CN105447188A - Knowledge learning based peer-to-peer social network document retrieval method - Google Patents
Knowledge learning based peer-to-peer social network document retrieval method Download PDFInfo
- Publication number
- CN105447188A CN105447188A CN201510955213.4A CN201510955213A CN105447188A CN 105447188 A CN105447188 A CN 105447188A CN 201510955213 A CN201510955213 A CN 201510955213A CN 105447188 A CN105447188 A CN 105447188A
- Authority
- CN
- China
- Prior art keywords
- node
- interest
- list
- knowledge
- recommended
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Abstract
The invention discloses a knowledge learning based peer-to-peer social network document retrieval method, which belongs to the technical field of the social network and mainly solves the problems of low document retrieval recall rate, high retransmission overhead and non-ideal performance in a peer-to-peer social network. The method comprises the following steps: A) a node establishes an interest index and a knowledge index; and B) the node obtains a recommendation node according to the interest index and the knowledge index to retransmit a query message. The node establishes the interest index according to a document amount obtained by the interest vector keyword of the node which has the same interest with the node, and establishes the knowledge index according to the document amount obtained according to query keywords. When the query message is retransmitted, the recommendation node is obtained from the interest index and the knowledge index, and the message is retransmitted to the recommendation node. The method learns knowledge and establishes the indexes in a query process and obtains better query according to the indexes. Compared with the traditional typical methods, the method disclosed by the invention obtains a high recall rate, a low network overhead and good performance.
Description
Technical field
The present invention relates to social networks technical field, knowledge learning equity social networks document retrieval method (IESLP) the present invention proposed is for the file retrieval of reciprocity social networking service.
Background technology
Social networks application is more and more extensive, and people can realize the doings of virtual community by social networks, as commercial advertisement of making friends, chat, help each other, issue, carry out resource sharing and retrieval etc.Social networks has polytype, has online social networks based on Client/Server pattern as FaceBook, Renren Network etc.; There is the mobile social networking based on honeycomb soverlay technique, as micro-letter; Also the reciprocity social networks based on P2P technology is had.
P2P technology also claims peer-to-peer, and the node in reciprocity social networks is server is also client.In the social networks using non-structural P 2 P technique construction, search file is a challenging problem.Existing certain methods may be used for the file retrieval of reciprocity social networks at present, breadth-first search technology (RBFS) [V.Kalogeraki as random in (1), D.Gunopulos, D.Zeinalipour-yazti.ALocalSearchMechanismforPeer-to-Peer Networks [C] .Proc.Ofthe11thACMConferenceonInformationandKnowledgeMan agement (CIKM ' 02) .NewYork:ACM, 2002:300-307.].(2) NeuroGrid technology [JosephS.NeuroGrid:Semanticallyroutingqueriesinpeer-to-pe ernetworks.ProceedingsoftheInternationalWorkshoponPeer-t o-PeerComputing, Pisa, Italy, 2002.].(3) ESLP technology [L.Liu, N.Antonopoulos, S.Mackin, J.Xu, D.Russell, EfficientResourceDiscoveryinSelf-organizedUnstructuredPe er-to-PeerNetworks, ConcurrencyandComputation:PracticeandExperience, Wiley, Vol23 (2), February2009, pp.159-183.] etc.RBFS method is when carrying out query messages process, and a Stochastic choice k neighbor node carries out query messages forwarding, and receive neighbor node Stochastic choice k neighbours' forwarding inquiries message again of message, until TTL exhausts, the method efficiency of this search file is low, time delay is long.NeuroGrid technology sets up knowledge base at network node, by query script learning to knowledge store in knowledge base, carry out message forwarding according to the knowledge-chosen recommended node in knowledge base, forward node number is between minimum forwarding degree and max-forwards degree.The method has had improvement than RBFS, and recall rate and network performance have had raising.ESLP is a kind of reciprocity social network search technology of novelty, it applies to the human relation theory in people's contacts process in file retrieval, the Social behaviors Fast Learning knowledge of simulation people, spontaneous formation social circle, improve the success ratio of file retrieval, the method has had larger improvement at aspect of performance than NeuroGrid.
