C1W2 - How X matrix represents m training examples

[humanbot]

Hi Everyone.

I am new to the course and need some help. I am confused about the notations mentioned in C1W2 on the slide:

What confuses me is that in X, each training example x1, x2 and x3 is stacked as column and X will have m columns.

My understanding is different. Let’s take a housing example with 3 input features (No. of Bedrooms, Size, Location) and 1 output feature (Price). Let’s also assume we have 10 housing examples. Now if we stack up the 10 rows in X, we will have 10 rows x 3 features. However, Dr. Andrew mentions that X will have m columns (10 in this case) and nx rows (3 in this case). If someone can give an example by drawing, that will really help.

[TMosh]

There is no standard on the orientation of the X matrix. For some datasets, the examples are in the rows. For others, they may be in the columns.

[humanbot]

Yes I understand that there is no standard but I am trying to understand Dr. Andrew’s explanation. To simplify it further, as per Dr. Andrew, which of the following will apply for both X and Y:

[rmwkwok]

and ,

in accordance with DLS’s convention, which is