Skip to content

amajji/Multi-class-classification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Multiclass classification.

Data scientist | Anass MAJJI


🧐 Description

  • This project aims to implement a multi-class classification model. we have four classes with a minority class (less than 1%), so we are in the case of unbalanced classes. We used regularization methods in order to penalize the errors made on this minority class. The model is trained to predict around 4 different class.

🚀 Repository Structure

The repository contains the following files & directories:

  • Dataset directory: It contains a data pre-processing notebook where the train.csv file is used for training the model. Il contains also the predictions of test.csv dataframe.

  • model_weights: It contains all the weights of the models : one-hot-encoder, target encoder, random forest model.

  • App directory: Code for the web application that was developed for the model deployment. It contains Flask API code for the Back-End, and HTML/CSS/Javascript code for the Front-End.

📈 Performance & results

  • The test dataset contains 25 000 samples. Each sample contains many features, and its corresponding label.

  • The model used for this multi-class classification task is a Random Forest model.

  • The metric used to measure the model's performance is F1-score. After testing the model, I obtained a test F1-score of 72 %


📪 Contact

For any information, feedback or questions, please contact me