CN105787800A

CN105787800A - Intelligent social platform potential contact retrieval device, system and method

Info

Publication number: CN105787800A
Application number: CN201610192265.5A
Authority: CN
Inventors: 陈树群
Original assignee: Chengdu Gaohesheng Technology Co Ltd
Current assignee: Guangdong Guangdong Marketing Group Co., Ltd
Priority date: 2016-03-30
Filing date: 2016-03-30
Publication date: 2016-07-20
Anticipated expiration: 2036-03-30
Also published as: CN105787800B

Abstract

The invention discloses an intelligent social platform potential contact retrieval system. The system comprises a client and a service end, wherein the client comprises a user background information generation module, a first sending module and a first reception module, the service end comprises a second reception module, a contact retrieval module and a second sending module. The invention further discloses an intelligent social platform potential contact retrieval method. The method comprises steps that the user background information and contact retrieval key words are sent by the client to the service end; the user background information is supplemented by the service end to a user background information summary table; the user background information of second-level contacts or second-level and third-level contacts of a user are retrieved by the service end, and a contact retrieval result is generated; the contact retrieval result is sent by the service end to the client; the contact retrieval result is received by the client. The system and the method during application have functions of artificial intelligence and self learning, through the method, potential contacts which the user is willing to know about can be rapidly searched from the second-level and third-level contacts, and the system is convenient for popularization and application.

Description

A kind of intelligentized social platform potential human connection retrieval device, system and method

Technical field

The present invention relates to field of social network, specifically a kind of intelligentized social platform potential human connection retrieval device, system and method.

Background technology

Social tool is a part indispensable in modern network life, can easily and timely contact other people by social tool people, thus social tool is widely used and quickly develops.Social tool actually used in usually can relate to the search of social object, existing social tool is in use, be not easy to the potential human connection to getting close to mutually with oneself and scan for, and this has had a strong impact on the popularization and application of social networks.

Summary of the invention

It is an object of the invention to overcome the deficiencies in the prior art, provide a kind of intelligentized social platform potential human connection retrieval device, system, it is easy to user during its application in two degree of human connections, three degree of human connections, is quickly found out the potential human connection oneself having a mind to recognize, be conducive to the popularization and application of social networks.The invention also discloses the human connection search method based on the potential human connection searching system of above-mentioned intelligentized social platform.

This invention address that the problems referred to above are achieved through the following technical solutions: the potential human connection retrieval device of a kind of intelligentized social platform, be applied to and the client of service end communication connection, described device includes:

User context information generating module, during for registering according to user, the information of typing generates user context information；

First sending module, for sending user context information, human connection search key to service end；

First receiver module, for receiving the human connection retrieval result that service end generates according to human connection search key.The client of the present invention be further equipped with routine Data Enter and display module, in order to the typing of user's registration information, human connection search key input, and human connection retrieval result display.When the present invention applies, user context information includes the information such as graduation universities and colleges, specialty, department, company, industry.

The potential human connection retrieval device of a kind of intelligentized social platform, is applied to the service end being connected with client communication, and described device includes:

Second receiver module, for receiving user context information, the human connection search key that client sends；Userspersonal information's summary sheet module, generates user context information summary sheet for the user context information received according to the second receiver module；

Human connection retrieval module, for according to human connection search key, two degree of human connections of this user or the user context information of two degree and three degree human connections being retrieved, generates the human connection retrieval result meeting described human connection search key；Wherein, two degree of human connections are the good friend of user good friend, and three degree of human connections are the good friend of the good friend of user good friend；

Second sending module, sends to client for human connection is retrieved result.

The Internet noise phenomena of current social platform is very serious, user cannot quickly recognize which two degree of human connection or/and three degree of human connections are oneself real needs, thus cannot be carried out selectively social, such as, assume that certain user has 200 good friends in social platform, 200 described good friends on average have 200 good friends respectively, this user has 40000 two degree of human connections altogether, some social platform can actively recommend some two degree of human connections to user at present, but cannot all push (because the quantity of two degree of human connections is too huge), and cannot recommend accurately very much according to the needs of user self, the present invention can solve this difficult problem.Such as user directly inputs " Tsing-Hua University " or " earthwork engineering " or " engineering department ", " Poly real estate " in client, during information such as " real estate industries ", service end is by analyzing, can find that some friend of this user just has the friend graduating from Tsing-Hua University, or the friend of study civil engineering, or the friend just gone to work in the engineering department of some company, or just the friend of Poly real estate working, or it is engaged in the friend of real estate industry.After the user's described information with reference to social platform offer, user can contact with these friends, and request the other side recommends to recognize the two degree of human connections just now retrieved.So, the utilization rate of two degree of human connections can be improved when the present invention applies.

Further, the potential human connection retrieval device of a kind of intelligentized social platform, it is applied to the service end being connected with client communication, described device also includes:

User context information classification module, according to the classification of information in user context information, classifies to user context information；

Intellectualized retrieval dictionary module, for storing mutual relation between sorted user context information and each described sorted user context information of registered users, dynamically improves, according to the sorted user context information of user and the mutual relation thereof of new registration, the content supplementing described intellectualized retrieval dictionary module and the human connection search key according to user's input generates other key words that described key word is relevant.The present invention is when specifically arranging, and user context information classification module is not limited to be located at service end, and it can also be located at client；

Described human connection retrieval module is additionally operable to other the relevant key words generated according to intellectualized retrieval dictionary module, and two degree of human connections of this user or the user context information of two degree and three degree human connections are retrieved.

The effect of the user context information classification module of the present invention is to assist intellectualized retrieval dictionary module to retrieve other keywords relevant with human connection search key, and enter intellectualized retrieval dictionary module by the user context information of new registration user dynamically being supplemented, so, can make the present invention that the dependency between the human connection search key of retrieval result and user's input when applying, can be improved constantly, it is achieved the intelligent functions of relevant human connection retrieval and machine learning function.Different from traditional dependency searching algorithm, in the present invention without operation personnel's craft typing dependency database, and, the present invention also has the feature of machine self-teaching, it is simple to popularization and application.

The present invention is after user inputs human connection search key, and intellectualized retrieval dictionary module is by generating relevant key word according to the information being stored in this module, thus realizing artificial intelligence.Further, this module is dynamically improved according to sorted user context information and the mutual relation thereof of the user of new registration and is supplemented storage content, it may be achieved self-teaching, can improve constantly the dependency between the human connection search key of retrieval result and user's input.

