PATENTSCOPE: latest developments
April 16, 2012
The PATENTSCOPE team is pleased to announce that
PATENTSCOPE was
migrated to the latest major version of the LUCENE search engine. We apologize for the few glitches
that some of you noticed in the last three weeks and that are now fixed. Please find below a
detailed description of the latest developments of the PATENTSCOPE search system:
1.
Response time: response times for queries have been improved significantly, especially for
complex queries and queries containing not null filters created using the “Search by Field
Combination” screen.
2.
New and enhanced operators: the NEAR operator has been extended so that you can combine it
with other Boolean operators. Moreover, if you specify an integer at the end of the NEAR operator,
it will be used as the maximum distance allowed between the predicates. Queries like the following
are now supported:
EN_ALLTXT:
((((“sound” OR “acoustic” OR “noise” OR “sonic” OR “voice”) NEAR3 (“insulation” OR “isolation”
or “proofing” OR “barrier”)) OR ( “soundproof” OR “silencing” ))
NEAR
(“wall” OR “panel” OR “bulkhead” OR “brickwork” OR “delimiting structure”))
EN_ALLTXT:( stock BEFORE exchange)
3.
Robust highlighting and hit maps: the highlighting mechanism is now robust and highlights
consistently results, taking into account the matched keyword positions in the document, as well as
stemming and fuzzy searching with the wildcard characters. Please note that keywords are
highlighted only:
- in the field/s in which the query was entered, for example: if you search disposable in the abstract field, disposable will only be highlighted in abstracts of retrieved patent documents, not in the titles, descriptions or claims.
- if their distribution in the text of patent documents matches the query: for example, if you search car NEAR electric, car will only be highlighted if electric is within a distance of 5 words from car in retrieved patent documents and vice versa (in the PATENTSCOPE search system, the default range within which search terms joined with the NEAR operator must lie is 5 words.
In addition, when displaying the description / claims or Full Text tab for a patent
application, a new icon is displayed
. Clicking on this icon displays a
map of the highlighted keywords in the text, allowing to quickly jump to the sections of interest
in the text.
4.
Display of Query trees: a new icon
is displayed next to the RSS icon in the result
list. If you click on it, the system parses again your last query, decomposes your query in
sub-clauses and executes each sub-clause one by one, letting you know the associated number
of intermediate results. For example, when searching “electric toothbrush” in abstracts in all
languages with the
CLIR interface, the query is the following:
EN_AB:("electric toothbrush") OR DE_AB:("elektrischen Zahnbürste" OR "Elektrozahnbürste") OR
ES_AB:("cepillo de dientes eléctrico" OR "cepillo dental eléctrico" OR "cepillo eléctrico dental")
OR FR_AB:("brosse à dents électrique") OR JA_AB:("電動歯ブラシ" OR "を用いた電動電子歯ブラシ" OR "これを用いた電動電子歯ブラシ") OR
KO_AB:("전동칫솔" OR "전동 칫솔용") OR PT_AB:("escova de dente elétrica") OR RU_AB:("электрическая зубная
щетка" OR "для электрической зубной щетки") OR ZH_AB:("电动牙刷")
and by clicking the query tree, the distribution per language of the 896 results is
displayed:
- 613 are in English
- 153 are in German
- 150 in Spanish
- 474 in French
- 24 in Japanese
- 5 in Korean
- 3 in Portuguese
- 48 in Russian
- 0 in Chinese