Skip to content

Latest commit

 

History

History
119 lines (106 loc) · 3.79 KB

README.md

File metadata and controls

119 lines (106 loc) · 3.79 KB
Raw

InformationRetrieval

This repo consists of all the assignments, and, tasks of Information Retrieval course of FAST NUCES Spring 2023.

Assignment 1

Implementation of Boolean Retrieval Model

Link to Article: Boolean Retrieval Model

Introduction

The Boolean retrieval model is a search model that retrieves documents that match a Boolean expression (a query) of terms, where the terms are connected by Boolean operators (AND, OR, NOT).

How to run?

  • install python 🐍 latest version and set it up on your computer.
  • clone this repository in a specific folder.
  • unzip the folder and open the assignment folder in a preferred code editor or IDE.

  • Install Dependencies

    • pip install nltk
    • pip install unidecode
    • pip install pycontractions
    • pip install AST
    • pip install tk

  • For Linux

    python3 main.py

  • For Windows

    python main.py

Queries Format

Simple Query

  • word

Complement Query

  • not word

Intersection Query

  • word1 and word2
  • word1 and word2 and word3
  • not word1 and word2
  • word1 and not word2
  • not word1 and not word2
  • not word1 and word2 and word3
  • word1 and not word2 and word3
  • word1 and word2 and not word3
  • not word1 and not word2 and not word3

Union Query

  • word1 or word2
  • word1 or word2 or word3
  • not word1 or word2
  • word1 or not word2
  • not word1 or not word2
  • not word1 or word2 or word3
  • word1 or not word2 or word3
  • word1 or word2 or not word3
  • not word1 or not word2 or not word3

Mixed Query

  • It includes query with mixed boolean opearators (AND, OR, NOT) and word limit upto 3 words max.

Proximity Query

  • word1 word2 \k, here 'k' represents the no. of words word2 is distant from word1.

Assignment 2

Implementation of Vector Space Model

Introduction

The Vector Space Model is a commonly used information retrieval technique where documents are represented as vectors in a high-dimensional space and ranked based on their similarity to a user's query.

How to run?

  • install python 🐍 latest version and set it up on your computer.
  • clone this repository in a specific folder.
  • unzip the folder and open the assignment folder in a preferred code editor or IDE.

  • Install Dependencies

    • pip install nltk
    • pip install unidecode
    • pip install pycontractions
    • pip install AST
    • pip install tk

  • For Linux

    python3 main.py

  • For Windows

    python main.py

Results

The results are written after applying threshold using alpha value= 0.35

Results Format: Query
Ranked Documents
Respective Cosine Similarities

alpha value=0.35
threshold formula= max(cosine sim)* alpha value

  • cricket politics
    [5, 26, 29, 14, 25, 3, 24]
    [0.10483768357856904, 0.0733026052948556, 0.05747687349155568, 0.051814006141766254, 0.04669640019650995, 0.044924892358261706, 0.04204193256553604]
  • dharamsala to indore
    [17]
    [0.5669900905084869]
  • retirement
    [14]
    [0.05884903818579923]
  • test captain
    [3, 6, 21, 17]
    [0.2169788724831718, 0.16412707621815958, 0.12857575540015573, 0.109537249881213]
  • pcb psl
    [11, 29]
    [0.3987613437849497, 0.23786386254409053]
  • hate
    []
    []
  • bowling coach
    [29, 6, 24]
    [0.15946207811500823, 0.10025156193891573, 0.06560991594156915]
  • relative comfort
    [19, 11, 2, 7]
    [0.034858940263823475, 0.03152703135309784, 0.013545725571459473, 0.0126389307857281]
  • possible
    [23, 15]
    [0.05534812091736982, 0.051743525151666955]
  • batter bowler
    [2, 3, 16, 15, 30, 9]
    [0.09290690948714346, 0.08726633903001002, 0.05206958459212497, 0.03995088457437994, 0.03776960712447397, 0.032756883132567645]