Name_Uasinn Suyluecdev Dae_Lee % Zoly Salt Lake &® Community a Linear Least Squares Approximation Lab College or Fitting a Polynomial Curve to a Set of Data Points. Mathematics Department dead , Leen, Gn MO af Part 1 Introduction Oy ‘One of the common situations that arise in the real world is as follows. You have a set of data that partially describe a given situation, but you either want to get an educated guess of a future value, or approximate data that lay between the measured data you have. In either case, what you need to do is find a curve (or function) that “best fits” the data. Once you have this function, you can evaluate it at different spots (inputs) and get approximations you are after (outputs). ‘So, what is meant by the “best fit?” One of the most well known ways is the method called Least Squares. To illustrate, let’s start with Figure A, a graph consisting of plotted points (the initial data). Figure A Figure B If we try to find the best line to fit this data using least squares, we would get the line in Figure B. To get an idea how we would measure how close this line fits, we would look at the vertical error associated with each point, as illustrated in figure C. Now, if we square each error and add up the total square errors, we get the measurement of how well this line fits. Graphically, the square errors are illustrated in figure D. a re | Figure C Figure D Different lines will yield a different sum of the square areas, and so the “best fit” is the one that has the “Least Squares”. The reason we use the square of the error is best understood with some knowledge of Calculus, so sullice to say this is how we do itThe following graph includes the 4 points and our linear approximation. Now it is your turn, Part II Vehicle Crashes ‘The following facts are based on analysis of data from the U.S. Department of Transportation's Fatality Analysis Reporting System (FARS). Year___Rate Motor vehicle crash deaths per 100,000 people, 1975-2005: 1975 20.6 1980 22.5 1985 18.4 1990 17.9 1995 15.9 2000. 14.9 From this graph, we can see that out of 100,000 people, 20.6 people died in a motor 2005147 vehicle crash in the year 1975, Use the method of least squares to find the best linear fit y= mx+b a) Write your system Ax=b b) Write what you have for A" c) Write your simplified normal equations ( A’ Ax = A"b) 4) Write your least squares solution. (RREF is a good general method for this lab) e) Graph the points and the linear approximation on the same coordinate axes. To graph your data and the approximate solution on the same coordinate axes, you can use maple, excel, virtual TI and a screen capture program or by hand with graphing paper. See last pages for details. If you find other methods, please share. 1) Use your approximation to estimate the number of deaths per 100,000 you expect in the year 2010.Part IV LINEAR or QUADRATIC? Golf Ball Distance When released on a hill with a constant slope, a golf ball falls at a speed determined by gravity, the slope, and friction from the grass. The below data table shows you the results of a specific experiment concerning how long it took in seconds for a golf ball to travel the given distance, in feet. What you need to do is to determine from the data whether you feel a linear model or a quadratic model would be a better fit. You can do this by finding both, or by sketching the data and doing a visual estimation. Once you determine which one you prefer, find that model and graph the resulting curve, along with the given data. a) b) °°) a ©) fy 2 Determine whether you will use y=mx+b or y=ax" +bx+c. Write your system Ax Write what you have for A’ Write your simplificd normal equations ( AT Ax = A*b) Write your least squares solution. Graph the points and the quadratic approximation on the same coordinate axes. Use your approximation to estimate how long it would take for a golf ball to travel 150 feet. time/sec | dist/feet 456 10 657 20 8.14 30 9.48 40. 10.65 50. 11.85 60. 12.46 70Part OL Vehicle Crashes 4a) Qstem Axe b Te beret bz 206 msfo 1 aes Grr lor 225 na 22s (Cea A ac iiees esd _ 1s F my! _ feta Ooms F ro lip] [8 aSm+ ema 439 3D) oer | 2 | wo o acer Levit b) AT ins” Bees Peer | carer oman cs os 10 15 20 29 a Shavting <¢ Zev Weve ae ata teen 5 ©) ptags Ate Simplified eaeiionisi go) 2) el ot m)= > 5 wis 22 30] [roc] vs fees | Ue ee aaa San aie ea ea ae io} [o | is | 3.4 Ber isa | aan Ca eas 47 | Simplified normal eqvations Taras os) fm], [1676 s] i lez | oe \ Lo J Hee eae d. Solurcon (mh) [t.-3oUVWNZS © Le) bn ne * ye Laouniizse + (12+Number ) ef deaths mn 2010 4 1 BoNTqagex - 12.116 GE anes: Y= i pougzee (35) - 12.116 4 33-6 So 33.0 dtats POY 100,000 Ww L010 Part OD A) System Axe B ae 2 86025 30 40 4S 4o 48 a4 | BO Soaeeiseia| eo 66 z00 | HO aR) Oe i Wee a) A zo 30 40 50 ae. 4S A Se Aiea act ra eS, Zo 30 HO so ee | 43 aa les Stepping clistance oe 27] [zo iH | uss_| [a0 50 co Jo 30 | {49 | | =3 el | 145 yss} | 293 ree 2 4o 4s os ee 24 eo) 38 | uss yHoppin, distance. Cont... c.) Simplify [B00 15,450 cuazo | [a | (80, 340 | eo ae \b |= |e asi | \eu,220 70,486 434,252 | Le] Lam eset D-) Solution ha Py coreg \4, 82852334 \e| [6.09919 54 545° ys T4s34x? +Ciss2ty) + O79 B) Graph B) skip mares @ 222 &. long 2 Luss (22)*- |. $8 27(ezz) t,o 62 mehPart TT A) ys metb BAK=b ae El 6 (2% \a[= ie 4] lea aus{ | be 32 | 85 {ee| [i244 Laos ©) Ar [Hse 53 SM 3.48 iocs LES wae] D) Atay 2 ath Smelted [ozeussi)(<] = [2n.i] Z) ae 4624925184 x FLY Graph SQ) Hous Leng ast H take fev He batt to trawel iso 4? Y= 4.684325 134 (iso) (Aaa Se nie Ret