https://github.com/mazensayed91/binary_classification

https://github.com/mazensayed91/binary_classification

Last synced: 11 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/mazensayed91/binary_classification
• Owner: Mazensayed91
• Created: 2020-08-08T22:04:59.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2020-11-16T07:29:26.000Z (over 5 years ago)
• Last Synced: 2025-04-04T08:47:31.985Z (over 1 year ago)
• Language: Jupyter Notebook
• Size: 289 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Binary_Classification

Cleaning and Dealing with im-balanced dataset using smote and achieving good results on both train-validation datasets

# Thoughts and observations:

all my thougths and observations are discussed through the process in the notebook

# Side-notes:



  
Variables 17,14 are the same variable but scaled

  
Variables 4,5 are the same variable but shifted

  
Variables 19 are the same as the label but they are not the same in the validation set

   
After PCA, we can drop more than 80% of the features and still have >90% of the variance of the data!

  
The data is imbalanced so the accuarcy metric is not the best metric , we should care for F-beta score



# To-Do



  
Adding microservice for prediction task

  
Getting more insights from the data



# Final Results-Using rfc with smote:



  
Training set has : F0.5 = 0.9965  Recall = 1  Precision = 1  Accuarcy = 1

  
Validation set has : F0.5 = 0.9745  Recall = 0.98  Precision = 0.97  Accuarcy = 0.97