Although these methods existing can realize the file retrieval of reciprocity social networks, also there are some defects.The forwarding degree of RBFS is a fixing constant; The forwarding degree of NeuroGrid, but can not adaptive change between minimum forwarding degree and max-forwards degree; Although the forwarding degree of ESLP considers the correlativity adaptive change of destination node and searching keyword, do not consider the relation of destination node number of documents and searching keyword, and do not consider the adaptive change of minimum forwarding degree and max-forwards degree yet.In social networks, user forms community by interest, and these algorithms do not have the interest attribute of explicit digging user yet, thus passes through the comparison self-teaching of interest vector similarity, and quick clustering becomes community.These reciprocity social networks file retrieval technical methods existing are in recall rate, the space forwarding lifting in addition in expense, network performance etc.
Summary of the invention
In order to solve above-mentioned reciprocity social network environment Documents retrieval recall rate compared with low, that network overhead large, performance is not high defect, the present invention proposes a kind of knowledge based and learn reciprocity social networks document retrieval method (being called for short IESLP), network node is according to interest vector and searching keyword self-teaching, automatic Community Formation cluster, to improve file retrieval recall rate under reciprocity social network environment and combination property.The effect adopting each node in the reciprocity social networks of IESLP document retrieval method is equal, node perform document retrieval routing algorithm divides two kinds of situations: produce when this querying node document and send initial query message, and sets up according to the feedback message of other node in network or upgrade interest index and knowledge index; When this node receives the query messages from other node by the number of documents of statistics this locality with Keywords matching, send feedback message to query node, and select the neighbor node forwarding inquiries message of oneself.Realizing technical scheme of the present invention comprises as follows:
A kind of knowledge based learns reciprocity social networks document retrieval method, comprises the steps:
Steps A, node sets up interest index and knowledge index, comprise: in file retrieval process, node obtains the similar knowledge store of interest local interest concordance list from the destination node that interest is identical, simultaneously obtains the knowledge store of mating with searching keyword in Indigenous knowledge concordance list according to searching keyword;
Step B, node obtains neighbor node as recommended node by local interest index and knowledge index, and to recommended node forwarding inquiries message, comprise: when node needs to forward the query messages that other nodes send over, inquire about local interest concordance list and knowledge index table and obtain the neighboring node list comprising matching inquiry keyword document, according to the correlation coefficient of neighbor node in coupling number of documents calculations list, and calculate self-adaptation forwarding degree in conjunction with minimum forwarding degree and max-forwards degree, then the self-adaptation forwarding degree according to neighbor node in list selects recommended node to carry out query messages forwarding.
As optimal technical scheme, the process that described steps A sets up interest index and knowledge index is as follows:
Step 1), when query node is identical with destination node interest, the number of documents that the interest vector keyword of query node foundation destination node feedback and destination node comprise interest vector keyword is set up or upgrades local interest concordance list;
Step 2), when query node is different with destination node interest, or query node is identical with destination node interest and destination node interest vector lists of keywords does not comprise searching keyword time, the number of documents that query node comprises searching keyword according to searching keyword and the destination node of destination node feedback is set up or upgrades Indigenous knowledge concordance list;
Step 3), destination node is according to searching keyword statistical match number of documents.
As optimal technical scheme, in described step B using neighbor node as the decision method of recommended node be: if the self-adaptation of neighbor node forwards number of degrees value and is greater than the quantity selecting recommended node in list, then this neighbor node is chosen as recommended node.
