Scribd
Upload
129 views2 pages

Solution Assignment2

This document summarizes key aspects of an assignment on information retrieval techniques. It includes 3 tasks - language modeling using maximum likelihood estimation to rank documents by probability, evaluating recall and precision of retrieval systems, and calculating interpolated precision at various recall levels. Equations and sample calculations are provided for ranking queries based on language models and evaluating retrieval performance metrics.

Uploaded by

xyz xyz25
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
129 views2 pages

Solution Assignment2

This document summarizes key aspects of an assignment on information retrieval techniques. It includes 3 tasks - language modeling using maximum likelihood estimation to rank documents by probability, evaluating recall and precision of retrieval systems, and calculating interpolated precision at various recall levels. Equations and sample calculations are provided for ranking queries based on language models and evaluating retrieval performance metrics.

Uploaded by

xyz xyz25
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

TDT4117 Information Retrieval - Autumn 2017

Solution - Assignment 2

Task 2 - Language Model


2. MLE

!
Since λ = 0.5, the formula can be transformed to: ∗ (𝑃𝑚𝑙𝑒(𝑡|𝑀𝑑) + 𝑃𝑚𝑙𝑒(𝑡|𝐶))
"

q1 = hurricane

! ! "
p(q1,d1) = ∗ ( + ) = 0.11
" 1 "2
! 3 "
p(q1,d2) = ∗ ( + ) = 0.04
" !" "2
! ! "
p(q1,d3) = ∗ ( + ) = 0.12
" 4 "2
rank: d3, d1, d2

q2 = hurricane florida

Uses the values for “hurricane” from q1.


! ! "
p(q2,d1) = 0.11 * ∗ ( + ) = 0.01
" 1 "2
! ! "
p(q2,d2) = 0.04 * ∗( + ) = 0.0033
" !" "2
! 3 "
p(q2,d3) = 0.12 * ∗( + ) = 0.0048
" 4 "2
rank: d1, d3, d2

q3 = hurricane katrina

! 3 3
p(q3, d1) = 0.11 * ∗ ( + ) = 0
" 1 "2
Katrina doesn’t appear in d2 or d3 either:
p(q3, d2) = 0
p(q3, d3) = 0
rank: d1 = d2 = d3

Task 3 - Evaluation of IR Systems


2. Precision and recall

Ret Rel Recall Precision


83 - 0/12 = 0 0/1 = 0
2 Yes 1/12 = 0.083 1/2 = 0.5
76 - 1/12 = 0.083 1/3 = 0.33
42 Yes 2/12 = 0.167 2/4 = 0.5
5 Yes 3/12 = 0.25 3/5 = 0.6
9 - 3/12 = 0.25 3/6 = 0.5
23 - 3/12 = 0.25 3/7 = 0.43
33 Yes 4/12 = 0.33 4/8 = 0.5
96 Yes 5/12 = 0.417 5/9 = 0.56
101 - 5/12 = 0.417 5/10 = 0.5
Task 4 - Interpolated Precision
2. Interpolated Precision

Recall iPrecision
0 0.6
1/12 0.6
2/12 0.6
3/12 0.6
4/12 0.56
5/12 0.56
6/12 0
7/12 0
8/12 0
9/12 0
10/12 0
11/12 0
12/12 0

Graph:

You might also like