Compute training error vs compute test error

Course Q&A Machine Learning Specialization Advanced Learning Algorithms
1

In the week3 Evaluating a model

Andrew introduced two ways to test your model’s performance without plotting.

We first split datasets into training and test sets.

Then we can use the test set data to check how’s our model’s performance.

image
image1105×597 119 KB

The second way is to reuse the training set to compute the model’s performance, however I am confused here, if we reuse the training set, isn’t it will not make any difference? How can we get info of how is our model’s performance from it since we used those data to train our model first? Isn’t the cost will be the “same”(not the same but very close to the cost since we didn’t use the regularization) What is the math behind it?

1 Like
2

Hello @Bio_J,

I think you have pointed out the problem of measuring model performance with the training set! However, that is not the intention. Andrew is just showing us how to carry out the train/test procedure, and he will talk about how we use these two values, for example, when he discusses diagnosing Bias and Variance in one of the next videos.

So, for now, please don’t consider that we are free to choose either of the two cost values as model performance measurement.

Cheers,
Raymond

1 Like