Preparing the y Vectors Using One-Hot Encoding (if Necessary) As explained in lectures, if the number of dasses K = 2, t

[answerhappygod]

Preparing The Y Vectors Using One Hot Encoding If Necessary As Explained In Lectures If The Number Of Dasses K 2 T 1 (26.59 KiB) Viewed 56 times

Preparing The Y Vectors Using One Hot Encoding If Necessary As Explained In Lectures If The Number Of Dasses K 2 T 2 (23.12 KiB) Viewed 56 times

Preparing the y Vectors Using One-Hot Encoding (if Necessary) As explained in lectures, if the number of dasses K = 2, then we will only use a single output node and the y vector will remain as it is. However, it 3. then we will have K output nodes and we will have to reformat our y vector(s) using one-hot-encoding Given a column vector of labels y with ascending labels starting from 0, the one-hot-encoded vector ŷ corresponding to y is a vector with the same number of rows as y and K columns in which the value of each row in y specifies the index of the column in 9 of the same row that should be set to 1 All other values are zeros. This can either be done in vectorized format with numpy Indexing magic, or a loop If you want to Up to you. Since we'l only ever do this operation a very small number of times, it isn't critical that it be vectorized Go ahead and complete the function below to produce the matrix yhat corresponding to a given y as per the spec desaibed [2 marks]

In [15]: def onehotenc(y): #y - a column vector with one Label per row, assumed to contain ascending Labels from 0 up to X. K - len(np.unique(y)) #You will need this. Don't change. Number of classes K. yhat - np.array([e]) #FILL I CODE BELOW: IM You need to set the variable yhat correctly yhat - np.zeros((y size, k)) Code to set the variable yhat correctly. I would supgest first setting the dimensions of what correctly and the proceeding yhat[np.arange(y size), y] - 1 STOP FILLING IN HERE return yhat.astype("int")