Further, user context information after described user context information classification module classification is as key word, difference according to generic is stored in different storage addresses, the storage address tag of key word its storage address subsidiary, for recording the type of key word, between any two difference storage address, the type according to storage information is preset with relationship between superior and subordinate；Wherein, relationship between superior and subordinate definition mode between any two key word is as follows: if described two key words are stored in different storage address, the then described relationship between superior and subordinate between two key words is equal to the relationship between superior and subordinate between the storage address of described two key word, if described two key words are stored in same storage address, then the relationship between superior and subordinate of two described key words is sane level.

When the present invention applies; user context information includes graduation universities and colleges, specialty, department, company and these five kinds of types of industry; but it is not limited to this five kinds of types when concrete application; any increase or reduce the similar approach of the above-mentioned type, or carry out, according to other standards, the similar approach classified and improve and each fall within protection scope of the present invention.In the specific implementation, due to graduation, universities and colleges are logically little with the dependency of specialty, department, company and industry, it is not necessary to typing intellectualized retrieval dictionary module for the present invention.Therefore, in the present invention, it is preferable that the user context information type of typing intellectualized retrieval dictionary module includes specialty, department, company, the big class of industry four.The relationship between superior and subordinate setting out four in the present invention in advance is followed successively by specialty, department, company and industry from the bottom to top, wherein, department, company and industry three are higher level's key word of specialty simultaneously, company and industry are higher level's key word of specialty and department simultaneously, and industry is higher level's key word of specialty, department and company three.

Described intellectualized retrieval dictionary module includes dictionary memory element, dictionary supplementary units and dictionary enquiring unit, wherein:

Dictionary memory element, for storing the relationship between superior and subordinate between the key word and the key word that comprise storage address tag, the relationship between superior and subordinate of each two key word is designed with accumulator register.The present invention setting by accumulator register so that be easy to when the present invention applies count the level of intimate between subordinate's key word and higher level's key word by big data means.

Dictionary supplementary units, dynamically improves the relationship between superior and subordinate between the key word and the key word that supplement in dictionary unit for the key word generated according to user context information classification module；Wherein, relationship between superior and subordinate between the dictionary supplementary units all key words to whether existing in dictionary memory element according to all key words of user context information classification module generation and generate according to the user context information of same user judges, if two key words in user context information have record in dictionary memory element, and the relationship between superior and subordinate between the two key word has corresponding record in dictionary memory element, then in dictionary memory element, the value of the accumulator register of the two key word relationship between superior and subordinate adds 1；If two key words in user context information have record in dictionary memory element, and the relationship between superior and subordinate between the two key word in dictionary memory element without respective record, in dictionary memory element, then it is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1；If two key words in user context information have record or two key word all no records in dictionary memory element only one of which, then by Unrecorded key word typing dictionary memory element, dictionary memory element is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1.

Dictionary enquiring unit, for generating relevant key word according to the human connection search key of user's input and feeding back to human connection retrieval module；Wherein, dictionary enquiring unit is generated relevant key word according to the human connection search key of user's input and is realized by two ways, concretely comprising the following steps of first kind of way:

A1, dictionary enquiring unit retrieve all higher level's key words of the human connection search key of user's input；

A2, according to the relationship between superior and subordinate between described all of higher level's key word, higher level's key word is ranked up, wherein, any two has described higher level's key word of relationship between superior and subordinate, the described higher level's key word being in junior position is arranged before the higher level's key word being in described in higher level status, described higher level's key word of any two sane level, and before the bigger higher level's key word of relationship between superior and subordinate accumulator register value between described human connection search key comes；

A3, one empty sequence of definition are as related keyword word sequence, the associative key generated for storing the human connection search key according to input；

A4, according to the higher level's keyword sequence after sequence, obtain all subordinates key word of each higher level's key word successively, and described all subordinates key word screened further, obtain and have subordinate's key word of identical storage address tag with human connection search key；According to the value of relationship between superior and subordinate accumulator register between described each subordinate's key word and described higher level's key word, described all subordinates key word is ranked up, before subordinate's key word that the value of relationship between superior and subordinate accumulator register is bigger comes, the subordinate's key word after sequence is stored in related keyword word sequence；

A5, repetition step A4, successively other higher level's key words are repeated above step, retrieve subordinate's key word of each higher level's key word successively to go forward side by side row filter and sequence, then related keyword word sequence it is stored in, if the subordinate's key word retrieved is existing record in related keyword word sequence, then need not repeat to be stored in related keyword word sequence；Wherein, in related keyword word sequence, key word sequence is gradually reduced with the dependency of the human connection search key of user's input from front to back.

Concretely comprising the following steps of the second way:

M1, dictionary enquiring unit retrieve all subordinates key word of the human connection search key of user's input；

M2, according to the relationship between superior and subordinate between described all of subordinate key word, subordinate's key word is ranked up, wherein, any two has described subordinate's key word of relationship between superior and subordinate, the described subordinate's key word being in higher level status is arranged before the subordinate's key word being in described in junior position, described subordinate's key word of any two sane level, and before the bigger subordinate's key word of relationship between superior and subordinate accumulator register value between described human connection search key comes；

M3, one empty sequence of definition are as related keyword word sequence, for storing the human connection search key according to input and the associative key of generation；

M4, according to sequence after subordinate's keyword sequence, obtain all higher level's key words of each subordinate key word successively, and described all higher level's key words are screened further, only to there being higher level's key word of identical storage address tag to be analyzed with human connection search key；According to the value of relationship between superior and subordinate accumulator register between described each higher level's key word and described subordinate's key word, described all higher level's key words are ranked up, before higher level's key word that the value of relationship between superior and subordinate accumulator register is bigger comes, the higher level's key word after sequence is stored in related keyword word sequence；

M5, repetition step M4, successively other subordinate's key words are repeated above step, retrieve higher level's key word of each subordinate's key word successively to go forward side by side row filter and sequence, then related keyword word sequence it is stored in, if the higher level's key word retrieved is existing record in related keyword word sequence, then need not repeat to be stored in related keyword word sequence；Wherein, in related keyword word sequence, key word sequence is gradually reduced with the dependency of the human connection search key of user's input from front to back.

Because there is certain dependency between human connection search key and its higher level's key word, it is possible to excavate out other key words relevant with human connection search key by this clue of higher level's key word；Such as, if the human connection search key word of input " civil engineering " this professional class, system can retrieve higher level's key word " engineering department " of this word, the specialty of a lot of colleagues in engineering department is all closely-related with " civil engineering ", for instance " engineering management ", " hydropower installation ", " plumbing " etc..By " engineering department " this clue, so that it may excavate out the key word of other the specialty classes relevant with " civil engineering ".Therefore, during the dictionary enquiring unit application of the present invention, retrieve all higher level's key words of the human connection search key of user's input, the retrieval range when present invention applies can be promoted.

