UIUC大学之Coursera课程Text Retrieval and Search Engines：Week 2 Practice Quiz

Week 2 Practice QuizHelp Center

Warning: The hard deadline has passed. You can attempt it, but you will not get credit for it. You are welcome to try it as a learning exercise.

In accordance with the Coursera Honor Code, I  certify that the answers here are my own work.

Question 1

Suppose a query has a total of 5 relevant documents in a collection of 100 documents. System A and System B have each retrieved 10 documents, and the relevance status of the ranked lists is shown below: 

	System A: [+ + - - - - - - - -]
	System B: [- + - - + - - - - +]

where the leftmost entry corresponds to the highest ranked document, and the rightmost entry corresponds to the lowest ranked document. A “+” indicates a relevant document and a “-” corresponds to a non-relevant one. For example, the top ranked document retrieved by System A is relevant whereas the top ranked one by B is non-relevant. 

What is the  precision at 10 documents of both systems?

P(A) = 2/5  P(B)= 3/5

P(A) = 8/100  P(B)= 7/100

P(A) = 2/100  P(B)= 3/100

P(A) = 2/10  P(B)= 3/10

Question 2

Assume the same scenario as in Question 1. What is the  recall of both systems?

R(A) = 2/10  R(B)= 3/10

R(A) = 8/100  R(B)= 7/100

R(A) = 2/5  R(B)= 3/5

R(A) = 2/100  R(B)= 3/100

Question 3

Assume the same scenario as in Question 1. What is the  average precision of both systems?

AP(A) = 3/10  AP(B) = 9/20

AP(A) = 2/100  AP(B) = 3/250

AP(A) = 2/5  AP(B) = 6/25

AP(A) = 2/10  AP(B) = 3/25

Question 4

Let w 1, w 2, and w 3 represent three words in the dictionary of an inverted index. Suppose we have the following document frequency distribution: 

Word	Document Frequency
w1	1
w2	5
w3	10

Assume that each posting entry of document ID and term frequency takes exactly the same disk space. Which word's postings list will occupy the largest disk space?

w1

w3

w2

Question 5

Assume we have the same scenario as in Question 4. If we enter a query Q= “w 1 w 2 w 3” then the  maximum possible number of accumulators needed to score all the matching documents is:

5

16

10

1

Question 6

Assume that d-gap between two documents is equal to 9. If you want to compress this d-gap with a  gamma code, what will be the binary representation of the code?

1110001

1110000

1110010

1110011

Question 7

Assume you have two retrieval systems X and Y. If X has a higher MAP (mean average precision), can Y have a higher gMAP (geometric mean average precision)?

Yes

No