Reflection

The first model, we developed, was using a logistic regression and, in my case, the wine data set. I felt that this was a good starting point to dive back into machine learning. However, one thing that irks me about this model is that there isn't really a metric on its performance.

The second model, around linear regression, felt like a logical progression from the logistic regression. However, the model still doesn't include a metric, and thus I'm not actually sure how well it performs.

In third model we had a look at k-means and we also learned how to recognise how many clusters a model has via the elbow method. I think my submitted model performed decently well. However, in the admittedly arbitrary category of how exciting the model felt, its behind the later models.

In the next assignment we started to delve into neural networks and how they can recognise images. This was the first topic that I really hadn't heard of before. This is not to say, that all the topics before were boring or that I haven't learned anything. On the scale of excitement, this definitely is more on the upper end. Sadly, the model I submitted didn't perform to well.

The next submitted model, was about labelling different clothing items and, with the newly learned convolutional neural network, my model hit an accuracy of 91.4%, which is the highest of the neural networks that I made myself. Because of this, this is the model that I've decided to submit. It's the best performing model and simultaneously also the most interesting model, in my opinion.