Advanced Database Management System - Tutorials and Notes: Solved Questions and Answers in NLP 03

# Natural language processing quiz questions with answers, NLP true false interview questions, NLP quiz questions for competitive exams

## NLP TRUE/FALSE Quiz Questions - SET 03

1. Models that assign probabilities to sequences of words are called Support Vector Models.
(a) TRUE                                          (b) FALSE

 Answer: (b) FALSE Models that assign probabilities to sequences of words are called language models or LMs.

2. Number of trigrams in the following sentence is 4; “calculates the similarity between two strings”.
(a) TRUE                                          (b) FALSE

 Answer: (a) TRUE There are four trigrams (3-grams) in the given sentence if we do not include the start and end markers. The trigrams are; “calculates the similarity”, “the similarity between”, “similarity between two” and “between two strings”.

3. We normalize the counts of words in an n-gram model to make the value to fall between 0 and 100.
(a) TRUE                                          (b) FALSE

 Answer: (b) FALSE We normalize the counts of words in an n-gram model to make the value to fall between 0 and 1. We get the maximum likelihood estimation (MLE) for the parameters of an n-gram model by getting counts from a normalize corpus, and normalizing the counts so that they lie between 0 and 1.

4. To calculate the bigram probability of a word wn given the previous word wn-1, we count the occurrence of word sequence “wn-1 wn” and normalize this by the count of wn-1.
(a) TRUE                                          (b) FALSE

 Answer: (a) TRUE To compute a particular bigram probability of a word y given a previous word x, we will compute the count of the bigram C(x y) and normalize by the count of unigram C(x). P(y|x) = count(y x)/count(y) This is called Maximum Likelihood Estimate (MLE).

5. It is better to compute the probabilities in a language model as log probabilities.
(a) TRUE                                          (b) FALSE

 Answer: (a) TRUE Since probabilities are (by definition) less than or equal to probabilities 1, the more probabilities we multiply together, the smaller the product becomes. Multiplying enough n-grams together would result in numerical underflow. By using log probabilities instead of raw probabilities, we get numbers that are not as small.

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