This repository contains a Python code that uses the MNIST dataset to train four different classification models to recognize handwritten digits.
Clone the repository and install the required packages via: pip install -r requirements.txt
To train the model of your choice run the following command: python KH-MNIST.py
- The RFC had the highest accuracy score and the shortest training time.
- The SVC had the second highest accuracy score and the longest training time.
- The KNC had the third highest accuracy score and the second shortest training time.
- The LR had the lowest accuracy score and the shortest latency.
Accuracy Score: measures how well a model predicts the correct output. Here it refers to the percentage of times that the model correctly predicted the label of each input.
Training Time: amount of time it takes for a model to process or "learn" a dataset.
Latency: amount of time it takes for a model to make a prediction on a new input or the time spent for predicting a single digit.