As optimal technical scheme, the process that described step B selection recommended node carries out query messages forwarding is as follows:
Step 1) be included in the interest vector of node when searching keyword, then from interest concordance list, node is selected to add in recommended node list according to searching keyword, when recommendation list interior joint lazy weight max-forwards number of degrees value, then from knowledge index table, node is selected to add in recommended node list according to searching keyword, when recommendation list interior joint lazy weight minimum forwarding number of degrees value, then from neighbor list, Stochastic choice residue joint adds in recommended node list;
Step 2) when inquiring about in the interest vector of subject key words not at present node, then from knowledge index table, node is selected to add in recommended node list according to searching keyword, when recommendation list interior joint lazy weight minimum forwarding number of degrees value, then from neighbor list, Stochastic choice residue joint adds in recommended node list;
Step 3) when all not meeting the recommended node of search request in interest concordance list and knowledge index table, expanding hunting zone, increasing the numerical value of minimum forwarding degree; From the connection neighbor node of present node, Stochastic choice meets node that up-to-date minimum forwarding degree requires as recommended node, adds in recommended node list;
4) from recommended node list, node forwarding inquiries message is selected successively, until recommended node list is empty.
As optimal technical scheme, calculate correlation coefficient in described step B, be specially:
The correlation coefficient of i-th node obtained from interest concordance list or knowledge index table according to searching keyword is
wherein i=1,2 ..., n, n are the number of the neighbor node that foundation searching keyword obtains from interest concordance list or knowledge index table, the molecule m of item on the right of equation
ibe the number of documents that i-th node mates with searching keyword, denominator
for the coupling number of documents of node that obtains from interest concordance list or knowledge index table according to searching keyword and.
As optimal technical scheme, the value calculating self-adaptation forwarding degree in described step B is:
The self-adaptation forwarding degree k of i-th node
i=Round (r
i λ(d
max-d
min))+d
min, wherein i=1,2 ..., n, r
ibe the correlation coefficient of i-th node, d
minfor minimum forwarding degree, namely minimum by the neighbor node number of selection forwarding inquiries message, d
maxfor max-forwards degree, namely maximum by the node number of selection forwarding inquiries message, λ is index regulatory factor, and scope is between 0 ~ 1, and Round function is bracket function.
As optimal technical scheme, described minimum forwarding degree d
min=2, max-forwards degree d
max=3, index regulatory factor λ=0.7.
As optimal technical scheme, described interest concordance list and described knowledge index table are the identical Hash table structure of structure, and keyword forms the key of Hash table, and the neighboring node list comprising matching inquiry keyword document forms the cryptographic hash of key.
As optimal technical scheme, the element in described list comprises two territories, and a territory stores information of neighbor nodes, and another territory stores the number of documents that this neighbor node comprises corresponding Keywords matching; Comprise the element of information of neighbor nodes in described list by number of files value Bit-reversed from big to small, the node acquiescence coupling number of files value of Stochastic choice is 0, comes list end.Compared with prior art, beneficial effect of the present invention:
The present invention is at reciprocity social networks joint structure interest index and knowledge index, in file retrieval process, same interest information of neighbor nodes is stored in oneself interest concordance list by number of files value by node from big to small, and the nodal information retrieved by searching keyword is stored in knowledge index table by number of documents from big to small, during forwarding inquiries message, node arrives the more neighbor node of knowledge acquisition number of documents as forwarding recommended node according to study.So, improve recall rate and the retrieval performance of reciprocity social networks file retrieval.
Accompanying drawing explanation
Fig. 1 is interest concordance list and knowledge index list structure figure;
Fig. 2 is the enforcement illustration setting up interest index and knowledge index;
Fig. 3 is that node obtains the enforcement illustration of recommended node forwarding inquiries message by interest index and knowledge index;
Fig. 4 is the comparison of technical solution of the present invention IESLP and ESLP, NeuroGrid method recall rate;
The comparison that Fig. 5 is technical solution of the present invention IESLP and ESLP, NeuroGrid method forwards expense;
Fig. 6 is technical solution of the present invention IESLP and ESLP, NeuroGrid method often forwards the comparison that a query messages obtains recall rate.
Embodiment
Below by the drawings and specific embodiments, further description is done to technical solution of the present invention.
