Practicing The Perceptron: Introducing Data Mining To The CS2 Classroom

摘要

We describe a 2-week team project designed for a typical CS2 class, giving students experiences crafting a complex object-class called a Perceptron. The basic perceptron is a statistical machine learning model resulting from a training algorithm capable of distinguishing between two or more categories of input vectors. In the process, students are introduced to basic concepts of data mining and data visualization, and use the model they build to explore datasets of interest to them. Several semesters of assigning variations of this project to students of varying levels in our Computer Science program refined the process for introducing this topic to students within our CS2 curriculum so it appears both relevant and meaningful. Initial observations indicated that having students work in small teams enhanced their individual understanding, and thus the ultimate success of their project. Reports from past students indicated having a choice in final datasets to use, as well as the teamwork aspects made this project particularly compelling to computer science majors in our program, a potential positive indicator for our departmental retention goals.