Decision Tree Analysis
Western Governors University
C207: Data-Driven Decision Making
Student Name:
Date: May 20, 2023
Decision Tree Analysis
A: Describe a business question that could be answered by applying decision tree analysis and is derived from the scenario in the attached “Decision Tree Analysis Resources.”
To maximize revenues, should the MPC create a new medicine, make likely adjustments to the present drug, or do nothing at all?
- Identify the relevant data values required for your decision tree analysis.
ALTERNATIVES | Probability of success | Profit Per Unit | Demand (Units) | Payoff | |||
Unfavorable market | Favorable market | Low (Unfavorable market) | High (Favorable market) | Unfavorable Market | Favorable Market | ||
New Drug | 29% | 71% | 0.63 | 1205 | 4341 | $759.15 | $2734.83 |
Changes to Existing Drug | 37% | 63% | 0.48 | 1807 | 5475 | $867.36 | $2628.00 |
Current/No change | 19% | 81% | 0.84 | 241 | 730 | $202.44 | $613.20 |
C1. Report how you analyzed the data using decision tree analysis
Find the attached decision tree analysis file for my analysis of the provided data and how I determined the best course of action.
C2. Justify why decision tree analysis is the appropriate analysis technique, including relevant details from the scenario to support your justification.
In order to make the best feasible choice, the decision tree analysis is used to examine many possibilities and different decisions. In this scenario, MPC has a choice between three distinct paths forward; it should choose the one that presents the least risk and high profitability. The first alternative is whether to develop a brand-new medicine, tweak an existing one, or do nothing at all; the second is the current market climate, including both positive and negative scenarios, which may be used to determine which course of action would provide the best results. Using this analytical tool, the aforementioned choices and their potential repercussions may be compared and contrasted, and the optimal course of action for the company’s productivity and profitability can be determined.
- Summarize the implications of your decision tree analysis by doing the following: D1:
Explain the role of probabilities and the role of demand for each branch.
The concept of probability is used to provide information on the likeliness of an occurrence. Probability has been utilized to illustrate the likelihood of each result and to compute expected value for each scenario. The first fork in the process of developing a new medicine has a 71% probability of success in a favorable market and a 29% chance in an unfavorable market. The payment for each option is based on the monthly demand for the medicine, which is 4341 units in a good market and 1205 units in an unfavorable market. In the second scenario, when the current medicine is modified, the option has a 63 percent probability of success in favorable markets and a 37 percent chance of success in unfavorable markets. Both markets have expected monthly demand of 5475 units and 1807, respectively. The numbers may be used in anticipated value calculations, which can aid in making choices.
If no adjustments are made, the option has an 81% chance of success in a favorable market and a 19% chance in an unfavorable market. Monthly forecasts for both favorable and unfavorable markets are 730 and 241 units, respectively. The predicted value is then calculated using the values, which aids decision making.
D2: Explain how the expected value of each node is determined based on payoffs.
Expected values are determined for each note’s state and nature to determine the possible predicted profit to help in decision-making. This makes it easy for the company executives to make the right choice for the company that will maximize their profits and minimizes risks of loss.
For the first node of making the new drug, the state of nature note was calculated as follows:
Expected Value = ((success change for favorable market %) * (payoff for favorable market)) + ((success change for unfavorable market %) * (payoff for unfavorable market)) (GENERAL FORMULA)
EV = (0.71*2734.83) + (0.29*759.15) =$ 2161.88
The same formula is used to calculate the expected value for the other nodes. The node in the second branch of developing the existing drug is given by:
EV = (0.63*2628.00) + (0.37*867.36) = $1976.56
For the last node of making no change, the expected value is given by: EV = (0.81*613.2) + (0.19*202.44) = $535.16
D3: Limitations
Data elements: One limitation is the predicted demand for each alternative option and the market possibilities. Since these values are based on market research, they are subject to change in the case of a real-world scenario. Hence, they cannot be relied on as effective predictors. In some cases, they can mislead an organization into making the wrong decision based on faulty information.
Decision tree analysis: The technique is unstable in nature. Minor changes in data can lead to a significant change in the output (Lee et al., 2022). It can be from inaccurate market research or input error during the analysis, and if not tackled with high accuracy can result in a misleading result; hence the wrong decision is made.
Also read: Buy Nursing Care Plan Online
- Recommend a course of action that addresses the business question from part A and is based on the results of your decision tree analysis.
The advice for this organization is to press ahead with developing a new drug line. This conclusion is based on the results of the analysis that show this alternative has the highest expected value of $2161.88.
Changing an existing drug line does indicate a higher demand in the unfavorable market in addition to its expected value of $1976.56, hence making this alternative the best back up action for the company. The option of doing nothing displays an expected value of just $535.16, making it less profitable. Therefore, if the company wants to maximize profits and minimize risk, developing a new drug is the company’s best bet.
References
Lee, C. S., Cheang, P. Y. S., & Moslehpour, M. (2022). Predictive analytics in business analytics: decision tree. Advances in Decision Sciences, 26(1), 1-29.