Soal UTS Information Retrieval

UJIAN TENGAH SEMESTER GANJIL 2009/2010

PROGRAM PINDAHAN KELAS D3 KE S1

Mata Kuliah : Information REtrieval [CS4323]

Hari, Tanggal : Kamis, 5-November-2009

Waktu : 75 Menit

Sifat : Tutup buku

Dosen : YFA

INSTRUKSI : Kerjakan dan jawablah pertanyaan-pertanyaan berikut pada lembar jawaban ujian

1. Berpengaruhkah proses indexing dalam sebuah Sistem Information Retrieval secara keseluruhan ? Berikan uraian yang membuktikan analisis anda tersebut!

2. Pencarian return hits dengan metode similarity antara qury dan dokumen dapat dilakukan dengan proses pembobotan maupun tanpa pembobotan.

a. Berikan penjelasan dari kedua proses tersebut!

b. Buatlah analisis tentang kapan proses pembobotan harus dilakukan maupun tidak dilakukan pembobotan dalam information retrieval!

3. Suppose we have a topic (i.e., query) with a total of 10 relevant document in a collection with 100 documents. A system has retrieved 6 documents whose relevance status is [+;-;-;+;-;+] in the order of ranking. A “+” (or”-”) indicates that the corresponding documents is relevant (or non-relevant). For example, the first document is relevant, while the second is non-relevant, etc. Compute the precision and recall for this result.

4. Now suppose our collection has total of 10.000 documents, instead of 100 documents, but the topic has the same number of relevant documents (i.e., 10) in the collection, and the system has retrieved exactly the same result as in the previous question for this topic. Intuitively, the system’s performance is better in this case than in the previous case since now we have more distracting non-relevant documents and thus the retrieval task is much harder. Explain why precision and recall can not replect this difference, and propose a reasonable new measure that would reflect this diference. How important is it to capture this diference in our evaluation of retrieval method ?