MCQ's on Linear Regession, Logistic Regression,Neural Network,Deep Neural Network and Convolution Neural Network

 MCQ's  on Linear Regession, Logistic Regression,Neural Network,Deep Neural Network,Convolution Neural Network

Q.1) Linear  Regression is a type of ---------

a)Supervised learning

b)Unsupervised learning

c) Reinforcement learning

Correct Answer a)Supervised learning

Q.2) ------------------variable is  Predictive variable

a)Dependent 

b)Independent

Correct Answer a)Dependent

Q.3) The name multinomial logistic regression is usually reserved for the case when the ------variable has three or more unique values, such as cloudy, Rainy, hotor cold.

a.)Dependent

b.)Independent

Correct Answer – a) Dependent

Q.4.Logistic regression does not assume a linear relationship between the dependents and the independents.

  

a)False

b)True      

Correct Answer – b)True       

Q.5.Dependent  variable is categorical:

a)Y Ԑ(0,1)

b) Y Ԑ(-1,1)

 Correct Answer – a) )Y Ԑ(0,1)

Q.6. To  Predict whether mail on mail box is spam or not spam------algorithm is suitable

a)Linear

b)Logistic

c) Reinforcement

d)Deep

Correct Answer – b)Logistic

Q.7.Most widely used transfer function in  deep learning is

a)Leacky  Relu

 b)Relu

c)Sigmoid         

d) linear

Correct Answer – b)Relu

Q.8.Guy went fishing five times in a week.He caught a fish two times and failed to catch 3 times. What is the odds values

a)2/3

b)3/2

c)2/5

d)3/5

Correct Answer – a)2/3

Q.9. Algorithm suitable to  predict categirically dependent variable based on independent variable which may be continuous orCategorical

a)Linear

b)Logistic

c) Reinforcement

d)Deep

Correct Answer – b)Logistic

Q.10.In case of logistic regression ----- function  is used

a)Step function

b)Ramp function

c)Sigmoid function

d)Gaussion

Correct Answer – c)Sigmoid function

Q.11.Suitable algorithm to predict binary values  is----

a)Linear

b)Logistic

 c) Reinforcement

Correct Answer – b)Logistic

Q.12.Suitable algorithm for image, end to end  prediction is-----

a)Linear

b)Logistic

c) Deep learning

Correct Answer – c)Deep learning

Q.13 Y=bo+ b1*x+e is the equation of ----- regression

a)Linear

b)Logistic 

c) Reinforcement

d)Deep

Correct Answer – b)Linear

Q.14.In back propagation ---learning rule is used

a)Delta

b)Winner take all

c)Peceptron

Correct Answer – a)Delta

Q.15 SOM algorithm is type of --------

a)Supervised learning 

b)Unsupervised learning         

c) Reinforcement learning

Correct Answer b)Unsupervised learning 

Q.16.SOM Architecture is -----layer network

a)One layer

b)Two layer

c)Multilayer

Correct Answer b)Two Layer

Q.17. Winner in SOM algorithm is determined by equation ------where x with weights wj for each neuron j to determine winner

   a)  i(x) = arg minj||x – wj ||

   b) )  i(x) = arg maxj||x – wj ||

Correct Answer  a)  i(x) = arg minj||x – wj ||

Q.18. To define a topological neigborhood that is neurobiologically correct will come under

a)Competition

b)Cooperation

c)Synaptic Adaptation

Correct Answer-   b)Cooperation

Q.19.To cluster the data of similar properties ------ algorithm is used

a)Linear

b)Logistic

c)SOM

d)Deep

Correct Answer-   c)SOM

Q.20. Characteristic feature of the SOM algorithm is that the size of the neighbourhood ---- with time.

a)Shrink

b)Expand

Correct Answer-   a)Shrink

Q.21.---------is affecting on speed of the gradient decent

a)Learning rate

b)Momentum

Correct Answer-   a)Learning rate

Q.22.For  back propogation algorithm -----architecture is used

a)Feed Forward

b)Feed backward

c)Recurrent

Correct Answer-   a)Feed forward

Q.23. Feature extraction and classification are integrated into one structure in -----architecture

a)ANN

 b)Deep learning

 c)SOM

Correct Answer-   b)Deep learning

Q.24.-----architecture is relatively invariant to geometric, local distortions in the image.

a )ANN

  b)Deep learning

c)SOM

Correct Answer-   b)Deep learning

Q25.Most suitable algorithm for  face detection, and face recognition is ----

a )ANN

  b)Deep learning

 c)SOM

Correct Answer-   b)Deep learning

Q.26.After Convolution of cnn network layer size---------- depending  on filter size

a)Increase

 b)Decreases

c)Do not change

Correct Answer-   b)Decrease

Q.27.You  are building a binary classifier for recognizing apple(Y=1)vs watermelon(Y=0),which one of these activation function would you recommend for output layer

a)Relu

b)Leaky relu

c)Sigmoid

d)Tanh

Correct Answer- c)Sigmoid

Q.28.Image of cat recognition is an example of structured data

a)True

b)False

Correct Answer-  b)False

Q.29Which of the reason Deep learning recently taking off

a)Access  to have a lot of data

b)It’s a brand new field

c)Architecture is very small

 

