A FC AC 4.4 4.5 FC 2 A 1 CP 1 1 2 4.6 [12] 5 500 R [13] API 1 4.4 Group Report of 2015 SISP - 10 - Group Number 19-

4.5 4.6 4.4.3 Web Web Web 3 HTML Twitter 3 4.1 4.1 4.1 Web HTML Group Report of 2015 SISP - 11 - Group Number 19-

4.5 4.5.1 1 1 1 3 1 ( ) 4 3 28pt 500 Web 2 Group Report of 2015 SISP - 12 - Group Number 19-

3 4.7 4.7 4.5.2 4 Web 1 1 4.4 4.5 4.6 4.4 1 Group Report of 2015 SISP - 13 - Group Number 19-

4.5 FC A CP 4.6 4.5 CP NP A FC AC N 4.6 2015 7 10 4.6.1 4.6.2 500 3200 Group Report of 2015 SISP - 14 - Group Number 19-

2 20 3200 Group Report of 2015 SISP - 15 - Group Number 19-

4.6.3 Web 1 4.7 4.7.1 API + 3200 3200 TwitterAPI 30 30 R exe exe python python Group Report of 2015 SISP - 16 - Group Number 19-

TwitterAPI 1. ID 15 ID 2. 15 100 3. 1. PHP TwitterAPI 100 Processing 2. Processing PHP 3. PHP ID : 5 ( ) 10 1 446 ( ) Group Report of 2015 SISP - 17 - Group Number 19-

R R [10] 221028 107944 R ( ) 49% 10 27.58% 6 Web 10 R python 0 python 1428 107944 445 2 107944 10 Web 10 ( TFIDF ) TFIDF 4.1 python 446 10 9 10 107944 Normal L Normal S TFIDF 107944 TFIDF L TFIDF Group Report of 2015 SISP - 18 - Group Number 19-

4.1 107944 TFIDF 107944 TFIDF 107944 TFIDF S 107944 random 4.2 (python) (python) (R) Normal L 29.21% 17.58% Normal S 21.33% 20.00% 17.39% TFIDF L TFIDF S 13.33% 23.33% 14.29% random 13.04% 12.02% 14.13% 1000 29.21% 21.33% 5% random R R 4.2 3 40 10 3 4.3 4.3 Normal L 51.61% 83.65% Normal S 58.06% 80.97% TFIDF L TFIDF S 51.61% 80.65% 5 50% 8 5 8 5 Group Report of 2015 SISP - 19 - Group Number 19-

8C5 56 3 10000 107944 2 46% 5 ( ) 50% 2 20000 48.9% 52% 50% 8 100 52.33 0.003358 0.0579 50% 107944 2 t (t=14.09,df=54,p<0.01) Web 50% 50% 3 1 1 1 3 1 1 8 9 9 10 1 50% Group Report of 2015 SISP - 20 - Group Number 19-

4.7.2 Web Web Twitter 3 Twitter Twitter Twitter Twitter? 2 3 Twitter? Web Twitter Twitter Twitter Twitter 5 (CP NP AF C AC) 5 4.8 5 5 CP NP AF C AC CP NP A FC AC CP NP A FC AC Group Report of 2015 SISP - 21 - Group Number 19-

4.8 5 2 Twitter Twitter ( ) ( ) 2 Twitter Twitter 2 Twitter 3 3 3 ( ) 2 Group Report of 2015 SISP - 22 - Group Number 19-

Twitter 3 3 3 (Twitter ID) Twitter Twitter ( ) 1 Twitter (Twitter ID) 5 (CP NP A FC AC) 5 Group Report of 2015 SISP - 23 - Group Number 19-

2 Twitter 10 Javascript 10 10 2 Twitter Web 10 2 php.ini PHP Web Web Group Report of 2015 SISP - 24 - Group Number 19-

FTP FTP WinSCP FTP Web PHP exe exe exe exe Web VPS Window server for VPS 8GB 2015 11 2015 3 VPS Windows server 2012 R2 IIS8.5 Windows server 2012 R2 Web (IIS) IIS ISAPI CGI IIS8.5 PHP5.4 Python2.7.10 Java8 php.ini Python Numpy scipy sklearn 500 500 php.ini session.save path 500 SSL certificate problem: unable to get local issuer certificate [14] curl SSL certificate problem: unable to get local issuer certificate HTTP 500.0 - Internal Server Error 500 c:\php5\php-cgi.exe - FastCGI php.ini IIS FastCGIsetting VPS ISAPI CGI USERS php.ini 500 1 PHP exec PHP exec 50 PHP exec Array() PHP exec VPS Array() python Array() Group Report of 2015 SISP - 25 - Group Number 19-

IIS IUSRS USERS CGI ISAPI php.ini UAC 3 4.8 4.8.1 1 1 1 3 1 2 24pt 2015 12 2 2 Group Report of 2015 SISP - 26 - Group Number 19-

Twitter Localization ケーション内部で実行されるシステムの流れである 診断者が診断の選択画面にて診断を選択する と それぞれ異なった分析方法でアプリ内部の操作が行われ ツイートの分析を行った後にそれぞ れの診断結果が表示されるようになっている アプリ内部の操作で行われる形態素解析 エゴグラ ムの特徴辞書で点数化 機械学習によるお勧め観光地診断に使われている画像は分析班のスライド 班と共有しているものである 現在 未来診断 では作成したエゴグラムから現在 未来の性格 の推移を見れるということがわかるように矢印のある画面になり 集団診断 では矢印はないが 気の合う人を表示するという項目がある画面になり それぞれの結果表示画面に違いがあることが わかるようにしている 最後には中間発表の時にもあったツイート共有機能もついている 以上が 性格診断の一連の流れであるが どこで 現在 未来診断 と 集団診断 の分析の違いが表れる かがこの図だけではわからないという問題点もあった 中間発表では現在までの取り組みと展望を記載していたが 分析班が目標としていたサービスの 作成が完了していたので 今回は記載しないこととした 図 4.9 が実際に最終発表で使用した分析 班ポスターである 図 4.9 最終発表のポスター 文責: 傳法谷強 4.8.2 スライド 最終成果発表で分析班が開発した Web アプリケーションの説明をするためのスライドを作成し た 初めに作成したスライドでは 手法を説明するスライドを多く作っていた しかし 手法が多 Group Report of 2015 SISP - 27 - Group Number 19-分析班

