Friday, January 10, 2014

IS2140_Reading notes_Unit 1

Date: 01/10/2014

1, definition of FOA

The process of Finding Out About (FOA) is a series of cognitive activities, which helps decision-makers focus on knowledge relevant to the topic of search and operate on the large collections of linguistic objects by FOA searching techniques.  The semantic meanings of linguistic objects are essential for FOA.  We usually draw on others' opinion by communicating with others, reading valuable writings in library,   searching the Internet and so on. Well-arranged linguistic expressions enable people to understand knowledge wholly and represent own opinion more clear.

2, the process of FOA

The FOA conversion loop includes three phases:
(1) Asking a question;
  Users turn their information need into external expression of questions, proposing query.
(2) Constructing an answer;

   Search engine is a computer program that algorithmically performs the task to answer users' questions or query.

(3) Assessing the answer.

This is a closing of the FOA conversion loop. Making use of asker's relevance feedback fully provides more information with their reaction to each retrieved document.


3, Information Retrieval
IR is concerned with representing, searching, and manipulating large collections of electronic text and other human-language data.

4, IR Applications:
(1) Web Search Engine
Web search engines are….., based on clusters of computers which work cooperatively to generate a ranked list of Web pages without redundant and duplicate pages to satisfy users' information need  embodied in the query. The rank of the Web pages usually abide by the search engine 's ranking algorithm, which keeps several features including the content and structure of pages, the relation to other pages, the content and structure of the Web as a whole, and characteristics of user such as geographic location or past searching behavior in balance.

(2)Desktop and File System Search Engines
(3)Document Management and Search Service

(4)Digital Libraries and Other Specialized IR Systems

(5)Other IR Applications Associated with Storage, Manipulation, and Retrieval of Human-Language    Data
For example, news aggregator, e-mail system, text clustering and categorization systems, summarization systems, information extraction systems, Topic detection and tracking systems, expert search systems, question answering systems, and multimedia information retrieval systems.

5, Basic Information Retrieval Systems Architecture

User constructs a topic depending on the information need, and issues a query which consists of terms to the IR system. Then a search engine accepts user's queries and processes them by maintaining and manipulating an inverted index for a document collection, which forms the principal data structure used by the search engine.  The search engine sorts relevant documents according to the score or the retrieval status value (RSV) for each document. After eliminating the redundant and duplicate results, the search engine reports the ranked result list for further processing. 

6, IR System Performance Evaluation

There are two aspects: efficiency and effectiveness.
Efficiency may be measured in terms of time and space.
Effectiveness may be measured in terms of relevance, depending on human judgment.


7,IR Problem
There are two crucial problems:
(1) How to exactly extract information from documents;
(2) How to decide the relevance which depends on use's assessment for solving query.

No comments:

Post a Comment