CN102915381B - Visual network retrieval based on multi-dimensional semantic presents system and presents control method - Google Patents

Abstract

The invention relates to a multi-dimensional semantics-based visualized network retrieval presentation system and a presentation control method, belonging to the technical field of network retrieval. The system includes a query server, a semantic matching and reasoning module, an index database, a semantic index result set, a fractal rule unit, and a multidimensional result presentation unit. The acquired semantic matching and reasoning result sets establish and save the semantic ontology index; the multi-dimensional rule unit clusters the index result set into multi-dimensional and multi-level data results according to the semantic distance of the keywords in the semantic index result set, so as to benefit users in the In the multi-dimensional candidate retrieval result presentation form, the retrieval target result can be quickly located, so as to effectively distinguish different semantics of the same text information and improve retrieval efficiency.

Description

Visual network retrieval presentation system and presentation control method based on multi-dimensional semantics

technical field

The present invention relates to the technical field of network retrieval, in particular to the technical field of network retrieval presentation, in particular to a multi-dimensional semantics-based visual network retrieval presentation system and presentation control method.

Background technique

With the rapid development of retrieval technology, various mature search engines such as Google (Google), Yahoo (Yahoo), and Baidu (Baidu) have emerged at home and abroad. These search engines are mainly based on text-based information retrieval technology, providing users with a complete and highly relevant information retrieval engine. Although the existing text search technology can search for files containing the user's text query information, the presentation form is mainly to sort according to the relevance of the search results, and return the results in the form of a link result set to the user. The biggest shortcoming of this retrieval technology is that the ambiguity of the retrieval keywords leads to a wide variety of semantic relationships in the search result sets. For example, when the search keyword submitted by the user to the search engine is "apple", the search engine cannot correctly judge " refers to the fruit "apple", or to the "Apple" company founded by Steve Jobs, or to the French film "The Apple". In the absence of context, the search engine cannot accurately determine which candidate content is most relevant to the searched "apple" keyword, so the search results often cannot meet the needs of users.

Contents of the invention

The purpose of the present invention is to overcome the shortcomings in the above-mentioned prior art, and provide a method of matching the user's text query information and file index information to present the retrieval results to the user in a layered and dimensional manner according to the logic of semantics, so as to facilitate In the presentation form of candidate retrieval results based on multiple dimensions, users can quickly locate the target results of retrieval, thereby effectively distinguishing different semantics of the same text and improving retrieval efficiency. The system structure is simple, the cost is low, and the method is easy to apply. A multi-dimensional semantic-based visualized network retrieval presentation system and presentation control method with a wide range.

In order to achieve the above-mentioned purpose, the multi-dimensional semantics-based visual network retrieval and presentation system of the present invention has the following components:

The system includes a query server, a semantic matching and reasoning module, an index database, a semantic index result set, a fractal rule unit and a multidimensional result presentation unit. Wherein, the query server is used to provide user search keyword input interface; the semantic matching and reasoning module is connected to the query server, and the keyword semantics is matched and reasoned according to the knowledge set in the related field; the index database is respectively connected to the query The server and the semantic matching and reasoning module are used to provide corresponding search results for the search keywords; the semantic index result set is connected to the index database to store the index result sets corresponding to the search keywords; the fractal rule unit is respectively connected to the index database The semantic index result set and the semantic matching and reasoning module described above cluster the index result set into multi-level data results in multiple dimensions according to the semantic distance of keywords in the semantic index result set; the multi-dimensional result presentation unit connects the The fractal dimension rule unit is used to present the multi-level data results on the multi-dimensional to the user.

In the multi-dimensional semantic-based visual network retrieval and presentation system, the semantic matching and reasoning module includes a standard ontology knowledge base, a semantic matching unit and a semantic reasoning unit. Wherein, the standard ontology knowledge base stores ontology knowledge sets in corresponding domains; the semantic matching unit connects to the standard ontology knowledge base, obtains semantic matching rules of keywords according to the ontology knowledge sets, and performs semantic matching; semantic reasoning The unit is connected to the standard ontology knowledge base, obtains semantic reasoning rules of keywords according to the ontology knowledge set, and performs semantic reasoning.

