In grade school I found that shortly after learning a new word I'd almost certainly hear it in conversation. Similarly, after writing the Goldman Algorithm post I came across an article that was very similar. Researchers at Emory created a model to predict a patient's chance they will suffer a heart failure within 5 years. Unlike the Goldman Algorithm which determines how to care for an individual already suffering from chest pains this model can help prevent or delay that trip to the ER... but will it? Before I go any further I applaud the efforts, however, this opens up an interesting question into human's risk tolerance. Will those who are diagnosed with a very high chance of having a heart failure change their behavior to prevent or delay it? Like Deal or No Deal, contestant in this game are also likely to make irrational choices (or lack of good choices). As a skeptic, I see the US population's obesity trends increasing in spite of the risks being well known, and I doubt this will make a significant change. On the other hand, given the right marketing, this model may be the start of a public awareness campaign that changes behaviors for the better, similar to the recent anti-smoking campaign.
Link to BUS 650: Application of models and risk tolerance
Link to article: http://www.eurekalert.org/pub_releases/2009-03/eu-hfr032609.php
Link to obesity and smoking reference:
http://www.cdc.gov/nccdphp/dnpa/obesity/trend/maps/
http://www.cdc.gov/mmwr/preview/mmwrhtml/mm5320a2.htm#fig2
Monday, April 20, 2009
Sunday, April 19, 2009
Application of Monte Carlo Analysis
4x4 Brake Design and Tolerance Study
I used a Monte Carlo simulation several years ago when I faced with a unique design challenge with multiple constraints and variable inputs. I was responsible for developing the brakes for a new four wheel drive utility vehicle. The challenge was to properly set the height adjustment for the brake pedal. The vehicle uses hydraulic brake system with two semi-independent systems so if one is damaged and inoperable the operator will still be able to stop the vehicle with the second. If one system fails the brake pedal will travel roughly half its stroke before any stopping force is generated. I had a balancing act, I had to position the pedal so that it could generate the pressure to stop the car in any situation before it bottomed out on the floorboard or was uncomfortably close to the driver. Also, each part involved has its normal manufacturing tolerances that must be considered.
Typically most design issues can be resolved using 3D computer aided drafting practice and a quick max and min analysis. In this case, potential failures resulting in a poor design were serious and my constraints were so tight that I was forced to use more advanced models.
The model I created used outputted the location of the pedal in the resting and fully depressed positions along with the potential interference with the floorboard and clearance to the driver's reach called the "zone of comfort". I was able to alter dimensions and tolerances for every part in the system to locate the pedal in the ideal position. The conclusion of the analysis resulted in a few design changes and a pedal that is optimally located for comfort and safety. The Monte Carlo modeling approach took plenty of time to develop but in the long run I was able to cut design time and cost.
As a rookie blogger I'm not sure how to upload the actual file so I clipped pictures of different sections of the analysis for your entertainment. If you'd like to see the actual model email me.
Deal or No Deal?
Why People Make Irrationally Risky Economic Decisions
The article "Why Game Shows Have Economists Glued to Their TVs" was published in the Wall Street Journal. The article contends that studying the contestants of game shows can help unlock some of the secrets behind why it seems people make illogical choices while making decisions. Evidence for studying risk tolerance and decision making can come from strange sources. In this case its the budget and setting of the source that enables Economists to see how people make decisions in the face of uncertainty with large economic impact to them. One interesting insight mentioned is that contestants tended to be more conservative following a large morale defeat. The application of this evidence can be used in almost any setting and may help unlock effective but poorly understood tactics of negotiations such as anchoring.
Link to BUS 650: Game shows like Deal or No Deal are essentially real life experiments into human risk tolerance for large monetary rewards. The game shows budget allows those studying decision modeling to get actual insight to questions previously only posed theoretically.
Article is available from Emory's library (password is required)
http://proquest.umi.com.proxy.library.emory.edu/pqdweb?did=967212221&sid=2&Fmt=3&clientId=1917&RQT=309&VName=PQD
The article "Why Game Shows Have Economists Glued to Their TVs" was published in the Wall Street Journal. The article contends that studying the contestants of game shows can help unlock some of the secrets behind why it seems people make illogical choices while making decisions. Evidence for studying risk tolerance and decision making can come from strange sources. In this case its the budget and setting of the source that enables Economists to see how people make decisions in the face of uncertainty with large economic impact to them. One interesting insight mentioned is that contestants tended to be more conservative following a large morale defeat. The application of this evidence can be used in almost any setting and may help unlock effective but poorly understood tactics of negotiations such as anchoring.