Correct Answer- a)Access  to have a lot of data 

Q.30.In deep neural network performance ----- for large data size

a)Improves

b)Decreasea

c)Saturated

Correct Answer-  a)Improves

Q.31.If image size is 6*6 and filter size is 3*3 what will be the output size after convolution

a)4*4

b)3*3

c)2*2

Coect Answer-  a)4*4

Q.32.In case of deep neural network after max polling new image size is -----

a)Increases

b)Decreases

Correct Answer-  a)Decreases

Q.33.  What are the steps for using a gradient descent algorithm?

a)Calculate error between the actual value and the predicted value

b)Reiterate until you find the best weights of network

c)Pass an input through the network and get values from output layer

d)Initialize random weight and bias

e)Go to each neurons which contributes to the error and change its respective values to reduce the error

a) a, b, c, d, e

b) e, d, c, b, a

c) c, b, a, e, f

d)d, c, a, e, b

Correct Answer- d)d, c, a, e,b

Q.34. The backpropagation law is also known as generalized delta rule, is it true?
a) Yes
b) No

Correct answer - a) Yes

Q.35.Which of the following neural networks uses  supervised learning

a)Multilay perceptron
b)Self organizing feature map
c) Hopfield network

Correct Answer- a ) Multilayer perceptron

Q.36. What are the general tasks that are performed with backpropagation algorithm?
a)Pattern mapping
b)function approximation
c)prediction
d) all of the mentioned
Correct answer - d) all of the mentioned

Q.37.What is the objective of backpropagation algorithm?
a) to develop learning algorithm for multilayer feedforward neural network
b) to develop learning algorithm for single layer feedforward neural network
c) to develop learning algorithm for multilayer feedforward neural network, so that network can be trained to capture the mapping implicitly
d) none of the mentioned

Correct answer- c) to develop learning algorithm for multilayer feedforward neural network, so that network can be trained to capture the mapping implicit

Q.38. Identify the following activation function :φ(V) = Z + (1/ 1 + exp (– x * V + Y) ),Z, X, Y are parameters

a)Step function

b)Ramp function

c)Sigmoid function

d)Gaussian function

Correct answer c)Sigmoid function

Q.39.An artificial neuron receives n inputs x1, x2, x3............xn with  weights w1, w2, ..........wn attached to the input links. The weighted sum_________________ is computed to be passed on to a non-linear filter  Φ called activation function to release the output.

a)Σ wi

b)Σ xi

c)Σ wi + Σ xi

d)Σ wi* xi

Correct answer - d)Σ wi* xi

Q.40 which of the following are the parameters of neural network

a)weights

b)gradient

c)error

Correct answer a)weights

Q.41.--------algorithm is suitable for audio signal recognition

a)Deep learning

b)Neutal  network

c)Recurrent network

d)Linear

Correct answer –a)Deep learning

Q.42.If dimension of matrix is 4*4 and filter is 2*2 with stride 2 what will be the output dimension matrix after max polling

a)2*2

b)3*3

correct answer- a)2*2

Q.43. Increase in size of a convolutional kernel would necessarily increase the performance of a convolutional neural network. 

a) True

b) False

Correct answer-b) False

Q.44.  Backpropagation works by first calculating the gradient of ___ and then propagating it backwards. 

a) Sum of squared error with respect to inputs

b) Sum of squared error with respect to weights

c) Sum of squared error with respect to outputs

d) None of the above

 Correct answer- c) Sum of squared error with respect to outputs

Q.45.  Which of the following is a bottleneck for deep learning algorithm?

a) Data related to the problem

b) CPU to GPU communication

c) GPU memory

d) All of the above

 Correct answer d) All of the above

Q.46.Sub sampling of pixels in deep neural network will not change image size

a)True

b)False

Correct answer - b)False

Q.47. The convolution and sub-sampling layers are considered as 2-D layers

a)True

b)False

Correct answer - a)True

Q.48. There are functions you can compute with a “small” L-layer deep neural network that shallower networks require exponentially more hidden units to comput.

a)True

b)False

Correct answer- a)True

Q.49.--------open sourse tool is widely used in machine learning algorithm

a)c++

b)java

c)Python

Correct answer - c)Python

Q.50.To cluster data ----learning is used

a)Supervised

b)Unsupervised

Correct answer - b)Unsupervised