next up previous contents index
Next: Dictionaries and tolerant retrieval Up: The term vocabulary and Previous: Combination schemes   Contents   Index


References and further reading

Exhaustive discussion of the character-level processing of can be found in Lunde (1998). Character bigram indexes are perhaps the most standard approach to indexing Chinese, although some systems use word segmentation. Due to differences in the language and writing system, word segmentation is most usual for Japanese (Luk and Kwok, 2002, Kishida et al., 2005). The structure of a character $k$-gram index over unsegmented text differs from that in Section 3.2.2 (page [*]): there the $k$-gram dictionary points to postings lists of entries in the regular dictionary, whereas here it points directly to document postings lists. For further discussion of Chinese word segmentation, see Tseng et al. (2005), Sproat and Emerson (2003), Sproat et al. (1996), and Gao et al. (2005).

Lita et al. (2003) present a method for truecasing . Natural language processing work on computational morphology is presented in (Sproat, 1992, Beesley and Karttunen, 2003).

Language identification was perhaps first explored in cryptography; for example, Konheim (1981) presents a character-level $k$-gram language identification algorithm. While other methods such as looking for particular distinctive function words and letter combinations have been used, with the advent of widespread digital text, many people have explored the character $n$-gram technique, and found it to be highly successful (Beesley, 1998, Dunning, 1994, Cavnar and Trenkle, 1994). Written language identification is regarded as a fairly easy problem, while spoken language identification remains more difficult; see Hughes et al. (2006) for a recent survey.

Experiments on and discussion of the positive and negative impact of stemming in English can be found in the following works: Salton (1989), Krovetz (1995), Hull (1996), Harman (1991). Hollink et al. (2004) provide detailed results for the effectiveness of language-specific methods on 8 European languages. In terms of percent change in mean average precision (see page 8.4 ) over a baseline system, diacritic removal gains up to 23% (being especially helpful for Finnish, French, and Swedish). Stemming helped markedly for Finnish (30% improvement) and Spanish (10% improvement), but for most languages, including English, the gain from stemming was in the range 0-5%, and results from a lemmatizer were poorer still. Compound splitting gained 25% for Swedish and 15% for German, but only 4% for Dutch. Rather than language-particular methods, indexing character $k$-grams (as we suggested for Chinese) could often give as good or better results: using within-word character 4-grams rather than words gave gains of 37% in Finnish, 27% in Swedish, and 20% in German, while even being slightly positive for other languages, such as Dutch, Spanish, and English. Tomlinson (2003) presents broadly similar results. Bar-Ilan and Gutman (2005) suggest that, at the time of their study (2003), the major commercial web search engines suffered from lacking decent language-particular processing; for example, a query on www.google.fr for l'électricité did not separate off the article l' but only matched pages with precisely this string of article+noun.

The classic presentation of for IR can be found in Moffat and Zobel (1996). Extended techniques are discussed in Boldi and Vigna (2005). The main paper in the algorithms literature is Pugh (1990), which uses multilevel skip pointers to give expected $O(\log P)$ list access (the same expected efficiency as using a tree data structure) with less implementational complexity. In practice, the effectiveness of using skip pointers depends on various system parameters. Moffat and Zobel (1996) report conjunctive queries running about five times faster with the use of skip pointers, but Bahle et al. (2002, p. 217) report that, with modern CPUs, using skip lists instead slows down search because it expands the size of the postings list (i.e., disk I/O dominates performance). In contrast, Strohman and Croft (2007) again show good performance gains from skipping, in a system architecture designed to optimize for the large memory spaces and multiple cores of recent CPUs.

Johnson et al. (2006) report that 11.7% of all queries in two 2002 web query logs contained phrase queries , though Kammenhuber et al. (2006) report only 3% phrase queries for a different data set. Silverstein et al. (1999) note that many queries without explicit phrase operators are actually implicit phrase searches.


next up previous contents index
Next: Dictionaries and tolerant retrieval Up: The term vocabulary and Previous: Combination schemes   Contents   Index
© 2008 Cambridge University Press
This is an automatically generated page. In case of formatting errors you may want to look at the PDF edition of the book.
2009-04-07