→ cindy6603 :已收錄 :D 12/15 17:14
課程名稱︰資訊檢索
課程性質︰圖資系大三必修
課程教師︰唐牧群
開課學院：文學院
開課系所︰圖資系
考試日期（年月日）︰98.12.15
考試時限（分鐘）：150分鐘
是否需發放獎勵金：是
（如未明確表示，則不予發放）
試題 :
Here is an imaginary database that contains the following 5 documents:
D1: "a dog barks at a cat and it fell from a tree"
D2: "an ant watches ants on the bark of a tree"
D3: "a dog watches another dog barks a cat"
D4: "a dog barks at a cat watches another cat"
D5: "the bark fell from the tree as an ant watches"
(Terms in the stop-word list have been grayed out.)
1a. Calculate DF(Document Frequency) and IDF(Inverse Document Frequency)
weight for each index term(simply use N/n without logarithm).
1b. Create an inverted file for the database where each cell contains the
TF*IDF weight of each term in the documents(treat singulars and plurals
as the same word stem).
1c. What will be the search results look like by a system based on Boolean
and vector space model, respectively. To rank the search results using
vector space model, first calculate relevance scores(inner product without
document length normalization) and rank the documents accordingly.
1d. After the initial iteration, the user examines the results and marks D2,
D5 as relevant, and D3 and D4 as non-relevant. Produce the new ranking
using Rocchio's method where α=1.0 β=1.0 γ=1.0
2. Unlike data retrieval where perfect precision and recall are guaranteed,
information retrieval is more of a trial-and-error process. What might be
the possible causes behind the uncertainty of IR?
3. Define 4 of the following 5 concepts:
a.SIP(Statistically Improbable Phrases)
b.The "explode" function for PubMed search
c.The "subheading" function for PubMed search
d.Relevance feedback
e."Building block" search strategy
4. From information behavior research we know that users' "information needs"
or "search tasks" come in different shapes and sizes. List a few ways(two
or more) by which users' information needs or search tasks can be
categorized and comments on how users' needs or tasks might impact
information retrieval.
5. How is interactive view of IR different from traditional view of IR? What
doew it consider to be the most crucial component of IR process? Name one
or two interactive IR inspired search techniques that we have talked about
in the class.
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 140.112.180.107