S.P.A.C.E. Exploration for Software Engineering
1 like1,053 views
![GLOBALISM
& RESEARCHERS
R. Glass, Facts and Falllacies of Software
Engineering. Addison- Wesley, 2002.
C. Jones, Estimating Software Costs, 2nd
Edition. McGraw-Hill, 2007.
B. Boehm, E. Horowitz, R. Madachy, D.
Reifer, B. K. Clark, B. Steece, A. W.
Brown, S. Chulani, and C. Abts, Software
Cost Estimation with Cocomo II. Prentice
Hall, 2000.
R. A. Endres, D. Rombach, A Handbook
of Software and Systems Engi- neering:
Empirical Observations, Laws and
Theories. Addison Wesley, 2003.
• 50 laws:
• “the nuggets that must be captured
24
12/1/2011
to improve future performance” [p3]](https://image.slidesharecdn.com/menziesdec12-111127105256-phpapp02/85/S-P-A-C-E-Exploration-for-Software-Engineering-24-320.jpg)
S.P.A.C.E. Exploration for Software Engineering
- 1. S.P.A.C.E. & COWS & SOFT. ENG. TIM MENZIES WVU DEC 2011
- 2. THE COW DOCTRINE • Seek the fence where the grass is greener on the other side. • Learn from there • Test on here • Don’t rely on trite definitions of “there” and “here” • Cluster to find “here” and “there” 12/1/2011 2
- 3. THE AGE OF “PREDICTION” IS OVER OLDE WORLDE NEW WORLD Porter & Selby, 1990 Time to lift our game • Evaluating Techniques for Generating No more: D*L*M*N Metric-Based Classification Trees, JSS. • Empirically Guided Software Development Time to look at the bigger picture Using Metric-Based Classification Trees. IEEE Software Topics at COW not studied, not • Learning from Examples: Generation and publishable, previously: Evaluation of Decision Trees for Software Resource Analysis. IEEE TSE • data quality In 2011, Hall et al. (TSE, pre-print) • user studies • reported 100s of similar • local learning studies. • conclusion instability, • L learners on D data sets What is your next paper? in a M*N cross-val • Hopefully not D*L*M*N The times, they are a changing: harder now to publish D*L*M*N 12/1/201 3
- 4. REALIZING AI IN SE (RAISE’12) An ICSE’12 workshop submission • Organizers: Rachel Harrison, Daniel Rodriguez, Me AI in SE research • To much focus on low-hanging fruit; • SE only exploring small fraction of AI technologies. Goal: • database of sample problems that both SE and AI researchers can explore, together Success criteria • ICSE'13: meet to report papers written by teams of authors from SE &AI community 12/1/2011 4
- 5. ROADMAP Some comments on the state of the art • Why so much SE + data mining? • Why research SE + data mining • But is data mining relevant to industry • The problem of conclusion instability Learning local • Globalism: learn from all data • Localism: learn from local samples • Learning locality with clustering (S.P.A.C.E.) • Implications 12/1/2011 5
- 6. ROADMAP Some comments on the state of the art • Why so much SE + data mining? • Why research SE + data mining • But is data mining relevant to industry • The problem of conclusion instability Learning local • Globalism: learn from all data • Localism: learn from local samples • Learning locality with clustering (S.P.A.C.E.) • Implications 12/1/2011 6
- 7. Q1: WHY SO MUCH SE + DATA MINING? A: INFORMATION EXPLOSION http://CIA.vc • Monitors 10K projects • one commit every 17 secs SourceForge.Net: • hosts over 300K projects, Github.com: • 2.9M GIT repositories Mozilla Firefox projects : • 700K reports 12/1/2011 7
- 8. Q1: WHY SO MUCH SE + DATA MINING? A: WELCOME TO DATA-DRIVEN SE Olde worlde: large “applications” (e.g. MsOffice) • slow to change, user-community locked in New world: cloud-based apps • “applications” now 100s of services • offered by different vendors • The user zeitgeist can dump you and move on • Thanks for nothing, Simon Cowell • This change the release planning problem • What to release next… • … that most attracts and retains market share Must mine your population • To keep your population 12/1/2011 8
- 9. ROADMAP Some comments on the state of the art • Why so much SE + data mining? • Why research SE + data mining • But is data mining relevant to industry • The problem of conclusion instability Learning local • Globalism: learn from all data • Localism: learn from local samples • Learning locality with clustering (S.P.A.C.E.) • Implications 12/1/2011 9
- 10. Q2: WHY RESEARCH SE + DATA MINING? A: NEED TO BETTER UNDERSTAND TOOLS Q: What causes the variance in our results? • Who does the data mining? • What data is mined? • How the data is mined (the algorithms)? • Etc 10 12/1/2011