The present invention also provides a method for realizing visual network retrieval presentation control based on multi-dimensional semantics by using the system, the method includes the following steps:

(1) The query server receives the query keyword, and judges whether the keyword is a complex sentence, if so, proceeds to step (2), if not, proceeds to step (3);

(2) The query server performs word segmentation and filtering processing, and outputs a keyword string containing a separator to the index database, and then enters step (3);

(3) The semantic matching and reasoning module performs semantic matching and reasoning on the keywords, and sends the semantic reasoning result set to the index database;

(4) The index database establishes and saves the semantic ontology index according to the acquired semantic matching and reasoning result sets, and sends the index result sets of the semantic matching and reasoning result sets to the fractal rule unit;

(5) The multi-dimensional rule unit clusters the index result set into a data form with multiple dimensions according to the semantic distance of the keywords in the semantic index result set, and the data form clusters multiple levels in each dimension data result;

(6) The multi-dimensional result presentation unit presents the multi-level data results in multiple dimensions to the user.

In the method for realizing visual network retrieval presentation control based on multi-dimensional semantics, the query server performs word segmentation and filtering, and outputs keyword strings containing separators to the index database, specifically: the query server according to Word segmentation and filtering are performed for different language types of keywords, and keyword strings containing separators are output.

In the method for realizing visual network retrieval and presentation control based on multi-dimensional semantics, the semantic matching and reasoning module includes a standard ontology knowledge base, a semantic matching unit and a semantic reasoning unit, and the standard ontology knowledge base stores ontologies in corresponding fields knowledge collection; both the semantic matching unit and the semantic reasoning unit are connected to the standard ontology knowledge base, and the step (3) specifically includes the following steps:

(31) After the semantic matching and reasoning module receives the query keyword, the semantic matching unit performs semantic matching processing on the keyword according to the standard ontology knowledge base, and submits the semantic matching result set to the semantic reasoning unit;

(32) The semantic reasoning unit performs semantic reasoning processing on the semantic matching result set to obtain a semantic reasoning result set, and sends the semantic reasoning result set to the index database.

In the method for realizing visual network retrieval and presentation control based on multi-dimensional semantics, the semantic matching process specifically includes: performing semantic similarity calculation with query keywords according to a specific keyword set in this field to realize semantic matching.

In the method for realizing visual network retrieval and presentation control based on multi-dimensional semantics, the semantic inference processing is specifically: according to the ontology knowledge in a specific field, the inference rules of the field are obtained, and the rules are used to reason the semantic matching results to obtain Semantic inference result set.

In the method for realizing visual network retrieval presentation control based on multi-dimensional semantics, the step (5) specifically includes the following steps:

(51) The multi-dimensional rule unit calculates the semantic distance between keywords in the semantic index result set;

(52) The multi-dimensional rule unit clusters the index result set into data results of multiple dimensions and multiple levels according to the semantic distance.

In the method for implementing visual network retrieval presentation control based on multi-dimensional semantics, the step (51) specifically includes the following steps:

(51-1) The multi-dimensional rule unit searches for the nearest common ancestor node of multiple keywords in the semantic index result set;

(51-2) The multi-dimensional rule unit calculates the distance between each keyword and the nearest common ancestor node;

(51-3) The multi-dimensional rule unit uses the sum of the distances between each keyword and the nearest common ancestor node as the semantic distance between the keywords in the semantic index result set.

In the method for realizing visual network retrieval presentation control based on multi-dimensional semantics, the step (52) specifically includes the following steps:

(52-1) The multi-dimensional rule unit analyzes the relationship between the search keywords and the semantic distance according to the semantic distance between the keywords;

The multidimensional rule unit described in (52-2) expands a certain dimension in the multidimensional data set, and determines the dimension and level to which the retrieval result belongs;

(52-3) Collect each retrieval result into a data result with multiple dimensions and multiple levels.

The multi-dimensional semantics-based visual network retrieval presentation system and presentation control method of the invention are adopted. The system includes a query server, a semantic matching and reasoning module, an index database, a semantic index result set, a fractal rule unit, and a multi-dimensional result presentation unit. The semantic matching and reasoning module can be used to carry out semantic matching and reasoning on the keywords, and the index database can establish and save the semantic ontology index according to the obtained semantic matching and reasoning result sets; the multidimensional rule unit can use the semantic distance of the keywords in the semantic index result set , to cluster the index result set into multi-dimensional and multi-level data results; finally, the multi-dimensional result presentation unit presents it to the user, so that the user can quickly locate the target result of the retrieval in the presentation form of candidate retrieval results based on multi-dimensional A visual network retrieval presentation system and presentation control method based on multi-dimensional semantics that can effectively distinguish different semantics of the same text information, improve retrieval efficiency, and have simple system structure, low cost, simple method application, and wide application range.