Link to BUS 650: Game shows like Deal or No Deal are essentially real life experiments into human risk tolerance for large monetary rewards. The game shows budget allows those studying decision modeling to get actual insight to questions previously only posed theoretically.
Article is available from Emory's library (password is required)
http://proquest.umi.com.proxy.library.emory.edu/pqdweb?did=967212221&sid=2&Fmt=3&clientId=1917&RQT=309&VName=PQD
Who Makes the Decisions
Because no portfolio could be complete without a Dilbert...
I read this Dilbert cartoon in an article about an employee that posted it at work then promptly got fired. The best part of the Dilbert cartoon is its ability to use satire to point out true conflicts we experience every day in the business word. This cartoon points out a structural weakness in how decisions are made in a hierarchy reporting structure. Far too often those who know and are effected the most by a decision are involved only indirectly in the decision that ultimately comes from above.
Link to BUS650: One dimension of decision making that we have not discussed in class is... "am I the right person to be making a decision?" When is it more effective to have a highly skilled (ninja) decision maker with limited exposure to the issue making decisions as opposed to less skilled decision maker who intimately knows the problem? I do know I couldn't land a plane safely in the Hudson River (see my previous post)
I read this Dilbert cartoon in an article about an employee that posted it at work then promptly got fired. The best part of the Dilbert cartoon is its ability to use satire to point out true conflicts we experience every day in the business word. This cartoon points out a structural weakness in how decisions are made in a hierarchy reporting structure. Far too often those who know and are effected the most by a decision are involved only indirectly in the decision that ultimately comes from above.
Link to BUS650: One dimension of decision making that we have not discussed in class is... "am I the right person to be making a decision?" When is it more effective to have a highly skilled (ninja) decision maker with limited exposure to the issue making decisions as opposed to less skilled decision maker who intimately knows the problem? I do know I couldn't land a plane safely in the Hudson River (see my previous post)
Presidential Bracket
Barack's Bracket
The bracket and results: http://games.espn.go.com/tcmen/entry?entryID=2813746
I was in line at a paint store waiting for the materials for my weekend chore to be mixed when ESPN's special was aired. I can't explain why it captivated me but I was drawn to the interview with President Barack Obama. It may have resulted from this being the first time I've heard anybody in his capacity rattle of sporting commentary like Dick Vitale, but more likely I was amused by his decision making process to arrive at his selections.
In contrast to the mathematical model presented in class, with alleged good results, the president decision process was wavering between instinct and evidence. This was evident by the discussion Barack had describing how he decided on each pick. He was quoted saying "VCU, I think, has been playing strong, and I hate to say this, because my brother-in-law is in the Pac-10 right now, but the Pac-10 has been looking pretty weak this year," Obama said. "I like that as an upset." In this case he identified an upset without using facts such as strength of schedule, overall record, or individual match-ups. On the other hand Barack did have an instinctual advantage after playing hoops with his correct selection for the winner, UNC. Overall, Barack's bracket correctly identified over 80% of the winners. Barack's bracket was a good example of how people actually make decisions. In this case, examining the method of the decisions is particularly important given the subject, the President of the United States.
Links to BUS650: Keeney's model on how people should decide is a general tool that could help define why different people approach problems differently. In this case, Barack's selections, based heavily off of instinct, likely reflected that this wasn't the most pressing issue on his agenda. On the other hand, I would expect Andy Katz's (ESPN expert) decision process to be completed in a different manner considering that he is an expert whose job relies on accurate and in depth analysis.
Resources and links:
The bracket and results: http://games.espn.go.com/tcmen/entry?entryID=2813746
I was in line at a paint store waiting for the materials for my weekend chore to be mixed when ESPN's special was aired. I can't explain why it captivated me but I was drawn to the interview with President Barack Obama. It may have resulted from this being the first time I've heard anybody in his capacity rattle of sporting commentary like Dick Vitale, but more likely I was amused by his decision making process to arrive at his selections.
