In this repository you will find the code that I used to create machine learning models for the Titanic problem. This project had two main purposes:
- To get more familiar with machine learning and the usage of the available algorithms
- To analyze the accuracy and speed of the different algorithms
Note that the results while not bad, could be better because I purposely did only limited pre-processing on the data-set. Together with the problem not being complex enough this is most likely also the reason that when optimized, most models have around the same accuracy.
The Titanic dataset posed a nice resource for my goals as it is not very complex and has been done a lot of times, thus there were a lot of resources to guide me in my journey.
While this was beneficial as my first project, it is a bit unfortunate that we were not able to see a lot of difference in the algorithms though.
Therefore my next project will be on a more complex problem.
I also learned that it likely returns better results to optimize the dataset rather than the model, although both would be best ofcourse.