Description of drawings

FIG. 1 is a schematic structural diagram of the multi-dimensional semantics-based visual network retrieval and presentation system of the present invention.

FIG. 2 is a flow chart of a specific embodiment of the method for implementing visual network retrieval presentation control based on multi-dimensional semantics in the present invention.

Fig. 3 is a flow chart of the retrieval and presentation module of the multi-dimensional semantic space in the embodiment of the present invention.

Fig. 4 is a sequence diagram of an embodiment of a visual retrieval and presentation system based on a multi-dimensional semantic space in the present invention.

Detailed ways

In order to understand the technical content of the present invention more clearly, the following examples are given in detail.

Please refer to FIG. 1 , which is a schematic structural diagram of the multi-dimensional semantics-based visual network retrieval and presentation system of the present invention.

In one embodiment, the system includes a query server, a semantic matching and reasoning module, an index database, a semantic index result set, a fractal rule unit and a multidimensional result presentation unit. Wherein, the query server is used to provide user search keyword input interface; the semantic matching and reasoning module is connected to the query server, and the keyword semantics is matched and reasoned according to the knowledge set in the related field; the index database is respectively connected to the query The server and the semantic matching and reasoning module are used to provide corresponding search results for the search keywords; the semantic index result set is connected to the index database to store the index result sets corresponding to the search keywords; the fractal rule unit is respectively connected to the index database The semantic index result set and the semantic matching and reasoning module described above cluster the index result set into multi-level data results in multiple dimensions according to the semantic distance of keywords in the semantic index result set; the multi-dimensional result presentation unit connects the The fractal dimension rule unit is used to present the multi-level data results on the multi-dimensional to the user.

The method for realizing visual network retrieval presentation control based on multi-dimensional semantics using the system described in this embodiment includes the following steps:

(1) The query server receives the query keyword, and judges whether the keyword is a complex sentence, if so, proceeds to step (2), if not, proceeds to step (3);

(2) The query server performs word segmentation and filtering processing, and outputs a keyword string containing a separator to the index database, and then enters step (3);

(3) The semantic matching and reasoning module performs semantic matching and reasoning on the keywords, and sends the semantic reasoning result set to the index database;

(4) The index database establishes and saves the semantic ontology index according to the acquired semantic matching and reasoning result sets, and sends the index result sets of the semantic matching and reasoning result sets to the fractal rule unit;

(5) The multi-dimensional rule unit clusters the index result set into a data form with multiple dimensions according to the semantic distance of the keywords in the semantic index result set, and the data form clusters multiple levels in each dimension data result;

(6) The multi-dimensional result presentation unit presents the multi-level data results in multiple dimensions to the user.

Wherein, the query server described in step (2) performs word segmentation and filtering processing, and outputs keyword strings containing separators to the index database, specifically: the query server separates the keyword strings according to different language types of keywords Perform word segmentation and filtering, and output keyword strings containing separators.

In a preferred embodiment, the semantic matching and reasoning module includes a standard ontology knowledge base, a semantic matching unit and a semantic reasoning unit. Wherein, the standard ontology knowledge base stores ontology knowledge sets in corresponding domains; the semantic matching unit connects to the standard ontology knowledge base, obtains semantic matching rules of keywords according to the ontology knowledge sets, and performs semantic matching; semantic reasoning The unit is connected to the standard ontology knowledge base, obtains semantic reasoning rules of keywords according to the ontology knowledge set, and performs semantic reasoning.

In the method for realizing visual network retrieval presentation control based on multi-dimensional semantics using the system described in this preferred implementation manner, the step (3) specifically includes the following steps:

(31) After the semantic matching and reasoning module receives the query keyword, the semantic matching unit performs semantic matching processing on the keyword according to the standard ontology knowledge base, and submits the semantic matching result set to the The semantic reasoning unit, the semantic matching processing, specifically: according to the specific keyword set in this field, carry out semantic similarity calculation with the query keywords to realize semantic matching;

(32) The semantic reasoning unit performs semantic reasoning processing on the semantic matching result set to obtain a semantic reasoning result set, and sends the semantic reasoning result set to the index database. Wherein, the semantic reasoning process specifically includes: according to the ontology knowledge in a specific field, obtain the reasoning rules in the field, use the rules to reason the semantic matching results, and obtain the semantic reasoning result set.

