Jumat, 08 April 2022

Koefisien Jaccard

 1. Jika diketahui A={1, 2, 3, 4}, B={1, 2, 4} dan C ={1, 2, 4, 5} berapakah Jaccard(A,B),

Jaccard(B,C) dan Jaccard(A,C)?


2. Berikutnya untuk kasus query dan document. Misalnya kita punya:

Q : ideas of march

Doc1 : Caesar died in march

Doc2 : the long march

Cari koefisien Jaccard antara query dengan doc1 dan doc2


3. Diketahui 3 dokumen:

D1 : Jack London traveled to Oakland

D2 : Jack London traveked to the city of Oakland

D3 : Jack traveled from Oakland to London

Nilai dari koefisien Jaccard J(d1,d2) dan J(d1,d3) jika dilakukan dengan n-gram analisis

dengan n=2(bigram) adalah


                                                                        Jawaban


1. Jaccard(A,B)

| A | = 4

| B | = 3

| A ∩ B | = 3

| A U B | = (|A| + |B| - | A ∩ B |) = 4 + 3 – 3 = 4

Jaccard(A,B) = | A ∩ B | / | A U B | = 3/4 = 0.75

Jaccard(B,C)

| B | = 3

| C | = 4

| B ∩ C | = 3

| B U C | = (|B| + |C| - | B ∩ C |) = 3 + 4 – 3 = 4

Jaccard(B,C) = | B ∩ C | / | B U C | = 3/4 = 0.75


Jaccard(A,C)

| A | = 4

| C | = 4

| A ∩ C | = 3

| A U C | = (|A| + |C| - | A ∩ C |) = 4 + 4 – 3 = 5

Jaccard(A,C) = | A ∩ C| / | A U C | = 3/5 = 0.6


2. Jaccard(Q, DOC1)

Q = 3

DOC1 = 4

| Q ∩ DOC1 | = 1

| Q U DOC1 | = 6

|Q ∩ DOC1| / | Q U DOC1| = 1/6 = 0.17

Jaccard(Q, DOC2)

Q = 3

DOC2 = 3

| Q ∩ DOC2 | = 1

| Q U DOC2 | = 5

| Q ∩ DOC2 | / |Q U DOC2| = 1/5 = 0.2


3. Jaccard(D1,D2)

D1 = 4 (Jack London, London traveled, traveled to, to Oakland)

D2 = 7 (Jack London, London traveled, traveled to, to the, the city, city of, of

Oakland)

| D1 ∩ D2 | = 3

| D1 U D2 | = 8

| D1 ∩ D2 | / | D1 U D2 | = 3/8 = 0.375


Jaccard(D1,D3)

D1 = 4(Jack London, London traveled, traveled to, to Oakland)

D2 = 5(Jack traveled, traveled from, from Oakland, Oakland to, to Londgon

| D1 ∩ D3 | = 0

| D1 U D2 | = 9

| D1 ∩ D2 | / | D1 U D2 | = 0/9 = 0