Realization of the present invention comprises two steps A and B:
Steps A: node sets up the step of interest index and knowledge index;
When node (query node) inquires about document, generated query message also sends query messages to neighbor node, query node obtains the similar knowledge store of interest local interest concordance list from the destination node that interest is identical, obtains required knowledge store in Indigenous knowledge concordance list according to searching keyword simultaneously.The process setting up interest index and knowledge index is as follows:
1) when query node is identical with destination node interest, the number of documents that the interest vector keyword of query node foundation destination node feedback and destination node comprise interest vector keyword is set up or upgrades local interest concordance list;
2) when query node is different with destination node interest, or query node is identical with destination node interest and destination node interest vector lists of keywords does not comprise searching keyword time, the number of documents that query node comprises searching keyword according to searching keyword and the destination node of destination node feedback is set up or upgrades Indigenous knowledge concordance list;
3) destination node is according to searching keyword statistical match number of documents.
Fig. 1 is the structural drawing of interest concordance list of the present invention and knowledge index table.Interest concordance list and knowledge index table are the identical Hash table structure of structure, and keyword forms the key of Hash table, and the neighboring node list of matching inquiry keyword document forms the cryptographic hash of key; Element in list comprises two territories, and territory stores neighbor node address information, and another territory stores this neighbor node and comprises the number of documents that corresponding keyword (key) mates.Comprise the element of information of neighbor nodes in list by number of files value Bit-reversed from big to small, the node acquiescence coupling number of files value of Stochastic choice is 0, comes list end.
In Fig. 1, left-hand line Topic_n in concordance list
i(wherein i=1,2 ..., y, y are the quantity of key) and be the key of Hash table, be made up of interest keyword or searching keyword; Right side is the ltsh chain table that key is corresponding, the Node_n in chained list node
i(i=1,2 ..., x) be the neighbor node of query node, m
i(i=1,2 ..., x) for neighbor node comprises the number of documents of keyword (key), and m
i>m
i+1, x is the quantity of ltsh chain table node.
Fig. 2 is the embodiment setting up interest index and knowledge index.In fig. 2, node A, B, D have same interest, and interest vector (InterestVector) is
node C, E interest is identical, and their interest vector is
node A is interested in theme " iPhone ", and produce query messages, encapsulate theme " iPhone " in message, query messages is broadcast to neighbor node B, C, D respectively by A.The interest of Node B, D with A is identical, and they read keyword C#, Java and Python in interest vector respectively, searches the local document comprising these three keywords, and according to keyword statistic document quantity.Statistical information is fed back to A by B and D, and the information received writes in interest concordance list (InterestIndex) by node A.The document B comprising C# theme has 9 sections, and D has 6 sections, and therefore, in interest concordance list, in the knot vector of corresponding C# keyword, B should come before D.And the document B comprising Java keyword has 3 sections, D has 8 sections, thus in interest concordance list corresponding Java keyword knot vector in D should come before B.Comprise each 6 sections of document B and D of Python keyword, A stores information according to the sequencing receiving feedback message.Suppose in Fig. 2, A first receives the feedback of B, after receive the feedback of D, then B comes before D.
The interest of node C with A is different, but has the interested theme of A " iPhone " in the local document of node C.Node C searches the document comprising theme " iPhone " in this locality, the document matches number of times often finding to meet the demands increases one.Find 8 sections of documents altogether, node C successful search number of times is 8.C by information feed back to A, node A by the information received with " iPhone " for keyword is stored in knowledge index table (KnowledgeIndex).Message is transmitted to neighbor node E by C.Node E finds the document 2 sections comprising theme " iPhone " in local document, then node E successful search number of times is 2.Now, in network successful search number of times by increase by 10 sections (8+2=10).E is by information feed back to A, and in the knowledge index table of A, in the knot vector being keyword with " iPhone ", C has 8 sections of documents, and node A will be inserted in after C, and is established to the connection of E node after receiving the feedback information of E.