In a further preferred embodiment, the step (5) specifically includes the following steps:

(51) The multi-dimensional rule unit calculates the semantic distance between keywords in the semantic index result set;

(52) The multi-dimensional rule unit clusters the index result set into data results of multiple dimensions and multiple levels according to the semantic distance.

In a more preferred embodiment, the step (51) specifically includes the following steps:

(51-1) The multi-dimensional rule unit searches for the nearest common ancestor node of multiple keywords in the semantic index result set;

(51-2) The multi-dimensional rule unit calculates the distance between each keyword and the nearest common ancestor node;

(51-3) The multi-dimensional rule unit uses the sum of the distances between each keyword and the nearest common ancestor node as the semantic distance between the keywords in the semantic index result set.

And the step (52) specifically includes the following steps:

(52-1) The multi-dimensional rule unit analyzes the relationship between the search keywords and the semantic distance according to the semantic distance between the keywords;

The multidimensional rule unit described in (52-2) expands a certain dimension in the multidimensional data set, and determines the dimension and level to which the retrieval result belongs;

(52-3) Collect each retrieval result into a data result with multiple dimensions and multiple levels.

In practical applications, in the multi-dimensional semantic space-based visual retrieval presentation system provided by the present invention, the semantic similarity calculation of the search keywords is extended, and the semantic distance between two keywords can be understood as two nodes, two The semantic distance between nodes refers to the sum of the paths from the nearest common ancestor nodes of two nodes to these two nodes. To calculate the minimum distance between two nodes is to find the nearest common ancestor node, then calculate the distance between the two nodes, and finally add the two distances to get the desired result.

In the semantic clustering algorithm, a multidimensional array is used to calculate the semantic distance of the retrieval keywords. After analyzing the semantic relationship between the two keywords, a certain dimension in the multidimensional data set can be expanded to determine the retrieval results. The data results on which levels of a dimension.

Fig. 1 schematically shows the principle diagram of the visual retrieval and presentation system based on multi-dimensional semantic space realized by the present invention, including query server, standard ontology knowledge base, semantic matching unit, semantic reasoning unit, index database, semantic index result set, fractal rules and multi-dimensional Results presentation unit. The query server is an interface that provides users with search keywords; the standard ontology knowledge base stores ontology knowledge collections in this field, and provides semantic matching and reasoning rules for the semantic matching unit and semantic reasoning unit; the index database provides corresponding search results for search keywords The semantic index result set stores the index result set corresponding to the search keyword; the fractal dimension rule unit clusters the index result set into a data form with multiple dimensions according to the semantic distance of the keywords in the semantic index result. The results of clustering multiple levels of data.

Fig. 2 shows a flow chart of an embodiment of the method of the present invention, which mainly includes the following steps.

Step 201, receiving query keywords, and judging whether the input keywords are complex sentences, if so, proceed to step 202; otherwise, proceed to step 203, and send to the index database.

In step 202, different word segmentation and filtering processes are performed according to different language types of the query keywords, and a series of delimited character strings such as Chinese words, English words, and number strings are output.

Step 203, according to the content of the index database, index to obtain a set of search results corresponding to the query keywords.

Step 204, Semantic Reasoning: According to the ontology knowledge in a specific field, get the reasoning rules of the field, use the rules to reason the description results, and get the reasoning result set; Semantic matching: according to the reasoning result set and the specific keywords in this field The collection performs semantic similarity calculation and semantic matching.

Step 205, the dimensionality rule unit clusters the index result set into a data form with multiple dimensions according to the semantic distance of keywords in the semantic index result, and clusters data results of multiple levels in multiple dimensions.

Step 206, the result presentation module presents the search results in a multi-dimensional data form.

Fig. 3 is a flow chart of the retrieval and presentation module of the multi-dimensional semantic space in the embodiment of the present invention, which mainly includes the following steps.

Step 301 , according to the established index content of the index database, a search result set corresponding to the query key word is obtained.

Step 302, the semantic matching module performs semantic similarity calculation and semantic matching according to the inference result set and the field-specific keyword set.

Step 303, the semantic reasoning module, sets the reasoning rules in this field, uses the rules to reason the description results, and obtains the reasoning result set.

