June 17, 2004 Allison Wiley, Megan Dameron, Sarah Grinnell Brian Brophy, Dick Coleman, Jessica Summerville Transitioning from Parametric to Buildup Estimates.

Slides:



Advertisements
Similar presentations
Statistical Techniques I EXST7005 Multiple Regression.
Advertisements

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display. 1 ~ Curve Fitting ~ Least Squares Regression Chapter.
Chapter 4: Trees Part II - AVL Tree
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and l Chapter 12 l Multiple Regression: Predicting One Factor from Several Others.
Simultaneous inference Estimating (or testing) more than one thing at a time (such as β 0 and β 1 ) and feeling confident about it …
Correlation and regression Dr. Ghada Abo-Zaid
SCEA June 2000 JRS, TASC, 5/7/2015, 1 BMDO Cost Risk Improvement in Operations and Support (O&S) Estimates J. R. Summerville,
Copyright (c) 2004 Brooks/Cole, a division of Thomson Learning, Inc. Chapter 13 Nonlinear and Multiple Regression.
HSRP 734: Advanced Statistical Methods July 24, 2008.
LINEAR REGRESSION MODEL
Chapter 3 Producing Data 1. During most of this semester we go about statistics as if we already have data to work with. This is okay, but a little misleading.
Section 4.2 Fitting Curves and Surfaces by Least Squares.
Data Structures Hash Tables
Statistics for Managers Using Microsoft® Excel 5th Edition
Data Sources The most sophisticated forecasting model will fail if it is applied to unreliable data Data should be reliable and accurate Data should be.
Correlation A correlation exists between two variables when one of them is related to the other in some way. A scatterplot is a graph in which the paired.
Chapter 5 Forecasting. What is Forecasting Forecasting is the scientific methodology for predicting what will happen in the future based on the data in.
1 Econometric Load Forecasting Peak and Energy Forecast 06/14/2005 Econometric Load Forecasting Peak and Energy Forecast 06/14/2005.
Basic Business Statistics, 11e © 2009 Prentice-Hall, Inc. Chap 15-1 Chapter 15 Multiple Regression Model Building Basic Business Statistics 11 th Edition.
Stat 112: Lecture 9 Notes Homework 3: Due next Thursday
Week 14 Chapter 16 – Partial Correlation and Multiple Regression and Correlation.
Chapter 6 (cont.) Regression Estimation. Simple Linear Regression: review of least squares procedure 2.
Business Forecasting Chapter 5 Forecasting with Smoothing Techniques.
Copyright ©2011 Pearson Education 15-1 Chapter 15 Multiple Regression Model Building Statistics for Managers using Microsoft Excel 6 th Global Edition.
Cost Analysis and Classification Systems
Forecasting using trend analysis
Objectives of Multiple Regression
Time-Series Analysis and Forecasting – Part V To read at home.
Copyright ©2011 Pearson Education, Inc. publishing as Prentice Hall 15-1 Chapter 15 Multiple Regression Model Building Statistics for Managers using Microsoft.
Graphical Analysis. Why Graph Data? Graphical methods Require very little training Easy to use Massive amounts of data can be presented more readily Can.
© 2004 Prentice-Hall, Inc.Chap 15-1 Basic Business Statistics (9 th Edition) Chapter 15 Multiple Regression Model Building.
1 Chapter 8 Sensitivity Analysis  Bottom line:   How does the optimal solution change as some of the elements of the model change?  For obvious reasons.
1 CSI5388: Functional Elements of Statistics for Machine Learning Part I.
1 Ch 3: Forecasting: Techniques and Routes. 2 Study objectives After studying this chapter the reader should be able to: Evaluate the suitability of several.
Sampling Methods. Definition  Sample: A sample is a group of people who have been selected from a larger population to provide data to researcher. 
Chapter 9 Analyzing Data Multiple Variables. Basic Directions Review page 180 for basic directions on which way to proceed with your analysis Provides.
DAVIS AQUILANO CHASE PowerPoint Presentation by Charlie Cook F O U R T H E D I T I O N Forecasting © The McGraw-Hill Companies, Inc., 2003 chapter 9.
Lesson Multiple Regression Models. Objectives Obtain the correlation matrix Use technology to find a multiple regression equation Interpret the.
Time series Decomposition Farideh Dehkordi-Vakil.
1 Chapter 10: Introduction to Inference. 2 Inference Inference is the statistical process by which we use information collected from a sample to infer.
CHAPTER 12 Descriptive, Program Evaluation, and Advanced Methods.
This material is approved for public release. Distribution is limited by the Software Engineering Institute to attendees. Sponsored by the U.S. Department.
Lecture 02.
Right Hand Side (Independent) Variables Ciaran S. Phibbs June 6, 2012.
Copyright © Cengage Learning. All rights reserved.
Right Hand Side (Independent) Variables Ciaran S. Phibbs.
Robust Estimators.
1 Another useful model is autoregressive model. Frequently, we find that the values of a series of financial data at particular points in time are highly.
Sampling Design and Analysis MTH 494 Lecture-22 Ossam Chohan Assistant Professor CIIT Abbottabad.
Statistics for Managers Using Microsoft Excel, 4e © 2004 Prentice-Hall, Inc. Chap 14-1 Chapter 14 Multiple Regression Model Building Statistics for Managers.
Analogy Technique Chapter Analogy - Method Comparative analysis of similar systems Adjust costs of an analogous system to estimate the.
“Teaching”…Chapter 11 Planning For Instruction
Basic Business Statistics, 10e © 2006 Prentice-Hall, Inc. Chap 15-1 Chapter 15 Multiple Regression Model Building Basic Business Statistics 10 th Edition.
Statistics for Managers Using Microsoft Excel, 4e © 2004 Prentice-Hall, Inc. Chap 14-1 Chapter 14 Multiple Regression Model Building Statistics for Managers.
The Mixed Effects Model - Introduction In many situations, one of the factors of interest will have its levels chosen because they are of specific interest.
Forecasting is the art and science of predicting future events.
Measures of Central Tendency: Just an Average Topic in Statistics.
Welcome Parents! FCAT Information Session. O Next Generation Sunshine State Standards O Released Test Items O Sample Test.
Demand Management and Forecasting Chapter 11 Portions Copyright © 2010 by The McGraw-Hill Companies, Inc. All rights reserved. McGraw-Hill/Irwin.
Chapter 11 – With Woodruff Modications Demand Management and Forecasting Copyright © 2010 by The McGraw-Hill Companies, Inc. All rights reserved.McGraw-Hill/Irwin.
Chapter Nine Hypothesis Testing.
Chapter 15 Multiple Regression Model Building
Week 14 Chapter 16 – Partial Correlation and Multiple Regression and Correlation.
8/31/16 Today I will discover facts about an element
Office of Education Improvement and Innovation
DSS-ESTIMATING COSTS Cost estimation is the process of estimating the relationship between costs and cost driver activities. We estimate costs for three.
15.1 The Role of Statistics in the Research Process
Scramble for Africa DBQ Writing Workshop.
AP U.S. History Exam Details
Forecasting Plays an important role in many industries
Presentation transcript:

