#
*Machine
learning quiz questions TRUE or FALSE with answers, important machine
learning interview questions for data science, Top 3 machine learning
question set*

##
__Machine Learning TRUE / FALSE Questions - SET 04__

###
*1. Training neural
networks has the potential problem of overfitting the training data.*

*1. Training neural networks has the potential problem of overfitting the training data.*

(a)
TRUE (b)
FALSE

**View Answer**Answer: TRUE##
Overfitting of
the training data happens if neural network model is suffering from high
variance. It means the trained parameters fits the training set well, but
performs poorly when tested on “unseen” data (the training or the validation
set).Solutions:More training
dataReducing the
number of hidden layersIncreasing
regularization parameter |

###
*2. A support vector
machine computes P(y|x).*

*2. A support vector machine computes P(y|x).*

(a)
TRUE (b)
FALSE

**View Answer**Answer: FALSESupport Vector
Machine is a linear model for classification and regression problems. SVM is
an algorithm that takes the data as an input and outputs a line that
separates those classes if possible.##
Objective of SVMThe objective of
the support vector machine algorithm is to find a hyperplane in an
N-dimensional space (N — the number of features) that distinctly classifies
the data points. |

###
*3. One drawback of
maximum likelihood estimation is that in some scenarios (for example,
multinomial distribution), it may return probability estimates of zero.*

*3. One drawback of maximum likelihood estimation is that in some scenarios (for example, multinomial distribution), it may return probability estimates of zero.*

(a)
TRUE (b)
FALSE

**View Answer**Answer: TRUE##
One drawback of
Maximum Likelihood Estimation (MLE) is that in some scenarios it may return
zero probability estimates. This happens when we try to evaluate MLE models
on unseen data. This may not
happen with equi-probable events like coin flips, dice etc. It usually occurs
in language models in Natural Language Processing.Example:Zero
probabilities are clearly a problem in language models, such as when
predicting the next word in a speech recognition application, because many
words will be sparsely represented in the training data. In such cases, the
next word may be unseen. Hence, this may end up in zero probability value. |

**************************

###
**Related links:**

**Related links:**