Content item retrieval based on a free text entry
View Patent ↗A method and apparatus for textual searching of a database is provided herein. During operation a user will input a letter into a search engine. The search engine will score words based on the letter and display results of the highest-scored words. Another letter will again be received and the process repeated. In situations where titles are returned to the user, additional steps of associating the words with a title and scoring the title take place. The highest-scored titles are provided to the user as the displayed results.
1. A method comprising the steps of:
receiving by an electronic processor, a textual input;
appending the textual input to a previously supplied textual input to form a search string;
accessing by the electronic processor, a database comprising a plurality of words; and
determining by the electronic processor, a score (sword) for words in the database based on the search string, wherein
sword=(μ 1 ( x )+μ 2 ( x )+ . . . +μ n ( x ),
where a match of an M-letter search string at depth x in a word produces a score of μ M (x),
determining, by the electronic processor, scores (Tscore) for a plurality of titles within the database, wherein a title within the database comprises a plurality of scored words, and wherein a score for each title is proportional to a summation of individual scores of each word in the title;
wherein for each of N words in the title (w 1 ,w 2 , . . . w N ), Tscore is calculated as:
Tscore
=
(
∑
n
=
1
N
sword
(
w
n
)
+
β
*
log
(
Z
/
i
(
w
n
)
)
,
where i(w n ) is a number of titles having the word w n , Z is a size of titles in the database and β is determined experimentally to control a reward score given to words appearing less often across titles; or Tscore is calculated as:
Tscore
=
∑
n
=
1
N
sword
(
w
n
)
)
.
2. The method of claim 1 wherein sword=sword*f+(μ 1 (x)+μ 2 (x)+ . . . +μ n (x)), where f is a forgetting factor to gradually discount the contribution of past inputs.
3. An apparatus comprising:
a database comprising a plurality of words;
a processor receiving a textual input (search string) and determining a score for words (sword) from database based on the search string, wherein a score for each word is proportional to a number of letters within the search string and the depth of the search string within the word;
wherein: the processor scores a plurality of titles within the database, wherein titles within the database comprise a plurality of scored words, and wherein a score for each title is proportional to a summation of the individual scores of each word in the title;
wherein for each of N words in a title (w 1 ,w 2 , . . . , w N ), Tscore for the title is calculated as:
Tscore
=
(
∑
n
=
1
N
sword
(
w
n
)
+
β
*
log
(
Z
/
i
(
w
n
)
)
,
where i(w n ) is a number of titles having the word w n , Z is a size of titles in the database and β is determined experimentally to control a reward score given to words appearing less often across the titles, or Tscore for the title is calculated as:
Tscore
=
∑
n
=
1
N
sword
(
w
n
)
)
.
4. The apparatus of claim 3 wherein the processor outputs a plurality of words having a highest score.
5. The apparatus of claim 3 wherein the processor outputs a plurality of titles having a highest score.
6. The apparatus of claim 3 wherein a score (sword) for each word is based on the search string, and wherein
sword=(μ 1 ( x )+μ 2 ( x )+ . . . +μ n ( x ),
where a match of an M-letter string at depth x in a word produces a score of μ M (x).