Letter-Recognition

Training different classification models in order to classify different letters.

Requirements

• Python 3.10+
• Python Libraries
  • pandas
  • numpy
  • matplotlib
  • sklearn

Use the following command to automatically install all dependencies:

pip install requirements.txt

How to use

Simply clone this repository and run the following commands:

python main.py

Once models are trained, navigate to the "/data/" directory and run the following command in order to generate performance charts:

python process.py

Expected Input

Modify the variable PAIR inside main.py.

• PAIR = 1 train and test models with the letters H & K
• PAIR = 2 train and test models with the letters M & Y
• PAIR = 3 train and test models with the letters A & B
• PAIR = 4 train and test the multi-class model with the letters H, K, M, Y, A & B

Expected Output

1. Various CSV files, inside the /data/ and /test/ directory, with statistics about each model
2. Various plots--after running /data/process.py
   • Example #1:
     
   • Example #2:
     

Folder structure

/data/

This folder will contain the raw output of the models: accuracy, training, and testing times, etc.

/test/

This folder will contain results after models are tested using the final validation set (10% of the input data set).

Libraries/Classes

/libs/const.py

This file is used for pre-processing the input dataset file (letters-recognition.data)

/libs/models.py

This file include all the implementations of the models used. The Model class include a function for each classifier.

/data/process.py

This file is used for generating plots

main.py

This file is the main driver