outlook sunny rain 1 rain temp 85 80

outlook sunny, rain, 1 rain, temp 85, 80, 72, 65, 71, 70, 68, rain, 75, sunny, overcast, NASA OSMA SAS '01 Sept. 5 -7, 2001 69, 75, 83, 64, 72, 81, humidity wind hours on course 86, 90, 95, 70, 96, 80, false, true, false, none none some 70, 88, 65, 90, 75, false, true, false, lots lots Humidity= 90. . 96% Harvest: try to improve the scores For “good”=lots do For “bad” {some, none} do Deltaf = X*Y number of examples X= #attribute. range in “good” - #attribute. range in “bad” Y = score(“good”) - score(“bad”) Outlook= overcast 2022 -02 -15

BTW- deltaf >> tree query (faster, simpler). Harvesting from data: golf 2 NASA OSMA SAS '01 Sept. 5 -7, 2001 If no change, Then lots of golf 6/(6+3+5)= 43% If outlook=overcast, Then lots of golf 4/(4)= 100% assumes causality, controllability Least change: bribe DJs to lie about the weather If humidity=90. . 96 Then lots of golf 2/(4)= 50% Least monitor: watch the humidity - alert if rising over 90% 2022 -02 -15

n Inputs = all ranges Harvesting from seed 2: COCOMO-R 3 • Sced=4: Time =160% of schedule • Pcap=4: programmer capability > 90 th percentile • Pmat=2: CMM level 2 n. Inputs = KC 1= changes 1 U now 1 NASA OSMA SAS '01 • Sced=2: Time = 100% of schedule • Acap=2: analyst capability 55 th percentile Sept. 5 -7, 2001 2022 -02 -15

Runs= 35228 Harvesting from seed 3: a qualitative model 4 n Qualitative constraint model of a circuit : n n E. g. bulbs may be blown/ok Score each run by # shining light bulbs (max=9) Run I: Find top treatment T learnt from run I. T must be: n n n NASA OSMA SAS '01 n Sept. 5 -7, 2001 Runs= 3264 And close Sw 1 c Runs= 648 And close Sw 3 c (not done) Runs= 32 9 switches, 9 bulbs, 3 batteries Many unknowns: n Leave Sw 2 c open Acceptable to users And Possible Constrain run I+1 with T 1 question culls 90% of options 2022 -02 -15

Generality 5 TAR 2 works when domains generates a small number of NASA OSMA SAS '01 outstandingly large deltaf values Sept. 5 -7, 2001 2022 -02 -15