During the dictionary enquiring unit application of the present invention, the reason that the higher level's key word having described in relationship between superior and subordinate is ranked up is as follows: analyze the relationship between superior and subordinate between all higher level's key words, being in the dependency between higher level's key word of higher level status and human connection search key can be more weak, because the scope that the higher level's key word being in higher level status includes is bigger, such as, if input " civil engineering ", service end retrieves its two higher level's key words " engineering department " and " Overseas Property ", and wherein " Overseas Property " is the higher level of " engineering department ".Because " Overseas Property " has subordinate's key word of department's classes such as " engineering department " " Finance Department " " sales department " " investment department " " Human Resource Departments ", and the work post of described department is completely different, thus it is multifarious for excavating specialty out.So the accuracy of the associative key excavating " civil engineering " according to " Overseas Property " this clue is just had a greatly reduced quality.

During the dictionary enquiring unit application of the present invention, the reason that higher level's key word of sane level is ranked up is as follows: mutually in higher level's key word of sane level, relatedness between higher level's key word and human connection search key that accumulator register value is bigger is bigger, also just more clue meaning.

The present invention makes corresponding sequence by dictionary enquiring unit, when the present invention applies, such as, input " civil engineering ", system excavates out some key words relevant to " civil engineering " according to its higher level's key word " engineering department ", such as " engineering management ", " plumbing ", " hydropower installation " etc., if the accumulator register value between " engineering management " and " engineering department " is the highest, then show that " engineering management " is the main specialty of the colleague that " engineering department " goes to work, so preferentially associating " engineering management " by " engineering department " this clue.

A kind of potential human connection searching system of intelligentized social platform, described system includes client and service end,

Client includes:

First receiver module, for receiving the human connection retrieval result that service end generates according to human connection search key；

Service end includes:

Further, described client or service end also include:

Described service end also includes:

Intellectualized retrieval dictionary module, for storing mutual relation between sorted user context information and each described sorted user context information of registered users, dynamically improves, according to the sorted user context information of user and the mutual relation thereof of new registration, the content supplementing described intellectualized retrieval dictionary module and the human connection search key according to user's input generates other key words that described key word is relevant；

Further, user context information after described user context information classification module classification is as key word, difference according to generic is stored in different storage addresses, the storage address tag of key word its storage address subsidiary, for recording the type of key word, between any two difference storage address, the type according to storage information is preset with relationship between superior and subordinate；Wherein, relationship between superior and subordinate definition mode between any two key word is as follows: if described two key words are stored in different storage address, the then described relationship between superior and subordinate between two key words is equal to the relationship between superior and subordinate between the storage address of described two key word, if described two key words are stored in same storage address, then the relationship between superior and subordinate of two described key words is sane level；

Described intellectualized retrieval dictionary module includes:

Dictionary memory element, for storing the relationship between superior and subordinate between the key word and the key word that comprise storage address tag, the relationship between superior and subordinate of each two key word is designed with accumulator register；

Dictionary supplementary units, dynamically improves the relationship between superior and subordinate between the key word and the key word that supplement in dictionary unit for the key word generated according to user context information classification module；Wherein, relationship between superior and subordinate between the dictionary supplementary units all key words to whether existing in dictionary memory element according to all key words of user context information classification module generation and generate according to the user context information of same user judges, if two key words in user context information have record in dictionary memory element, and the relationship between superior and subordinate between the two key word has corresponding record in dictionary memory element, then in dictionary memory element, the value of the accumulator register of the two key word relationship between superior and subordinate adds 1；If two key words in user context information have record in dictionary memory element, and the relationship between superior and subordinate between the two key word in dictionary memory element without respective record, in dictionary memory element, then it is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1；If two key words in user context information have record or two key word all no records in dictionary memory element only one of which, then by Unrecorded key word typing dictionary memory element, dictionary memory element is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1；

A5, repetition step A4, successively other higher level's key words are repeated above step, retrieve subordinate's key word of each higher level's key word successively to go forward side by side row filter and sequence, then related keyword word sequence it is stored in, if the subordinate's key word retrieved is existing record in related keyword word sequence, then need not repeat to be stored in related keyword word sequence；Wherein, in related keyword word sequence, key word sequence is gradually reduced with the dependency of the human connection search key of user's input from front to back.In the present invention, the value of accumulator register weighs the level of intimate between subordinate's key word and higher level's key word by being used to.

Concretely comprising the following steps of the second way:

Based on the potential human connection search method of above-mentioned a kind of intelligentized social platform potential human connection searching system, comprise the following steps:

When S1, client are registered according to user, the information of typing generates user context information, and sends user context information, human connection search key to service end；

S2, service end receive user context information, the human connection search key that client sends, and the user context information of reception is supplemented to user context information summary sheet；

Two degree of human connections of this user or the user context information of two degree and three degree human connections, according to human connection search key, are retrieved by S3, service end, generate the human connection retrieval result meeting described human connection search key；

Human connection is retrieved result and is sent to client by S4, service end；

S5, client receive the human connection retrieval result that service end generates according to human connection search key.

Further, described step S3 is further comprising the steps of:

According to the classification of information in user context information, user context information is classified；Two degree of human connections of this user or the user context information of two degree and three degree human connections are retrieved by described service end is additionally operable to the mutual relation between sorted user context information and each described sorted user context information of storage registered users, dynamically improve supplemental services end according to the sorted user context information of user and the mutual relation thereof of new registration content and other key words generating described key word relevant according to the human connection search key of user's input.

