data engineering for algorithms and giving report

I have five models with python for ddos detection, all of them work the same for the given data set and all values are 1. I wanted you to do something with the dataset, i.e. feature engineering or feature importance, so that all five of these algorithms do not give the same answer. After this, compare all three models in a report in English, evaluate and explain each one What reasons were better or worse? The performance report and evaluation of these three models is very important to me. I only have a day for this and I can't pay more than 15 dollars.