This is a simple implementation of the Perceptron algorithm, a supervised learning algorithm used for binary classification tasks. The project was created as part of the Artificial Intelligence Tools course at my university. It allows you to train a Perceptron on labeled data and test its performance on both training and test datasets. Additionally, you can classify a vector of values manually after the training phase.
-
Training: Trains the Perceptron using a learning rate and threshold.
-
Testing: Evaluates the accuracy of the model on the test dataset.
-
Manual classification: Allows the user to input new data for classification based on the trained model.
After starting the program, select a file containing labeled data with values separated by commas (a sample file called "iris_train.txt" is provided as an example). Then, enter the appropriate parameters for your classification task. The program will automatically shuffle and split the data based on the "testSplitRatio" provided by the user.
Once the model is trained, the program will display the accuracy for both the training and testing phases. Additionally, it will allow the user to manually enter new vectors for classification.