This application allows users to visualise the Perceptron Algorithm of a two-class classification problem. The linear boundary shifts and adjusts until either the number of iterations has reached a maximum, or the two classes have been classified correctly by the boundary.
This section displays the points, where blue dots represent class 1 and red crosses represent class 2.
Allows the user to set a lower-bound and upper bound on the x-axis and y-axis. Pressing 'Update Axes' updates the graph axes, and if points are already plotted, they will scale accordingly.
The x and y coordinate of each point can be specified manually and a class for that point can be chosen before adding it to the graph. The user can add as many points as they wish.
This feature allows the user to generate random points which is linearly seperable, therefore it guarantees the Perceptron to divide the two classes.
The initial weights must be specified, w1, w2, and w0 (bias). This is the initial boundary/line through the dataset. An appropriate learning rate must be set and an iteration limit. The iteration limit becomes important for classifications which are not linearly seperable as it will cause the algorithm to run infinitely. The user can choose the speed at which the linear boundary moves, using a slider, before running the algorithm. Click 'Run Perceptron' to watch the linear boundary move to seperate the two class classification.