Comparing Performance for different Machine Learning Models [closed]

I am reviewing Machine Learning Algorithms. When I tried to compare the results of two different algorithms, say Decision Trees & Random Forests. I noticed that even with the same dataset, if I do not specify the random_state for the model, the modelling results(confusion matrix, precision, accuracy, etc.) would be different each time. And sometimes the accuracy for DT would be higher than RF. So I set the random_statea for both models.

Here is my question. If I set the random_states to the same number for different models, can I safely say that the relationship between the results generated would be correlated so that I could compare their performance?

Here is the code I tried: [enter image description here](https://i.stack.imgur.com/I30DP.png)enter image description here

Many thanks in advance!