Training different classification models in order to classify different letters.
- Python 3.10+
- Python Libraries
- pandas
- numpy
- matplotlib
- sklearn
Use the following command to automatically install all dependencies:
pip install requirements.txt
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
Modify the variable PAIR
inside main.py
.
PAIR = 1
train and test models with the lettersH
&K
PAIR = 2
train and test models with the lettersM
&Y
PAIR = 3
train and test models with the lettersA
&B
PAIR = 4
train and test the multi-class model with the lettersH
,K
,M
,Y
,A
&B
- Various CSV files, inside the
/data/
and/test/
directory, with statistics about each model - Various plots--after running
/data/process.py
This folder will contain the raw output of the models: accuracy, training, and testing times, etc.
This folder will contain results after models are tested using the final validation set (10% of the input data set).
This file is used for pre-processing the input dataset file (letters-recognition.data
)
This file include all the implementations of the models used. The Model
class include a function for each classifier.
This file is used for generating plots
This file is the main driver