Further, the potential human connection search method of intelligentized social platform, further comprising the steps of: using described sorted user context information as key word, difference according to generic is stored in different storage addresses, the storage address tag of key word its storage address subsidiary, for recording the type of this key word, between any two difference storage address, the type according to storage information is preset with relationship between superior and subordinate；Wherein, relationship between superior and subordinate between any two key word adopts following methods definition: if described two key words are stored in different storage address, the then described relationship between superior and subordinate between two key words is equal to the relationship between superior and subordinate between the storage address of described two key word, if described two key words are stored in same storage address, then the relationship between superior and subordinate of two described key words is sane level；

Storage comprises the relationship between superior and subordinate between key word and the key word of storage address tag, and the relationship between superior and subordinate of each two key word is designed with accumulator register；

The relationship between superior and subordinate between the key word and the key word that supplement in dictionary unit is dynamically improved according to the key word of user context information classification module generation, to all key words generated according to user context information classification module whether dictionary memory element exists, and the relationship between superior and subordinate between all key words generated according to the user context information of same user judges, if two key words in user context information have record in dictionary memory element, and the relationship between superior and subordinate between the two key word has corresponding record in dictionary memory element, then in dictionary memory element, the value of the accumulator register of the two key word relationship between superior and subordinate adds 1；If two key words in user context information have record in dictionary memory element, and the relationship between superior and subordinate between the two key word in dictionary memory element without respective record, in dictionary memory element, then it is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1；If two key words in user context information have record or two key word all no records in dictionary memory element only one of which, then by Unrecorded key word typing dictionary memory element, dictionary memory element is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1；

Generate associative key according to the human connection search key of user's input, realized by two ways, concretely comprising the following steps of first kind of way: retrieve all higher level's key words of the human connection search key of user's input；According to the relationship between superior and subordinate between described all of higher level's key word, higher level's key word is ranked up, wherein, any two has described higher level's key word of relationship between superior and subordinate, the described higher level's key word being in junior position is arranged before the higher level's key word being in described in higher level status, described higher level's key word of any two sane level, and before the bigger higher level's key word of relationship between superior and subordinate accumulator register value between described human connection search key comes；Define an empty sequence as related keyword word sequence, the associative key generated for storing the human connection search key according to input；According to the higher level's keyword sequence after sequence, obtain all subordinates key word of each higher level's key word successively, and described all subordinates key word is screened further, acquirement and human connection search key have subordinate's key word of identical storage address tag, according to the value of relationship between superior and subordinate accumulator register between described each subordinate's key word and described higher level's key word, described all subordinates key word is ranked up, before subordinate's key word that the value of relationship between superior and subordinate accumulator register is bigger comes, subordinate's key word after sequence is stored in related keyword word sequence；Repeat previous step, retrieve subordinate's key word of each higher level's key word successively, described subordinate's key word is screened and sorts, and it is stored in related keyword word sequence, if the subordinate's key word retrieved is existing record in related keyword word sequence, then need not repeating to be stored in related keyword word sequence, wherein, in related keyword word sequence, key word sequence is gradually reduced with the dependency of the human connection search key of user's input from front to back；

Concretely comprising the following steps of the second way: dictionary enquiring unit retrieves all subordinates key word of the human connection search key of user's input；According to the relationship between superior and subordinate between described all of subordinate key word, subordinate's key word is ranked up, wherein, any two has described subordinate's key word of relationship between superior and subordinate, the described subordinate's key word being in higher level status is arranged before the subordinate's key word being in described in junior position, described subordinate's key word of any two sane level, and before the bigger subordinate's key word of relationship between superior and subordinate accumulator register value between described human connection search key comes；Define an empty sequence as related keyword word sequence, for storing the human connection search key according to input and the associative key of generation；According to the subordinate's keyword sequence after sequence, obtain all higher level's key words of each subordinate key word successively, and described all higher level's key words are screened further, only to there being higher level's key word of identical storage address tag to be analyzed with human connection search key；According to the value of relationship between superior and subordinate accumulator register between described each higher level's key word and described subordinate's key word, described all higher level's key words are ranked up, before higher level's key word that the value of relationship between superior and subordinate accumulator register is bigger comes, the higher level's key word after sequence is stored in related keyword word sequence；Repeat previous step, successively other subordinate's key words are repeated above step, retrieve higher level's key word of each subordinate's key word successively to go forward side by side row filter and sequence, then related keyword word sequence it is stored in, if the higher level's key word retrieved is existing record in related keyword word sequence, then need not repeat to be stored in related keyword word sequence；Wherein, in related keyword word sequence, key word sequence is gradually reduced with the dependency of the human connection search key of user's input from front to back.

The specialty that some user learns is completely uncorrelated with work, this situation is implicitly present in, described user is when social platform is registered, its user context information can be logged, the accuracy of later Dictionary retrieval function is adversely affected, in order to identify and delete these relationships between superior and subordinate having negative effect.Further, the potential human connection search method of intelligentized social platform, further comprising the steps of:

B1, retrieve all higher level's key words of a certain key word；

B2, respectively described each higher level's key word is done analysis below: obtain all subordinates key word of each higher level's key word successively, and described all subordinates key word is screened further, only obtain and the key word that is retrieved has subordinate's key word of identical storage address tag for analyzing；

B3, according to the value of relationship between superior and subordinate accumulator register between the subordinate's key word after described a certain higher level's key word and its screening, summation generates a summing value, if described in the value of relationship between superior and subordinate accumulator register that is retrieved between key word and described higher level's key word with the ratio of summing value less than the threshold value set, system is automatically deleted this relationship between superior and subordinate, or generation early warning information feeds back to operation personnel for operation personnel's reference, whether operation personnel's decision-making again deletes this relationship between superior and subordinate；

Other higher level's key words of the key word being retrieved are analyzed by B4, repetition step B2～B3.So, when the present invention applies, when " engineering department " work of the final Qu Liaoyijia property firm of user of " computer programming " specialty is learned in certain position, when this user registration, relationship between superior and subordinate between " computer programming " and " engineering department " can be logged dictionary, the accuracy of later Dictionary retrieval function can be adversely affected, the summing value of the value of the accumulator register that contrast " computer programming " is corresponding and the value of " engineering department " all of accumulator register, system easily finds, the former accounting is very little, almost can ignore, thus can be determined that the element being Dictionary retrieval function accuracy is had negative effect, such that it is able to delete described relationship between superior and subordinate.

Or/and based on step B1～B4 following steps developed:

X1, retrieve all subordinates key word of a certain key word；

X2, respectively described each subordinate key word is done analysis below: obtain all higher level's key words of each subordinate key word successively, and described all higher level's key words are screened further, only to there being higher level's key word of identical storage address tag to be analyzed with the key word being retrieved；

X3, according to the value of relationship between superior and subordinate accumulator register between the higher level's key word after described a certain subordinate key word and its screening, summation generates a summing value, if described in the value of relationship between superior and subordinate accumulator register between the key word and the described subordinate key word that are retrieved with the ratio of summing value less than the threshold value set, system is automatically deleted this relationship between superior and subordinate, or generation early warning information feeds back to operation personnel for operation personnel's reference, whether operation personnel's decision-making again deletes this relationship between superior and subordinate；

Other subordinate's key words of the key word being retrieved are analyzed by X4, repetition step X2～X3.