It is worth noting, although do not comprise inquiry subject key words in the interest vector of the destination node identical with query node interest, likely there is the interested document of query node in destination node, therefore, destination node, except searching by interest, also should be searched by searching keyword.In Fig. 2, have one section of document package in Node B containing " iPhone ", therefore the successful search number of times of B node is that the information received is stored in knowledge index to A, A by 1, B node feeding back information, is inserted into after node E.
Step B: node obtains neighbor node as recommended node by local interest index and knowledge index, and to the step of recommended node forwarding inquiries message.
When node receives the query messages that other node sends over, if need forwarding inquiries message, then inquire about local interest concordance list and knowledge index table and obtain the neighboring node list comprising matching inquiry keyword document, according to the correlation coefficient of neighbor node in coupling number of documents calculations list, and calculating self-adaptation forwarding degree in conjunction with minimum forwarding degree and max-forwards degree, the self-adaptation forwarding degree then according to neighbor node in list selects recommended node to carry out query messages forwarding.If the self-adaptation forwarding number of degrees value of neighbor node is greater than the quantity selecting recommended node in list, then this neighbor node is chosen as recommended node.
According to the correlation coefficient of neighbor node in coupling number of documents calculations list, be specially:
The correlation coefficient of i-th node obtained from interest concordance list or knowledge index table according to searching keyword is
wherein i=1,2 ..., n, the molecule m of item on the right of equation
ibe the number of documents that i-th node mates with searching keyword, denominator
for the coupling number of documents of node that obtains from interest concordance list or knowledge index table according to searching keyword and, n is the number of the neighbor node obtained from interest concordance list or knowledge index table according to searching keyword.
Calculate self-adaptation forwarding degree to be specially:
The self-adaptation forwarding degree k of i-th node
i=Round (r
i λ(d
max-d
min))+d
min, wherein i=1,2 ..., n, r
ibe the correlation coefficient of i-th node, d
minfor minimum forwarding degree, namely minimum by the neighbor node number of selection forwarding inquiries message, d
maxfor max-forwards degree, namely maximum by the node number of selection forwarding inquiries message, λ is index regulatory factor, and scope is between 0 ~ 1, Round function is bracket function, and n is the number of the neighbor node that foundation searching keyword obtains from interest concordance list or knowledge index table.。
The process that described selection recommended node carries out query messages forwarding is as follows:
1) when searching keyword is included in the interest vector of node, then from the local interest concordance list of node, the neighbor node comprising matching inquiry keyword document is selected to add in recommended node list according to searching keyword; When neighbor node lazy weight max-forwards number of degrees value in recommended node list, then from the Indigenous knowledge concordance list of node, the neighbor node comprising matching inquiry keyword document is selected to add in recommended node list according to searching keyword, if neighbor node lazy weight minimum forwarding number of degrees value in now recommended node list, then from the neighboring node list of node, Stochastic choice residue neighbor node adds in recommended node list, until neighbor node quantity reaches minimum forwarding number of degrees value in recommended node list;
2) when inquiring about in the interest vector of subject key words not at node, then from the Indigenous knowledge concordance list of node, the neighbor node comprising matching inquiry keyword document is directly selected to add in recommended node list according to searching keyword, when neighbor node lazy weight minimum forwarding number of degrees value in recommended node list, then from the neighboring node list of node, Stochastic choice residue neighbor node adds in recommended node list, until neighbor node quantity reaches minimum forwarding number of degrees value in recommended node list;
3) when all not meeting the recommended node of search request in the local interest concordance list and knowledge index table of node, expanding hunting zone, increasing the numerical value of minimum forwarding degree.From the neighbor node of node, Stochastic choice meets neighbor node that up-to-date minimum forwarding degree requires as recommended node, adds in recommended node list;
4) from recommended node list, neighbor node forwarding inquiries message is selected successively, until recommended node list is empty.
