Systems and methods for predicting usage of a web site using proximal cues

• US 6,907,459 B2
• Filed: 03/30/2001
• Issued: 06/14/2005
• Est. Priority Date: 03/30/2001
• Status: Expired due to Fees

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Chen, Kim K, Chi, Ed H.
• Primary Examiner(s): Harvey, Jack
• Assistant Examiner(s): Nguyen, Hai V.
• Application Number:
  US09/820,706
• Publication Number:
  US 20020143802A1
• Time in Patent Office: 1,537 Days
• Field of Search: 709/224, 715/501.1, 707/3, 707/5
• US Class Current: 709/224
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/958 Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

Techniques are provided for predicting the usage of a document collection given proximal cue information in the documents, a starting point and the user'"'"'s information needs. A document collection topology matrix is created indicating links between document content portions. The link entry documents are analyzed for proximal cue words based on link URL, surrounding text and title. For image links, the connected to document information may also be used. Proximal cue words are added to a matrix relating proximal cue words and links. The proximal scent matrix indicates a similarity between the user'"'"'s information need and the proximal cue word matrix. A distal scent information matrix is also calculated using distal document information and combined with the proximal scent matrix. Spreading activation is then applied to the resulting matrix using the starting location for a requested number of iterations and resulting in a predicted usage of the document collection.

Claims

1. A method for predicting usage of a collection of content portions comprising:

entering a user information need, a requested user starting location and a requested number of iterations;

determining a connection topology of the collection of content portions, the connection topology including a plurality of connections, each of the plurality of connections being contained in a respective one of the content portions included in the collection of content portions and connecting the respective one of the content portions to another of the content portions;

for each connection determined, determining proximal information cue words associated with each connection, storing at least one of the proximal information cue words based on the connection;

determining a predicted user presence in the collection of content portions by determining similarity between the stored proximal information cue words and the user information need;

determining a user path from the requested user starting location to a final destination based on the user information need based on spreading activation using the predicted user presence in the collection of content portions and the requested starting location for the requested number of iterations.

2. The method of claim 1, wherein determining the proximal information cue words comprises determining information cue words forming the connection.

3. The method of claim 1, further comprising:

if a connection is determined to be an image based connection, determining information cue words from the connected to content portion, the information cue words including at least one of title, words in the connected to content portion, and using the connected to content portion information cue words as proximal cue words for the image link.

4. The method of claim 1, wherein the connection topology information is stored in a matrix.

5. The method of claim 1, wherein the information about proximal information cue words is stored in a word/document matrix.

6. The method of claim 5, further comprising:

analyzing the proximal information cue word matrix after determining the user path for the requested number of iterations.

7. The method of claim 1, further comprising:

determining a weighted term frequency for each word in the collection of content portions; and

wherein determining a predicted user presence in the collection of content portions comprises;

determining a similarity between the stored proximal information cue words, the weighted frequency for each word and the user information need.

8. The method of claim 1, wherein the collection of content portions comprises a plurality of web pages, and

each connection in the connection topology comprises a URL embedded in a first web page that links between the first web page and a second web page.

9. A system for predicting usage of a collection of content portions comprising:

a controller;

a memory;

a topology determining circuit that determines a connection topology of a collection of content portions, the connection topology including plurality of connections, each of the plurality of connections being contained in a respective one of the content portions included in the collection of content portions and connecting the respective one of the collection portions to another of the content portions;

an input/output circuit for entering a user information need, user requested starting location in the collection of content portions and a requested number of iterations;

a connection determining circuit that identifies connections between content portions;

a proximal information cue word determining circuit that determines proximal cue words for each determined connection and stores at least one of the proximal information cue words based on the connection;

a user presence predicting circuit that determines a similarity between each stored proximal information cue word and the user information need to create an information scent array;

a spreading activation circuit that determines a user path based on user requested starting location in the collection of content portions to a final destination in the collection of content portions based on spreading activation applied the requested number of iterations to the information scent array and the user requested starting location in the collection of content portions.

10. The system of claim 9, wherein the proximal information cue word determining circuit determines information cue words that form the connection.

11. The system of claim 9, further comprising:

a circuit that determines if a connection is an image based connection; and

a circuit that determines information cue words from the content portion connected to by the connection, the information cue words including at least one of title and words in the connected to content portion, and using the connected to content portion information cue words as proximal cue words for the image link.

12. The system of claim 9, further comprising a topology information memory for storing the connection topology information in a matrix.

13. The system of claim 9, further comprising a proximal information cue word memory for storing information about proximal information cue words in a word/document matrix.

14. The system of claim 13, wherein the spreading activation circuit analyzes the proximal information cue word matrix after determining the user path for the requested number of iterations.

15. The system of claim 9, further comprising:

a weighted frequency determining circuit that determines a weighted frequency for each word in the collection of content portions; and

wherein the user presence predicting circuit determines the similarity between each stored proximal cue word, each stored proximal cue word'"'"'s respective weight term frequency and by the user information need.

16. The system of claim 9, wherein the collection of content portions comprises a plurality of web pages, and

each connection in the connection topology comprises a URL embedded in a first web page that links between the first web page and a second web page.