In order to promote retrieval rate further, the potential human connection search method of intelligentized social platform, further comprising the steps of:

C1, retrieve all higher level's key words of a certain key word, generate an empty sequence, for storing described higher level's key word；

C2, described higher level's key word is classified by storage address tag, respectively the higher level's key word comprising identical storage address tag is carried out analysis below: according to the value of the relationship between superior and subordinate accumulator register between the described a certain key word being retrieved and all higher level's key words having certain identical storage address tag, summation generates a summing value, if the value of relationship between superior and subordinate accumulator register between the described key word being retrieved and certain described higher level's key word and the ratio of summing value are less than the threshold value set, system is automatically deleted this relationship between superior and subordinate, or generation early warning information feeds back to operation personnel for operation personnel's reference, whether operation personnel's decision-making again deletes this relationship between superior and subordinate；

C3, repetition step C2, higher level's key word that the other types with the key word that is retrieved with relationship between superior and subordinate store address tag is analyzed.So, when the present invention applies, even if the situation of the relationship between superior and subordinate input system existed between " computer programming " and " engineering department ", but retrieval " computer programming " this key word, system may find " the research and development of products portion " of a lot of Computer Company, " software service portion ", " technical maintenance portion ", departments such as " engineering departments ", by the value of relationship between superior and subordinate accumulator is analyzed, being very easy to find " engineering department " is an abnormal type, thus can be determined that this is a relationship between superior and subordinate disturbing dictionary search function accuracy, can consider to delete.

Or/and based on step C1～C3 following steps developed:

Y1, retrieve all subordinates key word of a certain key word, generate an empty sequence, for storing described subordinate's key word；

Y2, described subordinate's key word is classified by storage address tag, respectively the subordinate's key word comprising identical storage address tag is carried out analysis below: according to the value of the relationship between superior and subordinate accumulator register between the described a certain key word being retrieved and all subordinate's key words having certain identical storage address tag, summation generates a summing value, if the value of relationship between superior and subordinate accumulator register between the described key word being retrieved and certain subordinate's key word described and the ratio of summing value are less than the threshold value set, system is automatically deleted this relationship between superior and subordinate, or generation early warning information feeds back to operation personnel for operation personnel's reference, whether operation personnel's decision-making again deletes this relationship between superior and subordinate；

Y3, repetition step Y2, subordinate's key word that the other types with the key word that is retrieved with relationship between superior and subordinate store address tag is analyzed.

When the present invention applies, the relationship between superior and subordinate that deleting has negative effect just should carry out after having a large number of users registration.

Further, described step S3 is further comprising the steps of: the user context information of all group friends of user once human connection is retrieved, and generates the human connection retrieval result meeting described human connection search key；Wherein, group friend is defined as: all add certain social group but two users not adding as a friend mutually.When the present invention applies, user may add certain industry personage social activity group using in social platform process, and by group friend is retrieved, the user that is more convenient for finds and oneself has a mind to the potential human connection got close to mutually of understanding.

In sum, the method have the advantages that and the degree of depth can excavate two degree of user, three degree potential human connections when (1) present invention applies, as certain user in the urgent need to recognize certain specific enterprise tenure, certain specific industry department or certain specific university specialty human connection time, go out whether this user has good friend just to have such human connection by fast searching of the present invention.So, it is easy to user when the present invention applies in two degree of human connections, three degree of human connections, is quickly found out the potential human connection oneself having a mind to recognize, be conducive to the popularization and application of social networks.

(2) present invention achieves the search function with artificial intelligence's characteristic in excavation process, after user inputs search key, other search keys relevant with described search key can be analyzed, and it is different with traditional associative key parser, operation team that need not be special when the present invention applies manually improves the data base of dependency between key word, can automatically set up dependency dictionary, there is self-teaching function, and accuracy along with the increase of number of users more and more accurate, in Cost Control and Consumer's Experience, there is greater significance.

(3) the invention provides two kinds of methods improving retrieval rate, after there is a large number of users registration, if finding, it is containing unworthy relationship between superior and subordinate, automatically this relationship between superior and subordinate is deleted, or generation early warning information feeds back to operation personnel for operation personnel's reference, whether operation personnel's decision-making again deletes this relationship between superior and subordinate.So, the present invention is avoided that when applying that key word the superior and the subordinate membership relation divides the unintelligible impact that retrieval accuracy is caused.

Accompanying drawing explanation

Accompanying drawing described herein is used for providing being further appreciated by the embodiment of the present invention, constitutes the part of the application, is not intended that the restriction to the embodiment of the present invention.In the accompanying drawings:

Fig. 1 is the structural representation of client in one specific embodiment of the present invention；

Fig. 2 is the structural representation of service end in one specific embodiment of the present invention；

Fig. 3 is the structural representation of one specific embodiment of human connection searching system in the present invention；

Fig. 4 is the relationship between superior and subordinate of background information classification in one specific embodiment of the present invention；

Fig. 5 is the flow chart of human connection search method in one specific embodiment of the present invention；

Fig. 6 is the flow chart generating other relevant key words of human connection search key in one specific embodiment of the present invention.

Detailed description of the invention

For making the object, technical solutions and advantages of the present invention clearly understand, below in conjunction with embodiment and accompanying drawing, the present invention is described in further detail, and exemplary embodiment and the explanation thereof of the present invention are only used for explaining the present invention, not as a limitation of the invention.

Embodiment:

As it is shown in figure 1, the potential human connection retrieval device of a kind of intelligentized social platform, being applied to and the client of service end communication connection, described device includes: user context information generating module, and during for registering according to user, the information of typing generates user context information；First sending module, for sending user context information, human connection search key to service end；First receiver module, for receiving the human connection retrieval result that service end generates according to human connection search key.The present embodiment is when specifically arranging, and client is additionally provided with the input module for Data Enter and the input of human connection search key and the display module shown for information.

As in figure 2 it is shown, the potential human connection retrieval device of a kind of intelligentized social platform, being applied to the service end being connected with client communication, described device includes: the second receiver module, for receiving user context information, the human connection search key that client sends；Userspersonal information's summary sheet module, generates user context information summary sheet for the user context information received according to the second receiver module；Human connection retrieval module, for according to human connection search key, two degree of human connections of this user or the user context information of two degree and three degree human connections are retrieved, generate the human connection retrieval result meeting described human connection search key, wherein, two degree of human connections are the good friend of user good friend, and three degree of human connections are the good friend of the good friend of user good friend；Second sending module, sends to client for human connection is retrieved result；User context information classification module, according to the classification of information in user context information, classifies to user context information；Intellectualized retrieval dictionary module, for storing mutual relation between sorted user context information and each described sorted user context information of registered users, dynamically improves, according to the sorted user context information of user and the mutual relation thereof of new registration, the content supplementing described intellectualized retrieval dictionary module and the human connection search key according to user's input generates other key words that described key word is relevant.

