Selene Communication Technologies LLC v. Google Inc.
Filing
1
COMPLAINT filed with Jury Demand against Google Inc. - Magistrate Consent Notice to Pltf. ( Filing fee $ 400, receipt number 0311-1503856.) - filed by Selene Communication Technologies LLC. (Attachments: # 1 Exhibit A, part 1, # 2 Exhibit A, part 2, # 3 Exhibit A, part 3, # 4 Exhibit A, part 4, # 5 Civil Cover Sheet)(cla, )
US 6,363,377 Bl
13
14
in h(t)'s hierarchy tree. The variable f is in the interval (0,1)
of queries generated by selecting each query item in the
and H is the height of h(t)'s hierarchy tree. In practical
initial search query for substitution to form a respectively
terms, it is desirable to generate a small but non-trivial
different one of the related queries.
number of related queries, say 5-10. Depending on the
5. A method for generating search queries according to
length and complexity of the original query and the value s, 5 claim 4 wherein the initial search query is modified by
f-family can produce quite a substantial number of queries.
restricting and broadening the search query by traversing
In such a case, only a reasonable number are forwarded to
one edge up and one edge down on the hierarchy tree for the
the search engine. The queries are chosen by ordering the
selected query item.
items to be modified from the outermost Boolean operator,
6. A method for generating search queries according to
to the innermost until the goal number is reached. The choice
10 claim 5 wherein the number of edges of the hierarchy tree to
of parameter f is based on experience of the particular user,
be traversed is increased to generate additional queries.
as well as the query specification. On the other extreme, if
7. A method for generating search queries according to
the query is very basic, such as a single word, all the possible
claim number 1 wherein the step of receiving an initial
related queries may be generated.
search query includes retreiving keywords from a prior
Although the invention is illustrated in terms of broad15 search and adding them to the search query.
ening and restricting one item at a time and determining the
8. A method for organizing a set of records into clusters
replacement by using the same variable fraction f for each
comprising the steps of:
item, it is contemplated that any number of items can be
a) receiving the set of records;
simultaneously changed and that the replacement for each
b) analyzing at least one text field from each record of the
can be determined independently by traversing any number
20
set of records to determine any patterns;
of edges in the item's hierarchy tree.
c) partitioning each record by the text field analyzed in
The results returned by the search engine contain the
step (b) into clusters based on a shared pattern, wherein
complete answer to the user's query, together with the
the step of partitioning a set of records consists of the
number of matches that each of the related queries would
steps of:
elicit. The relative number of matches that the related
25
cl) hashing all the single words, all the pairs of
queries produce is useful in providing the user with a
consecutive words and all long sequences of words
measure of the relative restrictiveness of the different items
for each record in the set of records; and
in the query. Based on this extra information, the user may
c2) partitioning a set of records by the text field hashed
be able to reformulate a query in a more intelligent way than
in step (cl) into clusters
he could otherwise.
30
An example implementation of the user interface is shown
d) analyzing the partitions to identify the clusters for
in FIG. 11. In addition to the query, the user can specify the
display; and
number of matches she would be most interested in seeing.
e) displaying the clusters to the user.
Out of all the related queries generated by the query tuner,
9. A method for organizing a set of records into clusters,
only those that yield close to the desired number of results
35 comprising the steps of:
are displayed to the user.
a) receiving the set of records;
Although illustrated and described herein with reference
b) analyzing at least one text field from each record of the
to certain specific embodiments, the present invention is
set of records to determine any patterns;
nevertheless not intended to be limited to the details shown.
c) partitioning each record by the text field analyzed in
Rather, various modifications may be made in the details
40
step (b) into clusters based on a shared pattern;
within the scope and range of equivalents of the claims and
d) analyzing the partitions to identify the clusters for
departing from the scope of the invention.
display; and
The invention claimed is:
e) displaying the clusters to the user;
1. A method for generating search queries to be sent to a
search engine for searching a information management
wherein the step of analyzing the partitions to identify the
45
system, comprising the steps of:
clusters for display includes analyzing a best cluster for
a) receiving an initial search query;
display by weighing both the number of documents that
contain the shared pattern and the length of the shared
b) converting the initial search query to general boolean
pattern.
language;
10. A method for organizing a set of records into clusters
c) identifying a level in a respective hierarchy tree for 50
according to claim 9, wherein each record is a document
each search query item in the initial search query;
having a uniform resource locator (URL), a title, a document
d) formulating additional related search queries by subexcerpt and date information, and the fields to be analyzed
stituting items from the respective hierarchy tree for
are selected from a group consisting of: the title, the URL,
selected items in the query, the substituted item having
the document excerpt and the date information.
a level in the hierarchy tree that is greater than or less 55
11. A system for generating search queries to be sent to a
than the level of the query item in the initial query; and
search engine for searching a information management
e) forwarding the initial search query and the additional
system, comprising:
search queries in parallel to the search engine.
a) means for receiving an initial search query;
2. A method for generating search queries according to
b) means for converting the initial search query to general
claim 1 wherein the information management system to be 60
Boolean language;
searched is a global information network.
c) means for placing each search query item in the initial
3. A method for generating search queries according to
search query into a hierarchy tree;
claim 1 wherein the search query items are selected from a
group consisting of: a meta-datum, a keyword and a Boolean
d) means for formulating additional related search queries
65
operator.
by replacing at least one search query item with an item
4. A method for generating search queries according to
either above or below the search query item in its
claim 1 wherein the additional related queries consist of a set
hierarchy tree; and
US 6,363,377 Bl
15
16
e) means for forwarding the initial search query and the
c) placing each search query item in the initial search
additional search queries in parallel to the search
query into a hierarchy tree;
engine.
d) means for formulating additional related search queries
12. A system for organizing a search engine's results
by replacing at least one search query item with an item
including a set of documents each document including a 5
either above or below the search query item in its
plurality of fields, comprising:
hierarchy tree; and
a) means for receiving the set of documents;
e) forwarding the initial search query and the additional
b) means for analyzing several fields from each document
search queries in parallel to the search engine.
to determine patterns
14. A carrier including a computer program which, when
10
c) means for partitioning each document by the fields
executed by a processor, causes the processor to organize a
analyzed in step (b) into clusters based on a shared
set of documents into clusters, by causing the computer to
pattern;
perform the steps of:
d) means for analyzing the partitions to identify clusters
a) receiving the set of documents;
for display;
15
b) analyzing at least one field from each document to
e) means for displaying the clusters of documents to the
determine a pattern;
user; and
c) partitioning each document by the fields analyzed in
f) means for user selection of the clusters for viewing;
step (b) into clusters based on a shared pattern;
wherein the means for analyzing the partitions to identify
20
d) analyzing the partitions to identify the clusters for
the clusters for display includes means for analyzing a
display;
best cluster for display by weighing both a number of
e) displaying the clusters to the user; and
documents that contain the shared pattern and a length
of the shared pattern.
f) allowing the user to select one or more of the clusters
13. A carrier including a computer program which, when
for viewing;
25
executed by a processor, causes the processor to generate
wherein the step of analyzing the partitions to identify the
search queries to be sent to a search engine for searching a
clusters for display includes the step of analyzing a best
information management system, the computer program
cluster for display by weighing both a number of
causing the computer to perform the steps of:
documents that contain the shared pattern and a length
a) receiving an initial search query;
of the shared pattern.
30
b) converting the initial search query to general Boolean
language;
* * * * *
Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.
Why Is My Information Online?