Week 12 Question: Due April 12
Answer one of the following questions as a comment on this post.
Submit your contribution by 11:00 AM on Thursday, April 12. It should be insightful (don’t just repeat back facts from the class), but it doesn’t need to be more than a few sentences. Most of all, I want to know what you think!
- Name and describe a business question that you could answer using a decision tree. What data would you collect to perform the analysis? Don’t use an example we’ve covered in class.
- What advice would you give someone regarding how to select the right predictor variables for a decision tree analysis?
David Schuff
#2
Select the predictors that separate best the transaction into two very distinct groups. For example: Graduation= Credits & Student Status. These two predictors describe best which students will graduate this coming year. Choosing Age and Gender will not help divide the student population into distinct groups or give us the estimated graduation date.
1.Name and describe a business question that you could answer using a decision tree. What data would you collect to perform the analysis? Don’t use an example we’ve covered in class.
One business question that one could ask is how likely is someone to delfault on credit card payments or likely to file for bankruptcy in the future. The credit card companies can use this business question as an indicator of whether to issue an individual a card or not. Some of the information that can be collected is gross income, the total amount of outstanding loans. Also, one could collect bill payment history and homeowner’s or rental history. Collecting this information can help give credit card companies a clear idea and strong correlation about the ability of an individual to pay back the purchases made using the credit card.
Question 2: To select the right predictor, my advice is to create as many alternative options for the decision tree analysis. This is because if a decision tree does not carry enough alternative options then the variables can result in wrong prediction. For example, if there are three managers with different ages and incomes then the decision tree should be separate by three managers with different positions, categorized by greater or less than within the specific age and incomes. So yet again, in order to have a better prediction, my advice is to create alternative options for a decision tree analysis.
1. You could figure out if a student will pass a test or not. You could collect hours studied, number of classes attended, overall gpa, graduation year, topic of test, types of questions, and if they passed or not.
#1. You could find out whether or not you should eat out instead of cooking food at home by:
- Time of day (how many stores are open at the specific time of day?)
- How busy you may be that day (do you have time to stop by home or go out to eat?)
- How much money you have (will you have enough money to purchase food?)
- Distance from home
- Distance from food service
- Closest grocery store (is it worth my time to travel for groceries?)
- Marital status (do I have a husband/wife that can cook for me while I’m going home?)
- Children? (depending on if you need to stay home)
2.
Using the classification method in the decision tree analysis, my suggestion is to identify the target categories of primary interest, then select and control which dependent variables are included in the analysis. For example, if we are primarily interested in identifying those individuals in high school most likely to perform well in college, we might select his or her SAT score as the key predictor because it has a closed cohesion with future success in a college. A good school is accounted for in another predictor, but a student’s age seems to not matter very much
#1 One business question you can answer with a decision tree is whether or not someone gets comped in a casino. The Root Node could be Comp Approval. The Child Nodes could be amount gambled over/under $30,000. Then the following node could be amount earned/lost. If a player earned more than a certain amount you could comp him a better room because you would want him to come back in the future. If he lost a certain amount you could comp him a lesser room because he might be more inclined to stay. This decision tree could be used as a tool to help casino employees determine comps.
Question # 1
A business question that you could answer using a decision tree is what insurance company you want to be covered by. You could collect data on what packages each company offers (auto, life, house,). How much each package cost, how they rank in customer satisfaction, number of customers insured (per year), how many customers leave (per year). With this information it will be easier for people to decide which insurance company offers the best overall service.
To find right predictor variables for a decision tree analysis it is important to have an idea about the business question. In order to find an appropriate variable you have to choose a predictor that will be relevant to the problem. Although it is good to base predictor on a statistics (what is a correlation between the result and particular variable) it is important that the statistic can be somehow applied to the problem. Sometime the variable can be irrelevant to the problem but will nicely divide the sample variable. You need to know the business event you trying to explain and pick variable that will be an intrinsic part of it.
Question#1
A business question that could be answer by using a decision tree is would a new product increase the company’s profit. The information for creating this decision tree is based on how similar products are in the current market. Also, the additional information would be the cost of producing the new product and breakeven point.
1.Name and describe a business question that you could answer using a decision tree. What data would you collect to perform the analysis? Don’t use an example we’ve covered in class.
The MIS program within Fox is very competitive and students are required to maintain a high GPA and become actively involved within the department so it can hold its ranking with elite students. The department can assess whether a student will perform well in this program or whether they will likely switch programs by using a decision tree. To perform this analysis, they can collect data such as the students current GPA, their involvement in clubs and organizations, their resume etc.
Question 1
A decision tree can be useful to answer many business questions. One example would be if a person was trying to start up a business and they narrowed the choices down to three, they can produce a decision tree to help make their final decision. They could collect information such as location, profitability, overall start up cost, etc. Putting all these data into the decision tree can better assist the person on deciding which business would be best for them.