As it is shown on figure 3, the potential human connection searching system of a kind of intelligentized social platform, described system includes client and service end, and client includes user context information generating module, the first sending module and the first receiver module.Wherein, user context information generating module, during for registering according to user, the information of typing generates user context information；First sending module, for sending user context information, human connection search key to service end；First receiver module, for receiving the human connection retrieval result that service end generates according to human connection search key.

Service end includes the second receiver module, human connection retrieval module, the second sending module, user context information classification module and intellectualized retrieval dictionary module.Wherein, the second receiver module, for receiving user context information, the human connection search key that client sends；Userspersonal information's summary sheet module, generates user context information summary sheet for the user context information received according to the second receiver module；Human connection retrieval module, for according to human connection search key, two degree of human connections of this user or the user context information of two degree and three degree human connections being retrieved, generates the human connection retrieval result meeting described human connection search key；Wherein, two degree of human connections are the good friend of user good friend, and three degree of human connections are the good friend of the good friend of user good friend；Second sending module, sends to client for human connection is retrieved result；User context information classification module, according to the classification of information in user context information, classifies to user context information；Intellectualized retrieval dictionary module, for storing mutual relation between sorted user context information and each described sorted user context information of registered users, dynamically improves, according to the sorted user context information of user and the mutual relation thereof of new registration, the content supplementing described intellectualized retrieval dictionary module and the human connection search key according to user's input generates other key words that described key word is relevant.The human connection retrieval module of the present embodiment is additionally operable to other the relevant key words generated according to intellectualized retrieval dictionary module, and two degree of human connections of this user or the user context information of two degree and three degree human connections are retrieved.The present embodiment is when specifically arranging, and user context information classification module is not limited to be located at service end, it is possible to be located at client.

User context information after the user context information classification module classification of the present embodiment is as key word, difference according to generic is stored in different storage addresses, the storage address tag of key word its storage address subsidiary, for recording the type of key word, between any two difference storage address, the type according to storage information is preset with relationship between superior and subordinate.Wherein, relationship between superior and subordinate definition mode between any two key word is as follows: if described two key words are stored in different storage address, the then described relationship between superior and subordinate between two key words is equal to the relationship between superior and subordinate between the storage address of described two key word, if described two key words are stored in same storage address, then the relationship between superior and subordinate of two described key words is sane level.

During the present embodiment application; user context information includes graduation universities and colleges, specialty, department, company and these five kinds of types of industry; but it is not limited to this five kinds of types when concrete application; any increase or reduce the similar approach of the above-mentioned type, or carry out, according to other standards, the similar approach classified and improve and each fall within protection scope of the present invention.In the specific implementation, due to graduation, universities and colleges are logically little with the dependency of specialty, department, company and industry, it is not necessary to typing intellectualized retrieval dictionary module for the present embodiment.Therefore, in the present embodiment, it is preferable that the user context information type of typing intellectualized retrieval dictionary module includes specialty, department, company, the big class of industry four.As shown in Figure 4, the relationship between superior and subordinate setting out four in the present embodiment in advance is followed successively by specialty, department, company and industry from the bottom to top, wherein, in figure, arrow points is for be pointed to higher level by subordinate, namely department, company and industry three are higher level's key word of specialty simultaneously, company and industry are higher level's key word of specialty and department simultaneously, and industry is higher level's key word of specialty, department and company three.

The intellectualized retrieval dictionary module of the present embodiment includes: dictionary memory element, and for storing the relationship between superior and subordinate between the key word and the key word that comprise storage address tag, the relationship between superior and subordinate of each two key word is designed with accumulator register；Dictionary supplementary units, dynamically improves the relationship between superior and subordinate between the key word and the key word that supplement in dictionary unit for the key word generated according to user context information classification module；Dictionary enquiring unit, for generating relevant key word according to the human connection search key of user's input and feeding back to human connection retrieval module.Wherein, dictionary memory element storage comprises the relationship between superior and subordinate between key word and the key word of storage address tag, and the relationship between superior and subordinate of each two key word is designed with accumulator register.

During the present embodiment application, according to the classification of information in user context information, user context information is classified.Two degree of human connections of this user or the user context information of two degree and three degree human connections are retrieved by service end is additionally operable to the mutual relation between sorted user context information and each described sorted user context information of storage registered users, dynamically improve supplemental services end according to the sorted user context information of user and the mutual relation thereof of new registration content and other key words generating described key word relevant according to the human connection search key of user's input.Sorted user context information is as key word, difference according to generic is stored in different storage addresses, the storage address tag of key word its storage address subsidiary, for recording the type of this key word, between any two difference storage address, the type according to storage information is preset with relationship between superior and subordinate.Wherein, relationship between superior and subordinate between any two key word adopts following methods definition: if described two key words are stored in different storage address, the then described relationship between superior and subordinate between two key words is equal to the relationship between superior and subordinate between the storage address of described two key word, if described two key words are stored in same storage address, then the relationship between superior and subordinate of two described key words is sane level.

When the potential human connection searching system of intelligentized social platform of the present embodiment is applied, the relationship between superior and subordinate between the key word and the key word that supplement in dictionary unit is dynamically improved according to the key word of user context information classification module generation, to all key words generated according to user context information classification module whether dictionary memory element exists, and the relationship between superior and subordinate between all key words generated according to the user context information of same user judges, if two key words in user context information have record in dictionary memory element, and the relationship between superior and subordinate between the two key word has corresponding record in dictionary memory element, then in dictionary memory element, the value of the accumulator register of the two key word relationship between superior and subordinate adds 1；If two key words in user context information have record in dictionary memory element, and the relationship between superior and subordinate between the two key word in dictionary memory element without respective record, in dictionary memory element, then it is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1；If two key words in user context information have record or two key word all no records in dictionary memory element only one of which, then by Unrecorded key word typing dictionary memory element, dictionary memory element is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1.

Now add the engineering department of Poly real estate in real estate industry for the Zhang San of Sichuan University's civil engineering graduation, during the present embodiment application, if " civil engineering " in Zhang San's background information and " engineering department " two key words have record in dictionary memory element, and the relationship between superior and subordinate between the two key word has corresponding record in dictionary memory element, then in dictionary memory element, the value of the accumulator register of the two key word relationship between superior and subordinate adds 1；If two key words in Zhang San's background information have record in dictionary memory element, and the relationship between superior and subordinate between the two key word in dictionary memory element without respective record, in dictionary memory element, then it is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1；If two key words in Zhang San's background information have record or two key word all no records in dictionary memory element only one of which, then by Unrecorded key word typing dictionary memory element, dictionary memory element is stored in the relationship between superior and subordinate of described the two key word, and it being equipped with an accumulator register for this relationship between superior and subordinate, this accumulator register initial value is set to 1.Adopt aforesaid way, the combining form of " civil engineering ", " engineering department ", " Poly real estate " and " real estate industry " any two key word in Zhang San's background information is all judged, and ensures to fill into when lacking application key word and relationship between superior and subordinate mutually.