June 17, 2004 Allison Wiley, Megan Dameron, Sarah Grinnell Brian Brophy, Dick Coleman, Jessica Summerville Transitioning from Parametric to Buildup Estimates

Outline Introduction Explanation of the Challenge Developing a Buildup Estimate Parametric Pullout Using a Toy Problem

Introduction Purpose of this Presentation – Discuss the challenge presented in transitioning from parametric to buildup cost estimates – Present some of the ideas and approaches considered to address the challenge – Generate ideas and discussion with the audience in order to advance our thinking

The Challenge The transition from a parametric to a buildup cost estimate – Develop cost estimates specific to a subsystem – Calculate the appropriate amount to remove from the parametric estimate for the system, in order to insert the buildup estimate Utility – Provides detailed information so that the cost of operations and support or specific production items can influence the system design – Buildups are instructive and can lead to improvements in parametric estimates

Developing a Buildup Estimate Why the transition? – There could be subsystems that behave in a way that is inherently different from the legacy subsystems in the cost data – It may not be possible to estimate the entire system at a detailed level, but there may be detailed information available for one or more subsystems Defining “buildup” – In this presentation, the word buildup refers to an estimate specific to a subsystem – Intuitively the authors had in mind traditional buildup estimates as well as analogies – Parametric estimates of a subsystem alone could also be considered

Parametric Pullout The term “parametric pullout” refers to the amount of the parametric estimate that is attributable to a specific subsystem – in short, the amount of the parametric estimate that is “pulled out” so that the buildup estimate may be inserted. The idea of parametric pullout and some of the methods considered to address this problem are explained in the next several slides with the use of a toy problem.

Parametric Pullout Introduction to the toy problem – You have a CER that estimates the cost of a car based on its weight – To reduce confusion, let’s call the car the TE-4 – You have recently developed a buildup estimate for the transmission alone, and you hope to eventually have buildup estimates for most of the major parts of the car – Let’s call the transmission the BE-2 – The challenge: Incorporate the buildup estimate for the BE-2 into the total cost estimate for the TE-4

Parametric Pullout (cont’d) Method 1: Historical Percentage – If the transmission of a car is historically 8% of car cost, remove 8% of the TE-4 estimated cost – Advantages: Easy to execute, also works for costs estimated using historical averages instead of CERs – Disadvantages: Method requires specific historical data that may not be available in all cases

Parametric Pullout Method 1: Historical Percentage – Toy Problem Historical Data $25,680 - $4,015 + $3,650 $25,680 x 15.6%

Parametric Pullout Method 1: Historical Percentages – Concerns – Consider the case where the BE-2 comprises a much larger percent of TE-4 weight than any legacy transmission and car. There is a concern that since the BE-2 weighs so much more, the percentage method may cause the removal of an inadequate amount of cost.