In contrast to the mathematical model presented in class, with alleged good results, the president decision process was wavering between instinct and evidence. This was evident by the discussion Barack had describing how he decided on each pick. He was quoted saying "VCU, I think, has been playing strong, and I hate to say this, because my brother-in-law is in the Pac-10 right now, but the Pac-10 has been looking pretty weak this year," Obama said. "I like that as an upset." In this case he identified an upset without using facts such as strength of schedule, overall record, or individual match-ups. On the other hand Barack did have an instinctual advantage after playing hoops with his correct selection for the winner, UNC. Overall, Barack's bracket correctly identified over 80% of the winners. Barack's bracket was a good example of how people actually make decisions. In this case, examining the method of the decisions is particularly important given the subject, the President of the United States.
Links to BUS650: Keeney's model on how people should decide is a general tool that could help define why different people approach problems differently. In this case, Barack's selections, based heavily off of instinct, likely reflected that this wasn't the most pressing issue on his agenda. On the other hand, I would expect Andy Katz's (ESPN expert) decision process to be completed in a different manner considering that he is an expert whose job relies on accurate and in depth analysis.
Resources and links:
Green Atlanta
Hyatt Regency
A friend posted an article on actions a local hotel took to become more environmentally friendly. The hotel has increased recycling and started composting food waste. The article also mentioned that the actions implemented either saved cost or were cost neutral. What struck me most was that the actions were taken only after loosing convention business to another city. I applied our learning on risk and utility to understand why it took threat of loosing business to take on cost saving actions.
The tradeoff the hotel faced reminded me of the discussions we had involving risk and also the article “Valuing Billions of Dollars”. The decline of natural resources and the environment is considered to be uncertain, indirect, non-economic, and a future cost. Each of these types of costs are typically discounted compared to the money lost by loosing a customer to a competitor. In this case the lost business had an enormous perceived value resulting from Atlanta’s position as a leading convention hub. Conventions are big business to Atlanta, in 2006 convention attendees spent an estimated $2.4 billion dollars.
Link to BUS650: Risk, utility, and tradeoffs were made by the Hyatt to retain convention business.
A friend posted an article on actions a local hotel took to become more environmentally friendly. The hotel has increased recycling and started composting food waste. The article also mentioned that the actions implemented either saved cost or were cost neutral. What struck me most was that the actions were taken only after loosing convention business to another city. I applied our learning on risk and utility to understand why it took threat of loosing business to take on cost saving actions.
The tradeoff the hotel faced reminded me of the discussions we had involving risk and also the article “Valuing Billions of Dollars”. The decline of natural resources and the environment is considered to be uncertain, indirect, non-economic, and a future cost. Each of these types of costs are typically discounted compared to the money lost by loosing a customer to a competitor. In this case the lost business had an enormous perceived value resulting from Atlanta’s position as a leading convention hub. Conventions are big business to Atlanta, in 2006 convention attendees spent an estimated $2.4 billion dollars.
Link to BUS650: Risk, utility, and tradeoffs were made by the Hyatt to retain convention business.
Resource links
- Article: http://www.facebook.com/ext/share.php?sid=74967942439&h=L5jGI&u=JOrcV&ref=nf
- Metro Atlanta Business-Convention Visitor Profile http://www.atlanta.net/ACVB/annualReports.html
The Goldman Algorithm
The Goldman Algorithm for Chest Pain
I have plenty of idle time while on interstate 20 between Augusta and Atlanta so I often listen to audiobooks. I was listening to Blink, the book by Malcom Gladwell, on the way home from class when I heard about the Goldman Algorithm. The author was using it as evidence that in certain occasions better decisions can be made from less information. Beyond this, the Goldman Algorithm is evidence that a simple well structured decision model can reach better result than experts. According to the book the algorithm yielded a 93% accuracy rate over a two year period while doctor’s diagnoses were accurate between 75-89% of the time. The more accurate diagnosis enabled hospitals to assign patients to the care unit that best fit their symptoms. This filtered those in the greatest need to the critical care unit more often, freed space in critical care units by reducing the number of patients that did not require that amount of attention, it increased efficiency of the ER, and reduced the overall cost of treatment.
Link to BUS650: the Goldman Algorithm is fundamentally a decision tree. On a larger scale it is an optimization model that reduces healthcare expense by assigning patients with chest pain to the treatment unit (often a constrained resource) that is best equipped to handle their symptoms.
Resource links
Link to Malcom Gladwell’s website:
•http://www.gladwell.com/index.html
Article on the algorithm:
•http://www.medscape.com/viewarticle/417246
•http://www.gladwell.com/index.html
Article on the algorithm:
•http://www.medscape.com/viewarticle/417246
Subscribe to:
Posts (Atom)