- 11. Q2: WHY RESEARCH SE + DATA MINING? A: NEED TO BETTER UNDERSTAND TOOLS Q: What causes the variance in our results? • Who does the data mining? • What data is mined? • How the data is mined (the algorithms)? • Etc Conclusions depend on who does the looking? • Reduce the skills gap between user skills and tool capabilities • Inductive Engineering: Zimmermann, Bird, Menzies (MALETS’11) • Reflections on active projects • Documenting the analysis patterns 11 12/1/2011
- 12. Inductive Engineering: Understanding user goals to inductively generate the models that most matter to the user. 12 12/1/2011
- 13. Q2: WHY RESEARCH SE + DATA MINING? A: NEED TO UNDERSTAND INDUSTRY You are a university educator designing graduate classes for prospective industrial inductive engineers • Q: what do you teach them? You are an industrial practitioner hiring consultants for an in-house inductive engineering team • Q: what skills do you advertise for? You a professional accreditation body asked to certify an graduate program in “analytics” • Q: what material should be covered? 13 12/1/2011
- 14. Q2: WHY RESEARCH SE + DATA MINING? A: BECAUSE WE FORGET TOO MUCH Basili • Story of how folks misread NASA SEL data • Required researchers to visit for a week • before they could use SEL data But now, the SEL is no more: • that data is lost The only data is the stuff we can touch via its collectors? • That’s not how physics, biology, maths, chemistry, the rest of science does it. • Need some lessons that survive after the institutions pass 14 12/1/2011
- 15. Its not as if we can embalm those researchers, keep them with us forever Unless you are from University College
- 16. PROMISE PROJECT 1) Conference, 2) Repository to store data from the conference: promisedata.org/data Steering committee: • Founders: me, Jelber Sayyad • Former: Gary Boetticher, Tom Ostrand, Guntheur Ruhe, • Current: Ayse Bener, me, Burak Turhan, Stefan Wagner, Ye Yang, Du Zhang Open issues • Conclusion instability • Privacy: share, without reveal; • E.g. Peters & me ICSE’12 • Data quality issues: • see talks at EASE’11 and COW’11 See also SIR (U. Nebraska) and ISBSG 16 12/1/2011
- 17. ROADMAP Some comments on the state of the art • Why so much SE + data mining? • Why research SE + data mining • But is data mining relevant to industry • The problem of conclusion instability Learning local • Globalism: learn from all data • Localism: learn from local samples • Learning locality with clustering (S.P.A.C.E.) • Implications 17 12/1/2011
- 18. Q3: BUT IS DATA MINING RELEVANT TO INDUSTRY? A: Which bit of industry? Different sectors of (say) Microsoft need different kinds of solutions As an educator and researchers, I ask “what can I do to make me and my students readier for the next business group I meet?” Microsoft research, Other studios, Redmond, Building 99 many other projects 18 12/1/2011
- 19. Q3: BUT IS IT RELEVANT TO INDUSTRY? A: YES, MUCH RECENT INTEREST POSITIONS OFFERED TO MSA GRADUATES: Credit Risk Analyst Business intelligence Data Mining Analyst E-Commerce Business Analyst Predictive analytics Fraud Analyst Informatics Analyst NC state: Masters in Analytics Marketing Database Analyst Risk Analyst Display Ads Optimization Senior Decision Science Analyst Senior Health Outcomes Analyst Life Sciences Consultant MSA Class 2011 2010 2009 2008 Senior Scientist Forecasting and Analytics graduates: 39 39 35 23 Sales Analytics %multiple job offers by Pricing and Analytics graduation: 97 91 90 91 Strategy and Analytics Range of salary offers 70K- 65K – 65K – Quantitative Analytics 140K 150K 60K- 115K 135K Director, Web Analytics Analytic Infrastructure Chief, Quantitative Methods Section 19 12/1/2011
- 20. ROADMAP Some comments on the state of the art • Why so much SE + data mining? • Why research SE + data mining • But is data mining relevant to industry • The problem of conclusion instability Learning local • Globalism: learn from all data • Localism: learn from local samples • Learning locality with clustering (S.P.A.C.E.) • Implications 20 12/1/2011