A business question that you could answer using a decision tree is if a company should continue selling a product. Some information you can collect could be amount sold at a specific time and/or day the placement of the product, which size of the product sells better etc.
1. Many answers can be found in the decision tree. A clear example is when we are leasing our apartments around campus. Landlords can answer questions such as: Would the student be able to pay the rent on time every month or not? Based on previous data he can answer this question, and then decide whether he should rent the apartment to this student or not.
1. A business question that could be answered using decision trees is if a company pays for an employees commuting expenses, whether they will pay them for their gas mileage or pay them for taking public transport. To determine this factors could include if the employee owns a car, how far they live from their work, which types and how many forms of public transport they would need to take, if the employee has a personal preference, where they would park their car, etc. This could determine whether it will be more costly to simply purchase a transport pass, or to pay for gas mileage.
#1
A business question that could be answered using a decision tree would be for a company’s recruiting process, what candidates they should select to meet for an interview. This information is based on each candidates background and qualifications. Information from resumes and job applications can be used for decision tree analysis to gather information such as GPA, prior experience, involvement in extra curricular activities, and major/minor. This will help make the decision on which candidates would be the best to interview.
1) When a company deliberates on whether to expand or not, it will have to use a decision tree. There could be two or more possible outcomes: profit of x amount or loss of x amount. To perform the analysis, the company has to collect information about interest rates, market attraction, different locations, government laws, tax policies…
Question #1
A business question that could be answered using a decision tree would be deciding which high school student applicants should be accepted into a college/university. Each leaf on the decision tree could be information that students fill out on their application for admission. This would include: GPA, SAT/ACT score, involvement in extra-curricular activities, decided major, and other key factors. This decision tree will help colleges/universities to narrow its pool of applicants, and allow them to find the most suitable students for admittance.
A bixuness question that you could answer using a decision tree is if you are trying to open a luxury restaurent in a city then you may have to collect data on what is the average income about that city, how many people in that city, how many simlilar restaurent in the city and so on . with those information it will help people to find out which city is befitting to open a lexury restaurent .
Question 1
A business question that could be answered using a decision tree would be what type of car to buy for a family. Factors would include family size, budget, amount of travel, mileage, as well as special features on the car. Business owners would be able to better sell cars to families if it had this type of data because they will be able to match-up the families to the correct cars.
#1 You could use the decision tree to estimate or predict the probability of meeting sales target quotas. Depending on the needs of the business they could track by employee or region. Using historical data and previous trends the information provided would give a good indication of the probabilities of meeting the quota or not. Information that they’ll need is past sales history, sales by day or week or even month, hours worked, amount of sales etc.
1
A common business question that can be answered using a decision tree is a businesses decision to expand into a larger facility, increase capacity at a current facility, or pay out excess cash to shareholders. You would need to analyze the market and potential room for expansion as well as all costs associated with that expansion. You would need to analyze the competition and see how much market share would be lost by continuing operations at the current level and not making additional capital expenditures. . You would need to analyze the location of a potential new facility and all transportation costs. By using probabilities of market share gains or losses as well as probabilities of success or failure of each decision on as many branches as possible, the decision tree would better assist in finding the actual cost of each possibility and help you analyze the best possible option.
#1
Which car a person will buy. For example a person married with kids will likely buy SUV. A single person will buy sedan. Person with more income will buy luxury car over someone with low income.
Data collected for this decision would be:
Gender
Married/Single
Level of income
No. of children
A pharmaceutical company that is trying to segment most profitable doctors can use decision trees model. Since only 20% of the customers are responsible for 80% of its profit, the pharmaceutical company needs to divide its current and potential consumers (doctors) into segments, and ultimately identify those 20% that contribute to the most profit. To perform the decision trees analysis, the pharmaceutical company needs to collect the following data: the demographics of most profitable customers (doctors), the geographic location of the hospitals that those customers work for, the listing of insurance companies that the doctors closely work with, the number of patients the doctors serve per day/week/month, and finally the number of prescriptions that doctors write per day/week/month.
#1
A decision tree could be used in a business setting when deciding to outsource a function such as production. You could find data on costs, cycle time, laws and regulations in outsourced country. You could also decide to what extent you would outsource; outsourcing certain components, outsourcing an entire project, or not outsourcing at all by looking at costs/time/security/risks.
1) One business question could be answered is when a company is trying to decide whether they should build a new facility to produce more products. The building will cost x amount of dollars and in the worst can scenario they will lose y amount of money and in the best case they will gain y amount of dollars. With certain projections and probabilities would it it beneficial for the company to build the facility or should they just continue to use the old facility?
Name and describe a business question that you could answer using a decision tree. What data would you collect to perform the analysis?
