Kaagel-titanic-problem

This repo is a try to solvz the famous titanic competition on kaggel, We try seviral ML models and Neural Network to find best F1-score possible .

Data Processing :

we started by EDA to check the data if it satisfies the challenge requirments, we made Droped some unmeaningfull data (like Names Ticket Number ..) then we filled null values with 0s or means according to each column, and finnaly we scalled the data.

Models Used :

K-NN :

1st model, choosed it to see where our accuracy ranges, 1st random try gave us 0.78 f1-score. after K-kroos validation we found out best k for the model = 6 .

Decision Tree :

Results :