4.10 4.11 4.12 4.10 4.11 Group Report of 2015 SISP - 28 - Group Number 19-

4.12 4.9 2015 12 11 1. 2. 3. Web 4. 5. 6. 7. Web PC Web PC 2 Web PC Group Report of 2015 SISP - 29 - Group Number 19-

A FC 3200 2 2 PC 50% 2 : Group Report of 2015 SISP - 30 - Group Number 19-

5 5.1 TwitterAPI TwistOAuth Web Web Processing Java 2 1 API 1 10 ID Group Report of 2015 SISP - 31 - Group Number 19-

15 15 TwitterAPI API API TFIDF R R python Java R python TwitterAPI php php API API Web mecab kuromoji TwitterAPI TwistOAuth Twitter4j TwitterOAuth Twitter4j Java Web Group Report of 2015 SISP - 32 - Group Number 19-

TwitterOAuth php TwistOAuth 5.2 API 2 Web HTML Twitter Web Web TwitterAPI HTML PHP eclipse HTML 2 TwitterAPI API TwitterID ID ID 2 ID ID 15 3200 100 1 PHP Processing PHP 2 PHP Processing Processing exe Group Report of 2015 SISP - 33 - Group Number 19-

PHP Processing PHP Processing PHP Processing 1 2 15 GoogleDrive Web : 5.3 Web (Twitter ) Web 3 Twitter Web Web HTML CSS Javascipt 3 Javascipt HTML CSS Twitter Twitter ID Twitter Twitter Twitter 2 3 Twitter 2 Twitter 2 Group Report of 2015 SISP - 34 - Group Number 19-

Twitter 2 2 2 2 10 AC CP 1 5.4 R 500 250 R API API R exe python exe numpy PHP 30 Group Report of 2015 SISP - 35 - Group Number 19-

Web API 30 python 2 49% 10 15% 4% API 4 5 6 30 3 5.5 Web Web Twitter Web 10 2 Group Report of 2015 SISP - 36 - Group Number 19-

php.ini PHP VPS Window server for VPS 8GB 2015 11 2015 3 Windows server 2012 R2 IIS8.5 IIS ISAPI CGI IIS8.5 PHP5.4 Python2.7.10 Java8 python Numpy scipy sklearn 3 5.6 Web Web Web HTML CSS Web Web Twitter Web 2 3 1 1 Web Group Report of 2015 SISP - 37 - Group Number 19-

1 1 2015 12 2 ( ) ( ) 4 Web QR Android QR URL 5.7 500 API 0 5 12 Group Report of 2015 SISP - 38 - Group Number 19-

Twitter Group Report of 2015 SISP - 39 - Group Number 19-

6 Web Web URL 6.1 3 1 Twitter 2 Web Twitter ID Twitter ID 6.2 6.1 6.2 2 6.3 10 6.4 TwitterID Group Report of 2015 SISP - 40 - Group Number 19-

Twitter 6.3 6.4 Firefox42.0 IE11 Windows server 2015 3 : Group Report of 2015 SISP - 41 - Group Number 19-

7 2016 2 Web 4 15 Firefox42.0 InternetExplorer11 Chrome Chrome Web URL : Group Report of 2015 SISP - 42 - Group Number 19-

8 Web Web API 3 Web Web API 2 2 Web Web API+ 2 Web Web API+ Web Group Report of 2015 SISP - 43 - Group Number 19-

[1],, http://appli-maker.jp/ (2016/1/8 ). [2] ktty1220,, http://tweet-profiling.ktty1220.me/ (2016/1/8 ). [3],, vol.47, no.8, pp.993-1000, September. 2006. [4] IT, SNS Social Networking Service, http://e-words.jp/w/sns.html (2015/12/18 ). [5] SNS LINE Twitter,, 2014. [6] nielsen, 92% SNS, SNS, http://www.netratings.co.jp/news_release/2015/01/ Newsrelease20150127.html (2016/7/24 ). [7] Twitter, Inc., Twitter Reports Third Quarter 2014 Results, https://investor. twitterinc.com/releasedetail.cfm?releaseid=878170 (2015/7/24 ). [8] Twitter Development of Speech Function for Communication Robot using Twitter,,, 2014. [9] mpyw, TwistOAuth/README EXAMPLES.md,https://github.com/mpyw/ TwistOAuth/blob/master/README\_EXAMPLES.md,(2015/7/19 ). [10] @sfchaos, R boost.interprocess,http://www.slideshare.net/sfchaos/rboostinterprocess (2015/11/8 ). [11] 243,, 1995, ( ; B-1063. ; part 2). [12] CiNii Articles,, http://ci.nii.ac.jp/naid/ 110008123560 (2015/7/22 ). [13] R., 2012. [14] Various Program Blog, windows curl,http://piji. daiwa-hotcom.com/wordpress/?p=1005 (2016/1/20 ). Group Report of 2015 SISP - 44 - Group Number 19-