As shown in Figure 5, when the potential human connection searching system of intelligentized social platform of the present embodiment is applied to retrieve potential human connection, comprise the following steps: when S1, client are registered according to user, the information of typing generates user context information, and sends user context information, human connection search key to service end；S2, service end receive user context information, the human connection search key that client sends, and the user context information of reception is supplemented to user context information summary sheet；Two degree of human connections of this user or the user context information of two degree and three degree human connections, according to human connection search key, are retrieved by S3, service end, generate the human connection retrieval result meeting described human connection search key；Human connection is retrieved result and is sent to client by S4, service end；S5, client receive the human connection retrieval result that service end generates according to human connection search key.

As shown in Figure 6, the step S3 of the present embodiment is in the specific implementation, further comprising the steps of: S3.1, retrieve all higher level's key words of human connection search key of user's input；S3.2, according to the relationship between superior and subordinate between described all of higher level's key word, higher level's key word is ranked up, wherein, any two has described higher level's key word of relationship between superior and subordinate, the described higher level's key word being in junior position is arranged before the higher level's key word being in described in higher level status, described higher level's key word of any two sane level, and before the bigger higher level's key word of relationship between superior and subordinate accumulator register value between described human connection search key comes；S3.3, one empty sequence of definition are as related keyword word sequence, the associative key generated for storing the human connection search key according to input；S3.4, according to sequence after higher level's keyword sequence, obtain all subordinates key word of each higher level's key word successively, and described all subordinates key word is screened further, acquirement and human connection search key have subordinate's key word of identical storage address tag, for the analysis of next step；S3.5, according to the value of relationship between superior and subordinate accumulator register between described each subordinate's key word and described higher level's key word, described all subordinates key word is ranked up, before subordinate's key word that the value of relationship between superior and subordinate accumulator register is bigger comes, the subordinate's key word after sequence is stored in related keyword word sequence；S3.6, repetition step S3.4～step S3.5, retrieve all subordinates key word of each described higher level's key word successively, and it is stored in related keyword word sequence to after its screening and sequence, wherein, in related keyword word sequence, key word sequence is gradually reduced with the dependency of the human connection search key of user's input from front to back；Two degree of human connections of this user or the user context information of two degree and three degree human connections are retrieved by S3.7, foundation human connection search key and other key words relevant to the human connection search key that is retrieved of generation.The present embodiment passes through step S3.1～S3.6, can excavate out other key words relevant with human connection search key according to this clue of higher level's key word.The step S3 of the present embodiment in the specific implementation, also adopts mode similar for step S3.1～S3.6, can excavate out other key words relevant with human connection search key according to subordinate's this clue of key word.The present embodiment generates relevant higher level's key word according to the human connection search key of user's input and is exemplified below: if engineering department's working of the student all Qu Liao property firms learning civil engineering of most registration, when retrieval " civil engineering ", it is engineering department that system can count maximally related department according to the graduation whereabouts of the student of civil engineering, professional dependency between colleague in engineering department is higher, so the degree of association between subordinate's specialty class keywords of department level key word can be higher, on the contrary, because having much different departments in company, such as the work post between Finance Department and engineering department is widely different, if so passing through to analyze higher level company class higher level's key word of the key word of certain specialty class, the result precision analyzed can be had a greatly reduced quality.

In order to promote accuracy during the present embodiment application, during the present embodiment application, following two mode is also adopted to delete the relationship between superior and subordinate of interference accuracy.First kind of way is:

After there is a large number of users registration, carry out following steps:

B1, retrieve all higher level's key words of a certain key word；

B3, according to the value of relationship between superior and subordinate accumulator register between the subordinate's key word after described a certain higher level's key word and its screening, summation generates a summing value, if described in the value of relationship between superior and subordinate accumulator register that is retrieved between key word and described higher level's key word with the ratio of summing value less than the threshold value (such as 0.1%) set, system is automatically deleted this relationship between superior and subordinate, or generation early warning information feeds back to operation personnel for operation personnel's reference, whether operation personnel's decision-making again deletes this relationship between superior and subordinate；

Other higher level's key words of the key word being retrieved are analyzed by B4, repetition step B2～B3.Wherein, all subordinates key word of the key word being retrieved, in use, be also based on the similar principle of step B1～B4, is analyzed by first kind of way.

The second way is:

After there is a large number of users registration, carry out following steps:

C3, repetition step C2, higher level's key word that the other types with the key word that is retrieved with relationship between superior and subordinate store address tag is analyzed.Wherein, all subordinates key word with the key word that is retrieved with relationship between superior and subordinate, in use, be also based on the similar principle of step C1～C3, is analyzed by the second way.

When the present embodiment deletes relationship between superior and subordinate, it is preferable that increase premised on 1,000,000 by number of users.The present embodiment less relationship between superior and subordinate of reduced value after a large number of users is registered is deleted, for the relationship between superior and subordinate relevant to specialty, and impact on dictionary accuracy when the people that changing majors is engaged in non-professional work can be prevented to register.

The relationship between superior and subordinate that two ways deletes interference accuracy is adopted to be exemplified below for the present embodiment: some user has learned computer major, the engineering department of result Qu Liao property firm, when he registers time, relationship between superior and subordinate between " computer programming " and " engineering department " the two key word also can be automatically replenished into dictionary, influencing whether the accuracy of whole Dictionary retrieval function, described the present embodiment adopts two kinds of methods for eliminating the negative effect to dictionary enquiring unit accuracy of this uncommon situation.The principle of first method is: retrieves higher level's key word " engineering department " of key word " computer programming " and analyzes all subordinates key word of " engineering department ", find professional class keywords mainly " civil engineering ", " engineering management " etc. that the degree of association is high between " engineering department " this key word, the degree of association of " computer programming " and " engineering department " can be much lower, therefore it is contemplated that the relationship between superior and subordinate deleted between " computer programming " and " engineering department ".The principle of second method is: retrieve the key word of whole departments class of key word " computer programming ", such as " computer development portion " " data base administration portion " " technical support department " " engineering department ", in like manner, by analyzing it appeared that the degree of association between degree of association ratio " computer programming " and other several key words between " computer programming " and " engineering department " is substantially much lower, therefore it is contemplated that relationship between superior and subordinate between deletion " computer programming " and " engineering department ".

