Answer Happy

Question 1:
Total Mean Squared Error is:
Select one:
a. Summation of : instant error at each output neuron for each
training input
b. Summation of : half mean squared of (instant error at each
output neuron)
c. Summation of : half mean squared of (instant error at each
output neuron for each training input)
d. none of the options
Clear my choice
question 2
In Back-Propagation algorithm, the weights are updated in the
backward propagation path.
Select one:
True
False
Question 3
In Back-Propagation algorithm, only weights between output
neurons and hidden neurons are updated
Select one:
True
False
Question 5
At each hidden layer in the feed-forward phase, the following is
calculated:
Select one:
a. V1[k,l,j+1]=V1[k,l,j+1] +w1[kk,j,i]*inputB[l,i]
b. Y1[k,l,j+1]=a*np.tanh(b*V1[k,l,j+1]
c. Both listed options
d. none of the listed options
question 6
In BP algorithm, the weights are initialized as:
Select one:
a. zeros
b. random values from the data set
c. uniformly distributed random values
d. none of the listed options
question 7:
For diferent initializations, the K-means algorithm will
denitely give the same clustering results.
Select one:
True
False
Question 9
The centroid of each cluster is calculated by the mean of points
belonging to this cluster.
Select one:
True
False
question 9:
Choose the correct statement(s).
Select one or more:
a. In K-means clustering, data points are assigned to the nearest
cluster.
b. In K-means clustering, data points are assigned to the farthest
cluster.
c. In K-means clustering, the algorithm must be repeated until no
further change in the cluster assignment
d. Depending on the initial choice of the centroids, the result may
converge to a local minimum resulting in a suboptimal solution.
question 10:
Too small learning rate results in very slow training
process.
Select one:
True
False
question 12:
For certain learning rates, LMS can give unstable results
Select one:
True
False
question 13:
Too high learning rate results in very slow training
process.
Select one:
True
False