- 21. The Problem of Conclusion Instability Learning from software projects So we can’t take on conclusions from • only viable inside one site verbatim industrial development • Need sanity checks +certification organizations? envelopes + anomaly detectors • e.g Basili at SEL • check if “their” conclusions work “here” • e.g. Briand at Simula Even “one” site, has many projects. • e.g Mockus at Avaya • e.g Nachi at Microsoft • Can one project can use another’s conclusion? • e.g. Ostrand/Weyuker at AT&T • Finding local lessons in a cost-effective manner! Conclusion instability is a repeated observation. • What works here, may not work there • Shull & Menzies, in “Making Software”, 2010 • Sheppered & Menzies: speial issue, ESE, conclusion instability
- 22. ROADMAP Some comments on the state of the art • Why so much SE + data mining? • Why research SE + data mining • But is data mining relevant to industry • The problem of conclusion instability Learning local • Globalism: learn from all data • Localism: learn from local samples • Learning locality with clustering (S.P.A.C.E.) • Implications 22 12/1/2011
- 23. GLOBALISM: BIGGER SAMPLE IS BETTER E.g. examples from 2 sources about 2 application types Source Gui apps Web apps Green Software Inc gui1, gui2 web1, web2, Blue Sky Ltd gui3, gui4 web3, web4 To learn lessons relevant to “gui1” • Use all of {gui2, web1, web2} + {gui3, gui4, web3, web4} 23 12/1/2011
- 24. GLOBALISM & RESEARCHERS R. Glass, Facts and Falllacies of Software Engineering. Addison- Wesley, 2002. C. Jones, Estimating Software Costs, 2nd Edition. McGraw-Hill, 2007. B. Boehm, E. Horowitz, R. Madachy, D. Reifer, B. K. Clark, B. Steece, A. W. Brown, S. Chulani, and C. Abts, Software Cost Estimation with Cocomo II. Prentice Hall, 2000. R. A. Endres, D. Rombach, A Handbook of Software and Systems Engi- neering: Empirical Observations, Laws and Theories. Addison Wesley, 2003. • 50 laws: • “the nuggets that must be captured 24 12/1/2011 to improve future performance” [p3]
- 25. GLOBALISM & INDUSTRIAL ENGINEERS Mind maps of developers Brazil (top) from PASSOS et al 20011 USA (bottom) 25 12/1/2011 See also, Jorgensen, TSE, 2009
- 26. (NOT) GLOBALISM & DEFECT PREDICTION 26 12/1/2011
- 27. (NOT) GLOBALISM & EFFORT ESTIMATION Effort = a . locx . y • learned using Boehm’s methods • 20*66% of NASA93 • COCOMO attributes • Linear regression (log pre-processor) • Sort the co-efficients found for each member of x,y 27 12/1/2011
- 28. CONCLUSION (ON GLOBALISM) 28 12/1/2011
- 29. ROADMAP Some comments on the state of the art • Why so much SE + data mining? • Why research SE + data mining • But is data mining relevant to industry • The problem of conclusion instability Learning local • Globalism: learn from all data • Localism: learn from local samples • Learning locality with clustering (S.P.A.C.E.) • Implications 29 12/1/2011
- 30. LOCALISM: SAMPLE ONLY FROM SAME CONTEXT E.g. examples from 2 sources about 2 application types Source Gui apps Web apps Green Software Inc gui1, gui2 web1, web2, Blue Sky Ltd gui3, gui4 web3, web4 To learn lessons relevant to “gui1” • Restrict to just this the gui tools {gui2, gui3, gui4 } • Restrict to just this company {gui2,web1, web2} Er… hang on • How to find the right local context? 30 12/1/2011
- 31. DELPHI LOCALIZATION Ask an expert to find the right local context • Are we sure they’re right? • Posnett at al. 2011: • What is right level for learning? • Files or packages? • Methods or classes? • Changes from study to study And even if they are “right”: • should we use those contexts? • E.g. need at least 10 examples to learn a defect model (Valerdi’s rule, IEEE Trans, 2009) • 17/147 = 11% of this data 31 12/1/2011
- 32. CLUSTERING TO FIND “LOCAL” TEAK: estimates from “k” nearest-neighbors • “k” auto-selected per test case • Pre-processor to cluster data, remove worrisome regions • IEEE TSE, Jan’11 T = Tim E = Ekrem Kocaguneli A = Ayse Bener K= Jacky Keung ESEM’11 • Train within one delphi localization • Or train on all and see what it picks • Results #1: usually, cross as good as within 32 12/1/2011
- 33. Results #2: 20 times, estimate for x in S_i. TEAK picked across as picked within 33 12/1/2011
- 34. CONCLUSION (ON LOCALIZATION) Delphi localizations • Can restrict sample size • Don’t know how to check if your delphi localizations are “right” • How to learn delphi localizations for new domains? • Not essential to inference Auto-learned localizations (learned via nearest neighbor methods) • Works just as well as delphi • Can select data from many sources • Can be auto-generated for new domains • Can hunt out relevant samples from data from multiple sources 34 12/1/2011