The step S3 of the present embodiment is further comprising the steps of: the user context information of all group friends of user once human connection is retrieved, and generates the human connection retrieval result meeting described human connection search key.User may add certain industry personage social activity group using in social platform process, and group friend is defined as: all adds certain social group but two users not adding as a friend mutually.Concrete example during the present embodiment retrieval group friend is as follows: user search " Poly real estate ", the good friend Zhang San demonstrating this user has participated in " Chengdu social club of real estate industry ", and this tissue has a member Li Si just to appoint marketing manager at Poly real estate.

During the present embodiment application, certain user inputs " Bank of China ", it has been found that he has a friend Zhang San just to have a friend Li Si working as customer manager in Chengdu company of Bank of China, and this user just can look for Zhang San to link up subsequently, and request Zhang San recommends understanding Li Si.

Above-described detailed description of the invention; the purpose of the present invention, technical scheme and beneficial effect have been further described; it is it should be understood that; the foregoing is only the specific embodiment of the present invention; the protection domain being not intended to limit the present invention; all within the spirit and principles in the present invention, any amendment of making, equivalent replacement, improvement etc., should be included within protection scope of the present invention.

Claims

1. the potential human connection retrieval device of intelligentized social platform, is applied to and the client of service end communication connection, it is characterised in that described device includes:

First receiver module, for receiving the human connection retrieval result that service end generates according to human connection search key.

2. the potential human connection retrieval device of intelligentized social platform, is applied to the service end being connected with client communication, it is characterised in that described device includes:

3. the potential human connection retrieval device of the intelligentized social platform of one according to claim 2, is applied to the service end being connected with client communication, it is characterised in that described device also includes:

4. the potential human connection retrieval device of the intelligentized social platform of one according to claim 3, is applied to the service end being connected with client communication, it is characterised in that

User context information after described user context information classification module classification is as key word, difference according to generic is stored in different storage addresses, the storage address tag of key word its storage address subsidiary, for recording the type of key word, between any two difference storage address, the type according to storage information is preset with relationship between superior and subordinate；Wherein, relationship between superior and subordinate definition mode between any two key word is as follows: if described two key words are stored in different storage address, the then described relationship between superior and subordinate between two key words is equal to the relationship between superior and subordinate between the storage address of described two key word, if described two key words are stored in same storage address, then the relationship between superior and subordinate of two described key words is sane level；

Described intellectualized retrieval dictionary module includes:

A5, repetition step A4, successively other higher level's key words are repeated above step, retrieve subordinate's key word of each higher level's key word successively to go forward side by side row filter and sequence, then related keyword word sequence it is stored in, if the subordinate's key word retrieved is existing record in related keyword word sequence, then need not repeat to be stored in related keyword word sequence；Wherein, in related keyword word sequence, key word sequence is gradually reduced with the dependency of the human connection search key of user's input from front to back；

Concretely comprising the following steps of the second way:

5. the potential human connection searching system of intelligentized social platform, it is characterised in that described system includes client and service end,

Client includes:

Service end includes:

6. the potential human connection searching system of the intelligentized social platform of one according to claim 5, it is characterised in that described client or service end also include:

Described service end also includes:

7. the potential human connection searching system of the intelligentized social platform of one according to claim 6, it is characterized in that, user context information after described user context information classification module classification is as key word, difference according to generic is stored in different storage addresses, the storage address tag of key word its storage address subsidiary, for recording the type of key word, between any two difference storage address, the type according to storage information is preset with relationship between superior and subordinate；Wherein, relationship between superior and subordinate definition mode between any two key word is as follows: if described two key words are stored in different storage address, the then described relationship between superior and subordinate between two key words is equal to the relationship between superior and subordinate between the storage address of described two key word, if described two key words are stored in same storage address, then the relationship between superior and subordinate of two described key words is sane level；

Described intellectualized retrieval dictionary module includes:

Concretely comprising the following steps of the second way:

8. based on the potential human connection search method of a kind of intelligentized social platform potential human connection searching system according to any one of claim 5～7, it is characterised in that comprise the following steps:

9. the potential human connection search method of intelligentized social platform according to claim 8, it is characterised in that described step S3 is further comprising the steps of:

10. the potential human connection search method of intelligentized social platform according to claim 9, it is characterized in that, further comprising the steps of: using described sorted user context information as key word, difference according to generic is stored in different storage addresses, the storage address tag of key word its storage address subsidiary, for recording the type of this key word, between any two difference storage address, the type according to storage information is preset with relationship between superior and subordinate；Wherein, relationship between superior and subordinate between any two key word adopts following methods definition: if described two key words are stored in different storage address, the then described relationship between superior and subordinate between two key words is equal to the relationship between superior and subordinate between the storage address of described two key word, if described two key words are stored in same storage address, then the relationship between superior and subordinate of two described key words is sane level；

11. the potential human connection search method of intelligentized social platform according to claim 10, it is characterised in that further comprising the steps of:

B1, retrieve all higher level's key words of a certain key word；

Other higher level's key words of the key word being retrieved are analyzed by B4, repetition step B2～B3；

Or/and based on step B1～B4 following steps developed:

X1, retrieve all subordinates key word of a certain key word；

12. the potential human connection search method of intelligentized social platform according to claim 10, it is characterised in that further comprising the steps of:

C2, described higher level's key word is classified by storage address tag, respectively the higher level's key word comprising identical storage address tag is carried out analysis below: according to the value of the relationship between superior and subordinate accumulator register between the described a certain key word being retrieved and all higher level's key words having identical storage address tag, summation generates a summing value, if the value of relationship between superior and subordinate accumulator register between the described key word being retrieved and certain described higher level's key word and the ratio of summing value are less than the threshold value set, system is automatically deleted this relationship between superior and subordinate, or generation early warning information feeds back to operation personnel for operation personnel's reference, whether operation personnel's decision-making again deletes this relationship between superior and subordinate；

C3, repetition step C2, higher level's key word that the other types with the key word that is retrieved with relationship between superior and subordinate store address tag is analyzed；

Or/and based on step C1～C3 following steps developed:

13. the potential human connection search method of the intelligentized social platform of one described in any one according to Claim 8～12, it is characterized in that, described step S3 is further comprising the steps of: the user context information of all group friends of user once human connection is retrieved, and generates the human connection retrieval result meeting described human connection search key；Wherein, group friend is defined as: all add certain social group but two users not adding as a friend mutually.