Step 304, calculate the semantic distance of two keywords, it can be assumed that the two keywords to be sought can be expressed as two nodes (and), and their common ancestor nodes have the following properties: the common ancestor node itself and its There must be "and" nodes in the left and right subtrees. Then visit itself, the left subtree and the right subtree sequentially from the head node, if there is an "or" node in it, the count symbol will be increased by 1. When it is found that the mark is 2 after the visit, it means that the current node contains both "and" nodes, that is, the current node is the nearest common node of the target, and the semantic distance between the two keywords is the "and" node Respectively to the sum of the nearest common nodes.

Step 305, classifying and clustering the search results, using a multidimensional array to calculate the semantic distance of the retrieval keywords, and after analyzing the relationship between the retrieval keywords and the semantic distance, a certain dimension in the multidimensional data set can be expanded to determine the retrieval results Which dimensions and which levels are the data results.

In step 306, the retrieval results are presented in subdimensions, and according to the semantic distance of keywords in the semantic index results, the index result sets are clustered into a data form with multiple dimensions, and data results of multiple levels are clustered in multiple dimensions.

FIG. 4 is a sequence diagram of an embodiment of a visual retrieval and presentation system based on a multi-dimensional semantic space in the present invention, which mainly includes the following steps.

Step 401, the query server sends a query request to the semantic extension module;

In step 402 and step 403, the semantic extension module expands the search keyword according to the standard ontology knowledge base, obtains the extended query request keyword, and sends it to the index database module;

Step 404, indexing the database module, and obtaining a search result set corresponding to the query keyword segment.

Step 405, the index database module sends the search result set to the fractal dimension presentation module.

Step 406, the fractal dimension presentation module calculates the semantic distance, and the method of calculating the semantic similarity of the retrieval keyword is to add up the paths from the nearest common ancestor node of the two nodes to the two nodes, so the calculation of the two The key to the minimum distance of a node is to find the nearest common ancestor node, and then calculate the distance between the two nodes, and add the distance to get the result.

Step 407, presenting module classification and clustering search results in fractal dimensions, using multidimensional arrays to calculate the semantic distance of retrieval keywords, and after analyzing the relationship between retrieval keywords and semantic distances, a certain dimension in the multidimensional data set can be expanded, Then it is determined which dimensions and which levels of data results the retrieval results are.

In step 408, the fractal-dimensional presentation module organizes the search results into semantic network relationships and displays them in a multi-dimensional data form.

The multi-dimensional semantics-based visual network retrieval presentation system and presentation control method of the invention are adopted. The system includes a query server, a semantic matching and reasoning module, an index database, a semantic index result set, a fractal rule unit, and a multi-dimensional result presentation unit. The semantic matching and reasoning module can be used to carry out semantic matching and reasoning on the keywords, and the index database can establish and save the semantic ontology index according to the obtained semantic matching and reasoning result sets; the multidimensional rule unit can use the semantic distance of the keywords in the semantic index result set , to cluster the index result set into multi-dimensional and multi-level data results; finally, the multi-dimensional result presentation unit presents it to the user, so that the user can quickly locate the target result of the retrieval in the presentation form of candidate retrieval results based on multi-dimensional A visual network retrieval presentation system and presentation control method based on multi-dimensional semantics that can effectively distinguish different semantics of the same text information, improve retrieval efficiency, and have simple system structure, low cost, simple method application, and wide application range.

In this specification, the invention has been described with reference to specific embodiments thereof. However, it is obvious that various modifications and changes can be made without departing from the spirit and scope of the invention. Accordingly, the specification and drawings are to be regarded as illustrative rather than restrictive.

Claims (8)

1. the visual network retrieval based on multi-dimensional semantic presents a system, and it is characterized in that, described system comprises:

Querying server, in order to provide user search keyword input interface;

Semantic matches and reasoning module, the querying server described in connection, mates and reasoning keyword semanteme according to the knowledge collection in association area;

Index data base, connects described querying server and semantic matches and reasoning module respectively, with thinking that search keyword provides corresponding Search Results;

Semantic indexing result set, the index data base described in connection, in order to preserve the indexed results collection corresponding with search keyword;

Point dimension rules unit, connects described semantic indexing result set and semantic matches and reasoning module respectively, according to the semantic distance of keyword in semantic indexing result set, indexed results clustering is become the many levels data result in multiple dimension;

Multidimensional consequence display unit, point dimension rules unit described in connection, in order to present the many levels data result in described multiple dimensions to user;

Described semantic matches and reasoning module comprise:

Standard body knowledge base, stores the ontology knowledge set in corresponding field;

Semantic matches unit, the standard body knowledge base described in connection, obtains the semantic matches rule of keyword, and carries out semantic matches according to described ontology knowledge set;

Semantic reasoning unit, the standard body knowledge base described in connection, obtains the semantic reasoning rule of keyword, and carries out semantic reasoning according to described ontology knowledge set.