Fig. 3 is the embodiment being obtained recommended node forwarding inquiries message by interest index and knowledge index.Destination node A has 7 connected node B, C, D, E, F, G, H, and wherein A and B, C, D have common interest vector
e and F has common interest vector
suppose that minimum forwarding degree is 2, max-forwards degree is 3, index regulatory factor λ=0.7.If inquiry subject key words is " Google ", then A obtains interdependent node B, C, D from interest concordance list, and falls to sort by coupling number of documents, calculates three node actual forwarding degree k
b=3, k
c=3, k
d=2.During beginning, recommended node quantity is 0 be less than k
b=3, therefore select B, now recommended node quantity is 1 be not more than k
c=3, C is selected.Select to recommend after C number of nodes to be 2 and k
d=2 is equal, stops selecting.Because recommended node quantity does not reach max-forwards degree 3, then inquire about knowledge index table, there is no the node of matching inquiry keyword " Google " in knowledge index table, and recommended node quantity reaches minimum forwarding degree 2, stop selection course, final B, C are chosen as recommended node.If inquiry subject key words is " Baidu ", then A obtains recommended node C from interest concordance list, because recommended node quantity does not exceed max-forwards degree, then from knowledge index table, recommended node E is obtained, now recommended node quantity is 2, reach the requirement of minimum forwarding degree, stop selection course, final C, E are chosen as recommended node.If inquiry subject key words is " Bing ", the node that interest concordance list and knowledge index table all do not meet the demands, then Stochastic choice G, H node is as recommended node.If inquiry subject key words is " iPhone ", it not in the interest vector of A, then obtains interdependent node F and E from knowledge index table, actual forwarding degree k
f=3, k
e=2, F and E is all selected as recommended node, and forward node quantity reaches minimum forwarding degree, selects to terminate.Query messages is forwarded to recommended node.
Be described below by the technique effect of experiment to the search method that the present invention proposes.
Experimental evaluation index: the recall rate (RecallPerTransferringQueryMessage) of average recall rate (AverageRecall), forwarding expense (NumberofTransferringQueryMessage), each query messages.
Average recall rate (AverageRecall): search all ratios meeting search request number of documents of line node in successful number of documents and network, average, this value is weighted mean value, weights coefficient is defined as each inquiry and obtains the ratio that recall rate accounts for all recall rate numerical value summations, and scope between zero and one.
Forward expense (NumberofTransferringQueryMessage): the mean value of the forwarding quantity of query messages in network, this value is weighted mean value.
The recall rate (RecallPerTransferringQueryMessage) of each query messages: recall rate with forward the ratio of expense, this index can the performance of evaluation technical proposal.
Simulating scenes is arranged:
Generate 1000 network nodes, other 4 nodes of each node Stochastic sum keep being bi-directionally connected, and each node has eight limits to be connected to other node, and initial network is UNICOM.Generate the vector comprising 1024 subject key words.Generate 32 interest vectors, each interest comprises 32 keywords, and the keyword in interest vector is random selecting from 1024 subject key words.From 32 interest vectors, random selecting one is assigned to a node in network, repeats this process, until each node has an interest vector in network.Generate 3000 documents, each document package is containing 8 keywords, and document keyword is random selecting from 1024 subject key words.All documents are stored into network node.Emulation number of days is 30 days, emulates 2000 every day, emulation total degree 60000 times.Every day, (2000 times) were for experimental evaluation index calculate mean value.Each emulation Stochastic choice network node produces query messages, and query messages hop count TTL is 3.
Fig. 4, Fig. 5, Fig. 6 are simulation result.As seen from Figure 4, the average recall rate of technical scheme IESLP that the present invention proposes is higher than ESLP and NeuroGrid.As seen from Figure 5, the forwarding expense of IESLP of the present invention is lower than ESLP, higher than NeuroGrid.As seen from Figure 6, the recall rate that each query messages of IESLP of the present invention obtains is higher than ESLP and NeuroGrid.
In sum, the IESLP technical scheme performance that the present invention proposes is better than ESLP and NeuroGrid.