- 35. ROADMAP Some comments on the state of the art • Why so much SE + data mining? • Why research SE + data mining • But is data mining relevant to industry • The problem of conclusion instability Learning local • Globalism: learn from all data • Localism: learn from local samples • Learning locality with clustering (S.P.A.C.E.) • Implications 35 12/1/2011
- 36. CLUSTERING + LEARNING Turhan, Me, Bener, ESE journal ’09 • Nearest neighbor, defect prediction • Combine data from other sources • Prune to just the 10 nearest examples to each test instance • Naïve Bayes on the pruned set Turhan et al. (2009) Me et al, ASE, 2011 Not scalable Near linear time processing No generalization to report to users Use rule learning 36 12/1/2011
- 37. CLUSTERING + LEARNING ON SE DATA Cuadrado, Gallego, Rodriguez, Sicilia, Rubio, Crespo. Journal Computer Science and Technology (May07) • EM on to 4 Delphi localizations • case tool = yes, no • methodology used = yes, no • Regression models, learned per cluster, do better than global But why train on your own clusters? • If your neighbors get better results… • … train on neighbors… • … test on local • Training data similar to test • No need for N*M-way cross val 37 12/1/2011
- 38. MUST DO BETTER Turhan et al. (2009) Me et al, ASE, 2011 Not scalable Near linear time processing No generalization to report to users Use rule learning Cuadrado et .al (2007) Me et al, ASE, 2011 Only one data set Need more experiments Just effort estimation Why not effort and defect? Delphi and automatic localizations ? Seek fully automated procedure Returns regression models Our users want actions, not trends. Navigators, not maps Clusters on naturally dimensions What about synthesized dimensions? Train and test on local clusters Why not train on superior neighbors (the envy principle) Tested via cross-val Train on neighbor, test on self. No 10*10-way cross val 38 12/1/2011
- 39. S.P.A.C.E = SPLIT, PRUNE SPLIT: quadtree generation PRUNE: FORM CLUSTERS Pick any point W; find X furthest from W, Combine quadtree leaves find Y furthest from Y. with similar densities XY is like PCA’s first component; found in Score each cluster by median O(2N) time, note O(N2) time score of class variable All points have distance a,b to (X,Y) Find envious neighbors (C1,C2) x= (a2 + c2 − b2)/2c ; y= sqrt(a2 – x 2) • score(C2) better than score(C1) Recurse on four quadrants formed Train on C2 , test on C2 from median(x), median(y) 39
- 40. WHY SPLIT, PRUNE? Unlike Turhan’09: LogLinear clustering time: i.e. fast and scales S.P. Turhan et al. (2009) Me et al, ASE, 2011 Not scalable Near linear time processing No generalization to report to users Use rule learning S. Cuadrado et .al (2007) Me et al, ASE, 2011 P. Only one data set Need more experiments Just effort estimation Why not effort and defect? Delphi & automatic localizations ? Seek fully automated procedure Returns regression models Our users want actions, not trends. Navigators, not maps Clusters on naturally dimensions What about synthesized dimensions? 40 Train and test on local clusters Why not train on superior neighbors (the envy principle) Tested via cross-val Train on neighbor, test on self. No 10*10-way cross val
- 41. S.P.A.C.E = S.P ADD CONTRAST ENVY (A.C.E.) . Contrast set learning (WHICH) Fuzzy beam search First Stack = one rule for each discretized range of each attribute Repeat. Make next stack as follows: • Score stack entries by lift (ability to select better examples) • Sort stack entries by score • Next stack = old stack • plus combinations of randomly selected pairs of existing rules • Selection biased towards high scoring rules Halt when top of stack’s score stabilizes Return top of stack 41
- 42. WHY ADD CONSTRAST ENVY? Search criteria is adjustable • See Menzies et al ASE journal 2010 Early termination S.P. A.C. Turhan et al. (2009) Me et al, ASE, 2011 E Not scalable Near linear time processing No generalization to report to users Use rule learning S.P. A.C Cuadrado et .al (2007) Me et al, ASE, 2011 .E. Only one data set Need more experiments Just effort estimation Why not effort and defect? Delphi & automatic localizations ? Seek fully automated procedure Returns regression models Our users want actions, not trends. Navigators, not maps Clusters on naturally dimensions What about synthesized dimensions? Train and test on local clusters Why not train on superior neighbors (the envy principle) Tested via cross-val Train on neighbor, test on self. No 10*10-way cross val 42 12/1/2011