You could use a decision tree to help select a new market/country you want to introduce a product in. Depending on the type of product, you can look at competitors success in such markets and weigh the costs and benefits of introducing your product in that market. You can look at different countries, the degree of penetration of the market, and the profits of the competitors already in the market. If profits were over a certain dollar amount you can opt to enter the market or if it was lower you can decide to go with another market/country.
1. You could figure out how how likely a buyer will qualify for a new house. . You could collect income, job stability, Savings, down payment, whether they already own a house or not, etc. This data could be analyzed to help figure out whether or not a person would be able to buy the house or not. It would essentially determine if they would be able to realistically afford the new home.
#1 a business question that you could answer using a decision tree.
For example, car retailers could use a decision tree when they want to know whether customesr would like to buy a certain type of SUV. They could collect the following data: Are the customers single or married? How many family members do they have? How many is the income, etc.
1. A business question you could answer by using a decision tree could be how likely someone is to get into a certain college. This information would be based on high school GPA, extracurriculars, recommendations from teachers, etc. People often predict whether or not they will get into a college solely based on these specific areas.
A business question that you could answer using a decision tree is whether or not to open up a new office. If a company wants to expand into a new location they need to collect the data representing the costs, benefits, and opportunities within that location that could come about from opening another office there.
Question #1
One business question that one could answer using a decision tree involves departmental analysis. As employees within a company are hired and/or promoted, a decision tree may be used to determine which jobs/overall departments would be the best fit for these individuals. Data that could be collected to build this tree might include relevant experience in the field, longevity, teamwork experience, leadership experiece, technical expertise, compatability, and so on. While this has the potential to greatly assist companies allocate their resources effectively through analysis, this type of business tree may only work in large businesses/corporations. This is because, the more data that can be collected, the more accurate the results of the tree.
One business question that could be applicable to a decision tree could be pertaining to risk exposure for a car insurance company. When analyzing potential applicants to cover, not cover, or cover with certain restrictions they can asses type of gender, age, car type, mileage, location, number of previous accidents/severity of them, and credit score to gather enough information to correlate these variables and understand who is a good candidate to represent.
#1
Big hotel and casino chains could use decision trees when offering yearly comps to valued customers. Guests could be grouped on how often they stay, how much they spend on casinos, and what type of amentities they opt for. You could group guest into categories based on spending brackets which could in turn decide the comps offered to each type of guest.
#1
A business question that could be answered using a decision tree is what hospital someone should go to have a dire procedure performed. Some information that should be considered in the process are:
-How much the procedure cost at the hospital
-What is their performance ranking
-How many fatalities and success they have
-How far the hospital is
-How soon the procedure needs to be performed
-How is their patient doctor/nurse relationships
All of these could be decisions someone would need to consider in what hospital they should get their procedure done and feel the safest and most comfortable at.
A decision tree analysis can be very useful for a mortgage broker when trying to assist a buyer or seller. In reference to a buyer, he/she can get a better insight of loan options that are more popular than others, what are the commonly known interest rates that are out there for new home owners (of course it is also determined by your credit score) but it gives the buyer an idea of what banks are currently offering. Also, can see a increase or decrease fluctuation based on debt ratio of loans and interest rates being distributed. What areas in which is nearby or where the buyer plans on living. Determine demographics based on gender, race, class of people that may reside within those areas.
A business question that could be answered using a decision tree is that “Is better to import a component from supplier or to product it by self?” we can create a decision tree include the estimate total cost, time and revenue of each project. Then compare them, looking for which one is advantage and disadvantage. I think this is extremely helpful to decision-makers.
1.
One question a decision tree could answer is for airline companies. They could find out
how many people use frequent flyers miles compared to flyers who don’t. Then they can determine how many miles they should give out with each flight and what type of rewards should be given.
1) A business question that could use a decision tree could be looking at purchasing/trading players in baseball. Teams will want to consider a variety of variables when considering the players and certain variables are more important than others. Choosing the best players for the team and for the money that each player may cost could be answered using a decision tree. The data collected to perform the analysis may include stolen bases, runs batted in, batting average, on-base percentage, and slugging percentage.
1. You can look at the likelihood of whether a student will pass a class or not. Some factors that can be consider can include attendance, participation, quiz grades, test grades, assignment grades, and frequency of office hour visits. Another factor that could be used is their semester credit load.
#2
I would advise to pick predictor variables that actually apply to the decision (should be slightly obvious) and are unique. The variables should begin with values that are more likely to be applicable to many situations and they should sequentially become more detailed and more likely to have more definitive outcomes (e.g. 0,1).
#2 The advice I would give someone regarding how to select the right predictors for a decision tree would be to choose them based on the given problem and choose predictors that have some type of correlation to each other.