The above is only for describing technical scheme of the present invention and specific embodiment; the protection domain be not intended to limit the present invention; be to be understood that; under the prerequisite without prejudice to flesh and blood of the present invention and spirit, those skilled in the art change, improve or be equal to replacement etc. all will fall within the scope of protection of the present invention.
Claims (9)
1. knowledge based learns a reciprocity social networks document retrieval method, it is characterized in that, comprises the steps:
Steps A, node sets up interest index and knowledge index, comprise: in file retrieval process, node obtains the similar knowledge store of interest local interest concordance list from the destination node that interest is identical, simultaneously obtains the knowledge store of mating with searching keyword in Indigenous knowledge concordance list according to searching keyword;
Step B, node obtains neighbor node as recommended node by local interest index and knowledge index, and to recommended node forwarding inquiries message, comprise: when node needs to forward the query messages that other nodes send over, inquire about local interest concordance list and knowledge index table and obtain the neighboring node list comprising matching inquiry keyword document, according to the correlation coefficient of neighbor node in coupling number of documents calculations list, and calculate self-adaptation forwarding degree in conjunction with minimum forwarding degree and max-forwards degree, then the self-adaptation forwarding degree according to neighbor node in list selects recommended node to carry out query messages forwarding.
2. a kind of knowledge based learns reciprocity social networks document retrieval method according to claim 1, it is characterized in that, the process that described steps A sets up interest index and knowledge index is as follows:
Step 1), when query node is identical with destination node interest, the number of documents that the interest vector keyword of query node foundation destination node feedback and destination node comprise interest vector keyword is set up or upgrades local interest concordance list;
Step 2), when query node is different with destination node interest, or query node is identical with destination node interest and destination node interest vector lists of keywords does not comprise searching keyword time, the number of documents that query node comprises searching keyword according to searching keyword and the destination node of destination node feedback is set up or upgrades Indigenous knowledge concordance list;
Step 3), destination node is according to searching keyword statistical match number of documents.
3. a kind of knowledge based learns reciprocity social networks document retrieval method according to claim 1, it is characterized in that, in described step B using neighbor node as the decision method of recommended node be: if the self-adaptation of neighbor node forwards number of degrees value and is greater than the quantity selecting recommended node in list, then this neighbor node is chosen as recommended node.
4. a kind of knowledge based learns reciprocity social networks document retrieval method according to claim 1, it is characterized in that, the process that described step B selection recommended node carries out query messages forwarding is as follows:
Step 1) be included in the interest vector of node when searching keyword, then from interest concordance list, node is selected to add in recommended node list according to searching keyword, when recommendation list interior joint lazy weight max-forwards number of degrees value, then from knowledge index table, node is selected to add in recommended node list according to searching keyword, when recommendation list interior joint lazy weight minimum forwarding number of degrees value, then from neighbor list, Stochastic choice residue joint adds in recommended node list;
Step 2) when inquiring about in the interest vector of subject key words not at present node, then from knowledge index table, node is selected to add in recommended node list according to searching keyword, when recommendation list interior joint lazy weight minimum forwarding number of degrees value, then from neighbor list, Stochastic choice residue joint adds in recommended node list;
Step 3) when all not meeting the recommended node of search request in interest concordance list and knowledge index table, expanding hunting zone, increasing the numerical value of minimum forwarding degree; From the connection neighbor node of present node, Stochastic choice meets node that up-to-date minimum forwarding degree requires as recommended node, adds in recommended node list;
4) from recommended node list, node forwarding inquiries message is selected successively, until recommended node list is empty.
5. a kind of knowledge based learns reciprocity social networks document retrieval method according to claim 1, it is characterized in that, calculates correlation coefficient, be specially in described step B:
The correlation coefficient of i-th node obtained from interest concordance list or knowledge index table according to searching keyword is
wherein i=1,2 ..., n, n are the number of the neighbor node that foundation searching keyword obtains from interest concordance list or knowledge index table, the molecule m of item on the right of equation
ibe the number of documents that i-th node mates with searching keyword, denominator
for the coupling number of documents of node that obtains from interest concordance list or knowledge index table according to searching keyword and.
