Personal Software Process /Team Software Process (PSP/TSP) Watts Humphrey PSP/TSP PSP/TSP TaskPit TaskPit PSP/TSP TaskiPit PSP/TSP 5 5

Size: px

Start display at page:

Download "Personal Software Process /Team Software Process (PSP/TSP) Watts Humphrey PSP/TSP PSP/TSP TaskPit TaskPit PSP/TSP TaskiPit PSP/TSP 5 5"

こうだいえんの
5 years ago
Views:

2 Personal Software Process /Team Software Process (PSP/TSP) Watts Humphrey PSP/TSP PSP/TSP TaskPit TaskPit PSP/TSP TaskiPit PSP/TSP 5 5

3 PSP/TSP TaskPit

4 1 Personal Software Process /Team Software Process (PSP/TSP) Watts Humphrey [1] PSP/TSP PSP/TSP [2][3][4] TaskPit[6] TaskPit WINWORD.EXE.docx.docx 1 1 Random Forests 5 1

5 2 2.1 PSP/TSP Watts PSP/TSP PSP/TSP PSP/TSP 1996 Process Dashboard[2] Task Coach[3] Slim Timer[4] TaskPit[6] Task Pit TaskPit 2.2 Michael [5] Michael Web PowerPoint Naive Bayes(NB) Linear Support Vector Machine(SVM) k-nearest Neighbor(k-NN) Michael NB -NN Michael Web 2

6 3

7 3 Random Forests TaskPit = [ B C ] (1) = [ ] = [ ] = [ ] 1 4

8 仕様書上での作業設計書上での作業開発環境上での作業作業目的 = 実装仕様書上での作業設計書上での作業作業目的 = 設計作業目的 = テスト仕様書上での作業設計書上での作業開発環境上での作業 :: クリックの多さ :: 打鍵の多さ TaskPit TaskPit PSP/TSP 2 TaskPit 1 Word Eclipse Word Word spec TaskPit 3 TaskPit TaskPit 4 TaskLog 5

TaskPit 2 TaskPit = TeraPad.exe notepad.exe WINWORD.EXE Acrobat.exe LabyTex.exe = POWERPNT.EXE = EXCEL.EXE Rgui.exe javaw.exe:weka javaw.exe:gemx = eclipse.exe devenv.exe EXCEL.

9 TaskPit 2 TaskPit = TeraPad.exe notepad.exe WINWORD.EXE Acrobat.exe LabyTex.exe = POWERPNT.EXE = EXCEL.EXE Rgui.exe javaw.exe:weka javaw.exe:gemx = eclipse.exe devenv.exe EXCEL.EXE:Microsoft Visual Basic = devenv.exe: TaskPit = TaskPit.exe TaskAnalyzer.exe = iexplore.exe firefox.exe Safari.exe chrome.exe = thunderbird.exe Outlook.exe iexplore.exe:gmail 3,,,,,, , ,0,0,0 TaskPit, , ,3,0,0, , ,2,0,0, , ,1,0,0, , ,5,0,2, , ,24,0,64, , ,2,0,0 TaskPit, , ,1,0,0, , ,2,0,0 4 6

10 3.3 Random Forests[7] Random Forests R Random Forests ntree mtry M Breiman ntree = 500 mtry = M Random Forests P M 1. P P N N B1 B2 Bi BN 1/3 OOB(Out- Of-Bag) 2. Bi Ti Random Forests OOB OOB Ti mtry 3. Bi OOB 7

11 4 5 TaskPit TaskPit TaskPit 12/16 12/22 5 OS CPU 5 3 Windows7 Core2 Duo E Windows Vista Core2 Duo E Microsoft Office Word Microsoft VisualStudio Eclipse 4.3 TaskPit 5 6 TaskPit 8

12 PSP/TSP TaskPit,,,, TaskPit,19,12,0, TaskPit,3,0,0,,35,3,153,,2,0,0,,12,0,0,,9,0,0,,9,0,4,,1,0,0,,1,0,0, 6 / 9

13 ) 2) 1 3) 2 3) 3 2-fold-cross-validation 1 1 recall( ) precision( ) F1-value(F1 ) X 2 A X X B X X C X X D X X 1 ( ) ( ) ( ) ( ) ( ) ( )

14 2 X X X A B X C D 2 X X precision A B 1 precision = A A + B precision 0 1 X X X recall A C 3 recall = A A + C recall 0 1 X X recall precision X X X X X X F1-value F1-value precision recall 4 F 1 value = 2 1 precision + 1 recall F1-value 0 1 precision recall F1-value 11 (2) (3) (4)

15 10 12

5 4 4 5.1 3 3 7 9 7 9 3 3 1 2 3 precision 0.521 0.693 0.752 0.

16 precision recall F1-value precision 13

17 8 recall 9 F1-value F1-value recall % Twitter Twitter 14

18 11 Twitter 3 Twitter 4 3 precision recall ,,,, Twitter,1,0,0,,1,0,1, Twitter,3,0,7,,11,0,14, Twitter,1,0,8,,26,0,0, Twitter,1,0,0,,53,10,52, Twitter,1,0,0,,1,0,0, Twitter,1,0,0,,1,0,0, Twitter,1,0,0,,19,2,1, Twitter,1,0,5,,1,0,0, Twitter,2,0,74, 10 ( ) 15

19 6 3 F1-value ,,,, Twitter,1,0,0, ( ),1,0,0, ( ),3,0,0, ( ) Twitter,1,0,2, ( ),1,0,0, ( ),4,0,0, ( ),1,0,0, ( ) or,1,0,0, ( ) Twitter,1,0,0, ( ),1,0,0, ( ) or,1,0,0, ( ),2,0,0, ( ),1,0,0, ( ) or,1,0,0, ( ),11,0,5, ( ),2,0,0, ( ),0,0,0, ( ),0,0,0, ( ) Random Forests

20 PSP/TSP TaskPit TaskPit 17

21 18

22 6 TaskPit TaskPit TaskPit PSP/TSP TaskPit 19

23 1 20

24 [1] Humphrey Watts S May 2001 [2] Process Dashboard [3] Task Coach Your friendly task manager [4] SlimTimer Time Tracking without the Timesheet [5] Granitzer Michael and Rath Andreas S and Kroll Mark and Seifert Christin and Ipsmiller Doris and Devaurs Didier and Weber Nicolas and Lindstaedt Stefanie N Machine Learning based Work Task Classification Journal of Digital Information Management October 2009 pp (2009) [6] XV FOSE2008 pp , Nov [7] Breiman, L Random Forests, Machine Learning Vol.45 No.1 pp.5 32 (2001) 21

Vol. 29 No. 1 Feb. 2012 79,,, PC TaskPit Manic Time [3] 4 TaskPit Task Task Aim Task Aim Task Aim Aim Task Task Task Random Forests Aim Task Task Aim

Vol. 29 No. 1 Feb. 2012 79,,, PC TaskPit Manic Time [3] 4 TaskPit Task Task Aim Task Aim Task Aim Aim Task Task Task Random Forests Aim Task Task Aim 78 PSP (Task) (Aim) Task Task Aim Task RandomForest Task 97 In this paper we propose a method to support Personal Software Process (PSP), which is a well known software process improvement framework for