Discussion 1: Tyler Klingler
Explained on a very simple level, logistic regression is a process that “uses math to evaluate the chances of something happening or not happening” (Bell, 2019). Given the right conditions within the data are me for logistic regression to be feasible, we can answer yes or no questions such as if someone will default on their home loans. With this foundational knowledge on what a logistic regression is and what its output will look like, we can provide a more clear result to our non-technical audiences. We’ll want to avoid giving too much how and now enough why, providing too much detail, difficult-to-understand terms and acronyms, and we want to avoid sounding big and clever (York, 2019).
There are many great ways in which the results of logistic regression can be explained to non-technical audience of which I’ll name three. One approach that one can take to explain these results to a non-technical audience includes cutting all but the most important information about your process and your conclusion. Another approach would be using graphical tools to paint the picture of your results. Another great approach is to provide more details on the process, important notes, and results to your audience but this information can be given through easy-to-understand metaphors.
There are many great benefits of interpreting analysis results simply for business leaders to understand. Perhaps the first and greatest benefit is that you can more easily display the value that you are providing to your company. This can provide you with job security, additional funding, and build the confidence of your work among the business leaders. Another benefit is that business leaders can more easily recognize where the company is succeeding and where changes need to be made. Your ultimate goal is to have your work steer the company in the right direction and reduce drag. If you can easily explain what is slowing the company down and what may be pulling it in the wrong direction, business leaders are more likely to take this advice and this can lead a company to thrive even more.
References
Bell, S. (2019, October 18). Explaining to a 5-year old: Logistic regression. Medium. Retrieved October 15, 2021, from https://towardsdatascience.com/explaining-to-a-5-year-old-logistic-regression-a6085d8e23a1.
York, K. (2019, December 8). Geeks: How to write for a non-technical audience. GeekBoss. Retrieved October 15, 2021, from https://geekboss.com/write-for-a-non-technical-audience/.
Discussion 2: Arcelia Rael
The results of logistic regression can be difficult for non-technical users to understand because of their technicality and complexity. The interpretability and explainability of these models are important because of how a model may come to drive business decisions, and how it can impact (for better or worse) consumers. Additionally, explainability may be mandated by legislation, like the Data Protection Regulation (GDPR), which includes a clause relating to explanation, where any models that may impact an individual must be readily explainable to the user (Sivek, 2020). Sivek (2020) notes that some models are more difficult to explain and interpret than others (logistic regression vs. Neural networks), but there are several steps that analysts can take to make them easier to digest.
One step that can be taken to make models easier to understand is to use visual tools that allow stakeholders to view the results of the model instead of having to view the raw output. These tools include graphs, plots, or even websites that walk the user through the model’s output. Sivek (2020) writes that quantitative measures can also be used but can be less impactful on users who are not familiar with the problem. In this case, it’s important to know your audience, as presenting these measures may detract from the explanation. Lastly, analysts should employ software that allows them to understand the inner workings of more complicated models (sometimes referred to as black boxes) so that they can ascertain the accuracy and be able to explain the model to stakeholders (Sivek, 2020).
References
Sivek, S. (2020). Interpretability, explainability, and machine learning – what data scientists need to know. KDnuggets. com/2020/11/interpretability-explainability-machine-learning.html” rel=”noopener noreferrer” target=”_blank”>https://www.kdnuggets.com/2020/11/interpretability-explainability-machine-learning.html
Discussion 3: Bria Jackson
“The Logistic Regression is a regression model in which the response variable (dependent variable) has categorical values such as True/False or 0/1. It actually measures the probability of a binary response as the value of response variable based on the mathematical equation relating it with the predictor variables (R – Logistic Regression, 2021).”
When it comes to a nontechnical audience, they truly depend on analysts to interpret logistic regression in a way they can understand so the appropriate decision can be made. Analysts can make this happen by plotting the data into graphs, chart and or tables. When using the three ways listed for analysts to make logistic regression results interpretable and meaningful for nontechnical audiences it is helpful to use symbols, colors, and legends with the visualizations.
It is extremely important when you are an analyst to know how to communicate with leaders and employees who are nontechnical. It’s important for an analyst to know how to communicate and break down the data that needs to be conveyed to the company, what good is it to know how to analyze but cannot communicate effectively. Typically, when analyzing the data, it is needed for important business decisions. Communication is KEY!
References:
R – Logistic Regression. (2021). Tutorialspoint. Retrieved October 14, 2021, from https://www.tutorialspoint.com/r/r_logistic_regres…
Le, J. (2018, April 10). Logistic Regression in R. DataCamp Community. Retrieved October 14, 2021, from https://www.datacamp.com/community/tutorials/logis…
0 comments