From the course: Artificial Intelligence Foundations: Neural Networks

Unlock the full course today

Join today to access over 24,900 courses taught by industry experts.

How do you improve model performance?

How do you improve model performance?

From the course: Artificial Intelligence Foundations: Neural Networks

Start my 1-month free trial Buy for my team

How do you improve model performance?

“

- [Instructor] As you may recall, the image on the left is of an underfitting model. In chapter four, you had a hands-on lab, where you built a simple Keras neural network. The image on the right shows the model's loss curves for the validation and training data for that model. The model loss curve appears to be very close to underfitting. You could tune the hyperparameters to improve your model's performance and in this example, hyperparameters are knobs that you can tune or change to see if the model improves. You could change the model's architecture, the optimizer, the loss function, and the number of epochs. Shown here is the code you use to build your simple neural network in the hands-on lab. The code shows a three-layer neural network with four input units in the first layer to receive the four features. A second hidden layer with three nodes, or units, and one output layer. Note the optimizer as "adam"…

Contents