ahayek84 / msr_alpha_2022 Goto Github PK

The Steps :

1. pip freeze > requirements.txt

possible problem :

• Merge conflict !!

Reuired Packages :

• SMOTE_VARIENT : pip install smote-variants

Preprocessing steps :

• Read the data from data folder ( Mongodb, react, socketio)
• Combine all in one data frame
• fill the null values with 0
• Use MultiColumnLabelEncoder to Label none numerical data ["login","name","email","source"]
• Source colmun is for the data source ( Mongodb, react, socketio)
• Use Smote_varient library to oversample minority classes

Classes

• preprocessors.py :: be called by create_dataset() , it will return a prepared dataset with classes balanced
• Link: https://github.com/ahayek84/msr_alpha_2022/blob/main/process/preprocessors.py
• Parameters : source = ["mongodb","react","socketio"] or defualt = "all" , use_smote=True