Tip: It is frequently best to take the CER result from the parameters of the whole system, and the CER result from the parameters of the whole system minus the subsystem and note the difference. This is especially important when CERs have multiple variables and/or are non-linear. Parametric Pullout Method 2: Parameter-based – Use the parameters of the CER to determine the correct piece to pullout – Advantage: Requires little data and is easily executed – Disadvantage: Subsystem may not have parameters comparable to the parent system. For example, there is not an intuitive way to pull the cost of a tire out of a car CER that uses the weight of the car’s electrical system.

Parametric Pullout Method 2: Parameter-based – Toy Problem – Run the car CER on the weight of the entire car – Run the car CER on the weight of the entire car, less the weight of the transmission – Subtracting the latter from the former yields the amount to be removed – Add in the buildup estimate for the transmission

Parametric Pullout Method 2: Parameter-based – Concerns – In a CER for a total system model, the parameters may mask or interplay with other parameters – For example, in a car CER based on electrical system weights, the electrical system weight acts as a proxy for other system weights (like the tires, the seats, and the frame of the car). – Intuitively, removing a few pounds of electrical system weight removes an electrical component, but mathematically it also removes all of the frame of the car that supports the electrical component. – It is important to understand the meaning of the CER.

Parametric Pullout Method 3: Obtain new CER – Example: – Y = legacy car cost – legacy transmission cost – X = (legacy car weight – legacy transmission weight)) – Results should yield reasonable F and t statistics, and a reasonable R 2 – Use TE-4 minus BE-2 weights in the new equation – Subtract the new result for a “transmission-less car” from the existing result for all of TE-4 – this is the pullout amount – Advantages: Provides a good check of the existing CER – Disadvantages: Requires significant resources and data to accomplish, some CERs will not respond well to this method without being completely reworked

Toy Problem (cont’d) Method 3: Re-run CER – Return to the original data that produced the CER – Subtract the cost of the transmission from the cost of the car, and the weight of the transmission from the weight of the car Car Cost = 4, (Car Weight) R 2 = 0.99 t and F significant Estimated weight of new car: 4,000 lbs Estimated cost of new car: $23,496

Toy Problem (cont’d) Method 3: Re-run CER – Rerun the regression (ensure that t and F statistics are still significant, and R 2 is reasonable) – Run the new CER on the weight of car to be estimated, minus the weight of the transmission “Transmission-less Car”Cost = 4, (Car Weight – Transmission Weight) R 2 = 0.99 t and F significant Estimated weight of new car w/o transmission: 4,000 – 500 = 3,500 lbs Estimated cost of new car w/o transmission: $19,791

Parametric Pullout (cont’d) Method 3: Obtain new CER (cont’d) – Concerns – This method may not behave intuitively – Changes are observed in the coefficients of unchanged parameters (when multiple parameters are present) – What is the expectation for behavior? Case 1: Method 2 and Method 3 yield the same result Car CER Possible “Transmission-less Car” CER If Method 2 and Method 3 are to yield the same result, the new CER must intersect the Car CER at the weight of the “transmission-less car”

Parametric Pullout (cont’d) Method 3: Obtain new CER (cont’d) – Concerns (cont’d) – What is the expectation for behavior? Car CER Possible “Transmission-less Car” CER Case 2: Method 3 removes more cost than Method 2 but the new CER behaves intuitively

Parametric Pullout (cont’d) Method 3: Obtain new CER (cont’d) – Concerns (cont’d) – What is the expectation for behavior? Case 3: Method 3 removes less cost than Method 2, the new CER may not behave intuitively Car CER Possible “Transmission-less Car” CER

Parametric Pullout (cont’d) Method 3: Obtain new CER (cont’d) – Concerns (cont’d) – What is the expectation for behavior? Case 4: The new CER yields an estimate higher than the existing CER – the method does not produce a parametric pullout Car CER Possible “Transmission-less Car” CER

Parametric Pullout Second order effects – In some models, particularly in operating and support cost, some elements are estimated using a parametric relationship to another cost element – If this is the case, it is important to keep careful track of order of operations, etc. … to be certain that the appropriate results are captured Warning: It is easy to get caught in the trap of thinking that the element which is being pulled out and put back in is “causing” changes to other elements in the model. This is not the case - second order changes to elements are attributable to a total change in the cost element value, not to the system being incorporated into the model.

Parametric Pullout Complexities beyond the toy problem – The toy problem demonstrates an example of a CER that is linear and uses just one variable – There are many complexities of the problem that occur in non-linear and/or multivariable CERs

Lessons Learned Buildups are useful because of the insight they may provide. The following has occurred during the authors’ experience with this topic: – A significant improvement in understanding of the nuances of a main source of historical data resulted – Led to an examination of several existing CERs and led directly to improvements in at least one – Allowed system designers to focus future cost and CAIV resources on specific areas of interest – The details of the process generated discussions with the system designers that provided further insight into design and cost issues of all types