Full course description
Course Overview
This course explores how to evaluate the performance of classification models and determine whether their predictions are reliable enough to support real-world decisions. Building on your understanding of logistic regression, you will learn how to move beyond simply building a classification model to assessing how effectively it distinguishes between different classes.
You will examine widely used performance measures, including the confusion matrix, sensitivity and specificity, precision and recall, the F1 score, and ROC curves. The course also explores the business consequences of classification errors, helping you understand how false positives and false negatives can have different costs depending on the situation. Through hands-on work in Python, you will apply these concepts to evaluate a logistic regression model and interpret its performance.
Why enroll in this course?
Building a classification model is only useful if you can determine how well it works. This course will give you the practical tools to evaluate classification models, compare different measures of performance, and understand the tradeoffs involved in deciding what constitutes a successful prediction.
Whether you are a business professional, analyst, or aspiring data scientist, this course will help you connect technical model performance with business impact. You will develop the skills to interpret classification results and make more informed decisions about how predictive models should be used in real-world settings.
What will you learn?
By the end of this course, you will be able to:
- Calculate and interpret key classification performance measures, including sensitivity and specificity
- Use confusion matrices and ROC curves to evaluate classification models
- Calculate and compare precision, recall, and the F1 score
- Evaluate how classification thresholds affect model performance
- Analyze the costs associated with false positives and false negatives
- Assess the business impact of classification errors using cost matrices
- Use Python to evaluate a logistic regression model and interpret its performance
Structure
The course is completely self paced. It will take you approximately 20 hours to complete all four modules. Activities include, video lessons, readings, and self reflection activities. Upon successful completion of this course, you will receive a certificate of completion.
Smith Executive Education Homepage | Contact us: rhsmith-execed@umd.edu