2. utilize the system described in claim 1 to realize visual network based on multi-dimensional semantic and retrieve the method presenting control, it is characterized in that, described method comprises the following steps:

(1) querying server described in receives searching keyword, and judges whether keyword is complex sentence, if so, then enters step (2), if not, then enters step (3);

(2) querying server described in carries out participle filtration treatment, and contains the keyword character string of separatrix to described index data base output packet, then enters step (3);

(3) semantic matches described in and reasoning module carry out semantic matches and reasoning to described keyword, and semantic reasoning result set is sent to described index data base;

(4) index data base described in is set up according to the semantic matches obtained and the reasoning results collection and preserves Ontology index, and by the indexed results collection of semantic matches and the reasoning results collection, indexed results collection is sent to described point and ties up rules unit;

(5) multidimensional rules unit is according to the semantic distance of keyword in described semantic indexing result set, indexed results clustering is become to have the data mode of multiple dimension, the data result of described data mode cluster many levels in each dimension;

(6) multidimensional consequence display unit presents the many levels data result in multiple dimension to user;

Described semantic matches and reasoning module comprise standard body knowledge base, semantic matches unit and semantic reasoning unit, and described standard body knowledge base stores the ontology knowledge set in corresponding field; Described semantic matches unit is all connected described standard body knowledge base with described semantic reasoning unit, and described step (3) specifically comprises the following steps:

(31) after the semantic matches described in and reasoning module receive searching keyword, described semantic matches unit carries out semantic matches process according to described standard body knowledge base to keyword, and semantic matches result set is submitted to described semantic reasoning unit;

(32) the semantic reasoning unit described in carries out semantic reasoning process to described semantic matches result set, obtains semantic reasoning result set, and described semantic reasoning result set is sent to described index data base.

3. according to claim 2ly realize visual network retrieval based on multi-dimensional semantic and present the method for control, it is characterized in that, described querying server carries out participle filtration treatment, and contains the keyword character string of separatrix to described index data base output packet, is specially:

Described querying server carries out participle and filtration treatment respectively according to the different language type of keyword, and output packet is containing the keyword character string of separatrix.

4. according to claim 2ly realize visual network retrieval based on multi-dimensional semantic and present the method for control, it is characterized in that, described semantic matches process, is specially:

According to the specific keyword set in this area, itself and searching keyword are carried out Semantic Similarity Measurement, realizes semantic matches.

5. according to claim 2ly realize visual network retrieval based on multi-dimensional semantic and present the method for control, it is characterized in that, described semantic reasoning process, is specially:

According to the ontology knowledge in specific area, draw the inference rule in this field, utilize rule to carry out reasoning to semantic matches result, obtain semantic reasoning result set.

6. realizing visual network retrieval based on multi-dimensional semantic and present the method for control according to any one of claim 2 to 5, it is characterized in that, described step (5) specifically comprises the following steps:

(51) semantic distance between the keyword in the semantic indexing result set described in the multidimensional rules unit described in calculates;

(52) indexed results clustering is become the data result of multiple dimension many levels by the multidimensional rules unit described according to described semantic distance.

7. according to claim 6ly realize visual network retrieval based on multi-dimensional semantic and present the method for control, it is characterized in that, described step (51) specifically comprises the following steps:

(51-1) the multidimensional rules unit described in searches the nearest public ancestor node of the multiple keywords in described semantic indexing result set;

(51-2) the multidimensional rules unit described in calculates the distance between each keyword and described nearest public ancestor node;

(51-3) the multidimensional rules unit described in using the distance sum between each keyword and described nearest public ancestor node as the semantic distance between the keyword in semantic indexing result set.

8. according to claim 7ly realize visual network retrieval based on multi-dimensional semantic and present the method for control, it is characterized in that, described step (52) specifically comprises the following steps:

(52-1) the multidimensional rules unit described in, according to the semantic distance between described keyword, analyzes the relation between search key and semantic distance;

(52-2) the multidimensional rules unit described in launches a certain dimension that multidimensional data is concentrated, the dimension belonging to deterministic retrieval result and level;

(52-3) each retrieval set is synthesized the data result with multiple dimension many levels.