- 43. DATA FROM HTTP://PROMISEDATA.ORG/DATA Find (25,50,75,100)th percentiles of class values • in examples of test set selected by global or local Express those percentiles as ratios of max values in all. Effort reduction = { NasaCoc, China } : COCOMO or function points Defect reduction = { lucene, xalan, jedit, synapse,etc } : CK metrics(OO) When the same learner was applied globally or locally • Local did better than global • Death to generalism As with Cuadrado ‘07: local better than global (but for multiple effort and defect data 43 12/1/2011 sets and no delphi-localizations)
- 44. EVALUATION S.P. A.C. COW Turhan et al. (2009) Me et al, ASE, 2011 E Not scalable Near linear time processing No generalization to report to users Use rule learning S. A.C CO Cuadrado et .al (2007) Me et al, ASE, 2011 P. .E. W Only one data set Need more experiments Just effort estimation Why not effort and defect? Delphi & automatic localizations ? Seek fully automated procedure Returns regression models Our users want actions, not trends. Navigators, not maps Clusters on naturally dimensions What about synthesized dimensions? Train and test on local clusters Why not train on superior neighbors (the envy principle) 44 12/1/2011 Tested via cross-val Train on neighbor, test on self. No 10*10-way cross val
- 45. ROADMAP Some comments on the state of the art • Why so much SE + data mining? • Why research SE + data mining • But is data mining relevant to industry • The problem of conclusion instability Learning local • Globalism: learn from all data • Localism: learn from local samples • Learning locality with clustering (S.P.A.C.E.) • Implications 45 12/1/2011
- 46. IMPLICATIONS: GLOABLISM Simon says, no 46 12/1/2011
- 47. IMPLICATIONS: DELPHI LOCALISM Simon says, no 47 12/1/2011
- 48. IMPLICATIONS: CLUSTER-BASED LOCALISM Simon says, yes 48 12/1/2011
- 49. IMPLICATIONS: CONCLUSION INSTABILITY From this work • Misguided to try and tame conclusion instability • Inherent in the data • Don’t tame it, use it • Built lots of local models 49 12/1/2011
- 50. IMPLICATIONS: OUTLIER REMOVAL Remove odd training items Examples: • Keung & Kitchenham, IEEE TSE, 2008: effort estimation • Kim et al., ICSE’11, defect prediction • case-based reasoning • prune neighboring rows containing too many contradictory conclusions. • Yoon & Bae, IST journal, 2010, defect prediction • association rule learning methods to find frequent item sets. • Remove rows with too few frequent items. • Prunes 20% to 30% of rows. Assumed, assumes a general pattern, muddle by some outliers But my works says “its all outliers”. 50 12/1/2011
- 51. IMPLICATIONS: STRATIFIED CROSS-VALIDATION Best to test on hold-out data • That is similar to what will be seen in the future • E.g. stratified cross validation This work: “similar” is not a simple matter • select cross-val bins via clustering • Train on neighboring cluster • Test on local cluster Why learn from yourself? • If the grass is greener on the other side of the fence • Learn from your better neighbors 51 12/1/2011
- 52. IMPLICATIONS: STRUCTURE LITERATURE REVIEWS ? 52 12/1/2011
- 53. IMPLICATIONS: SBSE-1 (A.K.A. LEAP, THEN LOOK) When faced with a new problem • Jump off a cliff with roller skates and see where you stop. That is: • Define objective function and use it to guide a search engine. 53 12/1/2011
- 54. IMPLICATIONS: SBSE-2 (LOOK BEFORE YOU LEAP) • Split • data on independent variables • Prune • leaf quadrants using dependent variables • Contrast. • Sort data in each cluster • Contrast intra-cluster data between good and bad examples • Add Envy: • For each cluster C1… • Find C2; i.e. the neighboring clustering you most envy • Apply C2’s rules to C1 54 12/1/2011
- 55. THE COW DOCTRINE • Seek the fence where the grass is greener on the other side. • Learn from there • Test on here • Don’t rely on trite definitions of “there” and “here” • Cluster to find “here” and “there” 55 12/1/2011
- 56. 56 12/1/2011