AD Teaching Wiki:

Results for Exercise Sheet 2 (Ranking)

These results should be based on the file movies.test-benchmark.tsv. Click "Edit" in the upper bar of this page and add your row to the table below, following the examples already there. Please write down all values with exactly two digits of precision. In the first column, write your account name.

Name

MP@3

MP@R

MAP

BM25 Parameters

Refinements

np1042

0.73

0.54

0.55

b=0.2, k=0.8

None (baseline)

mr488

0.72

0.57

0.56

b=0.1, k=1

log(imdb)

nw164

0.76

0.56

0.55

b=0.1, k=1.05

Penalty for short and common words, removed words from query, bonus for higher ranked movies

mb791

0.61

0.56

0.55

b=0.12, k=0.9

log(# imdb votes), removed words from queries

do67

0.79

0.54

0.55

b=0.2, k=0.8

multiply with number votes mapped to a number between 1 and 2

as1936

0.79

0.57

0.58

b=0.15, k=0.85

Filtering commonly used words, bonus if the words in query are part of the title, adding the imdb score, adding the log(number_of_votes)

mn279

0.81

0.54

0.55

b=0.1, k=0.75

bonus if the words in query are part of the title, removed words from queries

jm700

0.79

0.56

0.56

b=0.08, k=0.9

ignore words with 2 characters or less

gf52

0.75

0.54

0.56

b=0.15, k=1.35

added blacklist of common english words, included parameters for weighting the impact of IMDb votes, ratings and Wikimedia pages

ek223

0.76

0.55

0.56

b=0.08, k=0.70

ignore short words, reduce influence of common words

cw441

0.85

0.54

0.56

b=0, k=0.5

ignore short words, bonus for higher ranked movies

ls1369

0.70

0.56

0.53

b=0.1, k=0.9

remove single word delimeter words from queries and add the log of the iMDB votes to the score

ap367

0.79

0.54

0.54

b=0.04, k=1.15

None

cd100

0.76

0.51

0.54

b=0.09, k=0.81

10% score boost for docs containing all words (which unfortunately seems to be bad for the metrics)

bo30

0.84

0.52

0.52

b=0.02, k=0.59

ignore short words, random search for b and k values

bs249

0.79

0.55

0.53

b=0, k=1.1

None

sa328

0.75

0.60

0.61

b=0.3, k=0.85

None

hv30

0.73

0.54

0.56

b=0, k=1.15

ignore short words, blacklist for some common words, reward containing multiple words, log(number of Wikimedia pages)

tg241

0.76

0.55

0.55

b=0.03, k=1.35

ignore short and common words, boost score for movies with high popularity, imdb_rating and votes, boost queries with all keywords

sr530

0.72

0.55

0.55

b=0.04, k=1.15

None

eh169

0.70

0.52

0.49

b=0.25, k=1

Removed short words, multiplied score by rating and popularity

kd130

0.76

0.53

0.54

b=0, k=1.1

Removed common words

rs532

0.76

0.53

0.56

b=0.13, k=0.75

None

ch557

0.67

0.52

0.5

b=0.09, k=0.8

Removed words from queries

mz226

0.85

0.55

0.56

b=0.08, k=0.9

Removed short words from queries, multiplied document scores by a value depending on the number of votes

ak913

0.85

0.56

0.56

b=0, k=1.1

Removed short words from queries, used bonus score from votes, rating and no.of wiki pages

jh594

0.76

0.55

0.53

b=0.008, k=1.11

None

je249

0.72

0.52

0.53

b=0.001, k=1

Removed all words shorter than 4

bs426

0.76

0.54

0.54

b=0.2, k=0.8

Removed the 100 most common english words

ja162

0.81

0.55

0.56

b=0.05, k=0.9

Slight boost of results containing all keywords

me284

0.76

0.53

0.55

b=0.14, k=0.96

None

zh11

0.82

0.56

0.54

b=0.0, k=0.9

None (baseline)

ec141

0.78

0.54

0.53

b=0.0, k=1.0

None (baseline)

jn159

0.42

0.41

0.37

b=0.1, k=1.5

None

yh77

0.76

0.57

0.57

b=0.0, k=1.2

Removed words from queries, word in title 3x times more important than in description

sp360

0.79

0.55

0.54

b=0.001, k=1.1

None

mb1431

0.85

0.56

0.55

b=0.01, k=0.77

Bonus points for highly ranked movies

dl121

0.61

0.48

0.46

b=0.5, k=1.3

None

mm1460

0.79

0.55

0.55

b=0.40, k=2.00

Removed some words, added some words, added score as ln(IMDb)

sa347

0.85

0.56

0.54

b=0.001, k=0.75

None (tried removing words shorter than 3, adding synonym for "movies" & "films", and multiplying rating by rank, but results did not improve

kn115

0.67

0.49

0.47

b=1, k=0.1

Removed common words and adding synonyms

ca189

0.85

0.56

0.53

b=0.01, k=0.7

None

bg162

0.79

0.57

0.58

b=0.2, k=1.0

Removed "redundant" words, used ratings and popularity in the score

sb986

0.76

0.56

0.56

b=0.03, k=0.9

Removed words less than 2 letters from queries

rm243

0.81

0.55

0.53

b=0, k=0.9

none

aa272

0.82

0.56

0.57

b=0.0, k=0.9

Remove pairs of words, taking certain grammatical structures into account. Remove short words.

dg262

0.73

0.54

0.46

b=0.2, k=0.9

None

eg128

0.73

0.51

0.51

b=0.05, k=1.75

r * (log(num_ratings) + rating) + (1 - r) * bm25, r=0.2

rs476

0.75

0.54

0.53

b=0.01, k=1.2

None

gm133

0.79

0.53

0.54

b=0.1, k=0.7

Removed words shorter than 4 characters

js1344

0.58

0.50

0.49

b=0, k=0.95

remove: short words, generic terms (“movie”, “film”), add: (poor man's) pluralized version, (poor man's) de-pluralized version

ts574

0.58

0.41

0.42

b=0.2, k=0.8

None

hm132

0.78

0.54

0.54

b=0.2, k=0.85

removed short words

mh821

0.82

0.56

0.54

b=0, k=0.9

None

as1699

0.85

0.54

0.54

b=0.001, k=0.68

Removed short words

AD Teaching Wiki: InformationRetrievalWS2223/ResultsES2 (last edited 2023-03-27 17:30:43 by 10)