This question is on IR. Consider the Vector Space IR system discussed in lectures. For this...

Transcribed Image Text:

This question is on IR. Consider the Vector Space IR system discussed in lectures. For this question we use the following weighing scheme. If d is a document and t is a term, then N weight(t, d)= TFid log10 If q is a query and t is a term weight(t. q) is similarly defined. Recall that given a query q and a document d. v(q)-v(d) score (q, d)= || (q)|||v(d)|| Consider a document D₁ in which the word hash appears 10 times, table appears 3 times and has 10 other distinct words. Let D₂ be a document in which the word hash appears 10 times and the word table appears 2 times. When the query is hash table which document is more relevant? Assume that df, all the terms from D₁ and D₂ is N/10. Explain your answer. (20 Points) This question is on IR. Consider the Vector Space IR system discussed in lectures. For this question we use the following weighing scheme. If d is a document and t is a term, then N weight(t, d)= TFid log10 If q is a query and t is a term weight(t. q) is similarly defined. Recall that given a query q and a document d. v(q)-v(d) score (q, d)= || (q)|||v(d)|| Consider a document D₁ in which the word hash appears 10 times, table appears 3 times and has 10 other distinct words. Let D₂ be a document in which the word hash appears 10 times and the word table appears 2 times. When the query is hash table which document is more relevant? Assume that df, all the terms from D₁ and D₂ is N/10. Explain your answer. (20 Points)

Answer rating: 100% (QA)

Anger for vector space in IR system weighig scheme Coeight 1d x do... View the full answer