System and method for natural language processing and using ontological searches

• US 7,739,104 B2
• Filed: 05/30/2006
• Issued: 06/15/2010
• Est. Priority Date: 05/27/2005
• Status: Expired due to Fees

• Current Assignee: Hakia, Inc.
• Original Assignee: Hakia, Inc.
• Inventors: Raskin, Victor, Berkan, Riza C.
• Primary Examiner(s): Vo; Huyen X.
• Application Number:
  US11/443,898
• Publication Number:
  US 20060271353A1
• Time in Patent Office: 1,477 Days
• Field of Search: 707 1- 3, 707/5, 707/7, 707/4, 707/102, 704 1- 10, 704/251, 704/257, 704/270
• US Class Current: 704/9
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking

Abstract

A search engine is disclosed that utilizes natural language processing (NLP) techniques. The search engine utilizes meaning-based natural language processing using ontological semantics in analyzing the meaning of queries and the searched text. This system analyzes Web pages and queries. The NLP method produces equivalent meanings to a sequence of user initiated words, wherein relevance parsing of the original query produces a display of queries/questions as hot links to the next round of searching without additional typing by the user.

Claims

1. A method performed on a computer for data searching, comprising:

collecting data utilizing a computer network;

parsing the data by language type;

expanding the data using off-line ontological semantics;

categorizing the data and expanded data by anticipated queries;

a processor scoring the data and the expanded data based on anticipated queries;

receiving a query;

parsing the query for question type and expanding the query using ontological semantics and;

comparing the parsed and expanded user query with the anticipated queries; and

providing the matching data and the expanded data based in part on the comparison by adding a weight for each word and dividing by a total number of words, wherein the scoring is calculated such that

$\mathrm{Sx}=\frac{\underset{i=1,M}{\sum <br><br>}<br><br><br><br>R_i<br><br>W_i}{\underset{i=1,M}{\sum <br><br>}<br><br><br><br>W_i}\le <br><br>1.0$

 

wherein M is a number of significant words in the data, R is 1 for each word found in the data, and 0 otherwise, W is the weight for each word and S_x is a presence score;

storing the data and the expanded data based on anticipated queries;

receiving a query;

parsing the query for question type and expanding the query using ontological semantics;

the processor comparing the parsed and expanded user query with the anticipated queries; and

providing the matching data and the expanded data based in part on the comparison.

2. A method performed on a computer for data searching, comprising:

collecting data utilizing a computer network;

parsing the data by language type;

expanding the data using off-line ontological semantics;

categorizing the data and expanded data by anticipated queries;

scoring the data based on the parsed data;

a processor scoring the data based on a universal resource locator and title of the data, wherein the universal resource locator scoring is calculated such that;

$S_{\mathrm{dom}}={\mathrm{Ss}<br><br>\left(\frac{J}{Z}\right)}^M+c$

 

wherein J is a number of characters of query words found in a domain in the universal resource locator, Z is a total number of characters in the domain, M is a length of the query, c is a credibility factor, S_s is a sentence score and S_dom is the universal resource locator scoring;

scoring the data based on the query;

storing the data and the expanded data based on anticipated queries;

receiving a query;

parsing the query for question type and expanding the query using ontological semantics;

the processor comparing the parsed and expanded user query with the anticipated queries; and

providing the matching data and the expanded data based in part on the comparison.

3. A method performed on a computer for data searching, comprising:

collecting data utilizing a computer network;

parsing the data by language type;

expanding the data using off-line ontological semantics;

categorizing the data and expanded data by anticipated queries;

scoring the data based on the parsed data;

scoring the data based on a universal resource locator and title of the data;

a processor scoring the data based on the query, wherein the score is degraded by distance between data words and an order of the data words such that;

$\mathrm{St}=\underset{\underset{\underset{i\ne <br><br>j}{j=1,M}}{i=1,M}}{\sum <br><br>}<br><br><br><br>{\mathrm{Sd}}_{i,j}<br><br>{\mathrm{So}}_{i,j}<br><br>{\mathrm{Sw}}_{i,j}$

 

wherein M is a length of the query, S_d is a pairwise distance, S_o is a pairwise order score, S_w is a pairwise weighting score, and S_t is a pairwise total score;

storing the data and the expanded data based on anticipated queries;

receiving a query;

parsing the query for question type and expanding the query using ontological semantics;

the processor comparing the parsed and expanded user query with the anticipated queries; and

providing the matching data and the expanded data based in part on the comparison.