From the course: Building Recommender Systems with Machine Learning and AI

Unlock this course with a free trial

Join today to access over 22,600 courses taught by industry experts.

Exercise results: Tuning a RBM recommender

Exercise results: Tuning a RBM recommender - Python Tutorial

From the course: Building Recommender Systems with Machine Learning and AI

Start my 1-month free trial Buy for my team

Exercise results: Tuning a RBM recommender

“

- [Narrator] As you probably learned the hard way, it takes a really long time to tune this. What I learned from the process however, was a couple of things. First of all, it seems that as far as accuracy is concerned, reducing the number of hidden units, and increasing the learning rate helped, but not by much. Even after days of fiddling with it, I couldn't get the RMSE below 1.18 or so. And that's not a significant improvement over the 1.19 we started with. And these supposedly better recommendations don't look subjectively better if you scroll down and look at the results. So while hyperparameter tuning might squeeze out some gains in our RBM, it seems like our problems are deeper than the parameters. I suspect it's really a problem of not having enough data to properly train it. Later in this course we'll scale a base sparse neural networks similar to an RBM and run it in the cloud, which will allow us to experiment with a much larger data set to see if we get better results that…

Contents