6. a kind of knowledge based learns reciprocity social networks document retrieval method according to claim 1, it is characterized in that, the value calculating self-adaptation forwarding degree in described step B is:
The self-adaptation forwarding degree k of i-th node
i=Round (r
i λ(d
max-d
min))+d
min, wherein i=1,2 ..., n, r
ibe the correlation coefficient of i-th node, d
minfor minimum forwarding degree, namely minimum by the neighbor node number of selection forwarding inquiries message, d
maxfor max-forwards degree, namely maximum by the node number of selection forwarding inquiries message, λ is index regulatory factor, and scope is between 0 ~ 1, and Round function is bracket function.
7. a kind of knowledge based learns reciprocity social networks document retrieval method according to claim 6, it is characterized in that, described minimum forwarding degree d
min=2, max-forwards degree d
max=3, index regulatory factor λ=0.7.
8. a kind of knowledge based learns reciprocity social networks document retrieval method according to claim 1, it is characterized in that, described interest concordance list and described knowledge index table are the identical Hash table structure of structure, keyword forms the key of Hash table, and the neighboring node list comprising matching inquiry keyword document forms the cryptographic hash of key.
9. a kind of knowledge based learns reciprocity social networks document retrieval method according to claim 8, it is characterized in that, element in described list comprises two territories, and a territory stores information of neighbor nodes, and another territory stores the number of documents that this neighbor node comprises corresponding Keywords matching; Comprise the element of information of neighbor nodes in described list by number of files value Bit-reversed from big to small, the node acquiescence coupling number of files value of Stochastic choice is 0, comes list end.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510955213.4A CN105447188B (en) | 2015-12-17 | 2015-12-17 | A kind of reciprocity social networks document retrieval method of knowledge based study |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510955213.4A CN105447188B (en) | 2015-12-17 | 2015-12-17 | A kind of reciprocity social networks document retrieval method of knowledge based study |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105447188A true CN105447188A (en) | 2016-03-30 |
| CN105447188B CN105447188B (en) | 2018-11-06 |
Family
ID=55557364
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510955213.4A Active CN105447188B (en) | 2015-12-17 | 2015-12-17 | A kind of reciprocity social networks document retrieval method of knowledge based study |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105447188B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106534481A (en) * | 2016-09-28 | 2017-03-22 | 努比亚技术有限公司 | Image or video sharing system and method |
| CN109658277A (en) * | 2018-11-30 | 2019-04-19 | 华南师范大学 | A kind of science social networks friend recommendation method, system and storage medium |
| CN111694919A (en) * | 2020-06-12 | 2020-09-22 | 北京百度网讯科技有限公司 | Method and device for generating information, electronic equipment and computer readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080033905A1 (en) * | 2006-08-05 | 2008-02-07 | Terry Lee Stokes | System and Method for the Capture and Archival of Electronic Communications |
| CN101232415A (en) * | 2007-01-22 | 2008-07-30 | 华为技术有限公司 | Equity network node visit apparatus, method and system |
| CN102929914A (en) * | 2012-09-19 | 2013-02-13 | 浙江大学 | Mobile map service searching method based on P2P (point to point) node scheduling |
-
2015
- 2015-12-17 CN CN201510955213.4A patent/CN105447188B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080033905A1 (en) * | 2006-08-05 | 2008-02-07 | Terry Lee Stokes | System and Method for the Capture and Archival of Electronic Communications |
| CN101232415A (en) * | 2007-01-22 | 2008-07-30 | 华为技术有限公司 | Equity network node visit apparatus, method and system |
| CN102929914A (en) * | 2012-09-19 | 2013-02-13 | 浙江大学 | Mobile map service searching method based on P2P (point to point) node scheduling |
Non-Patent Citations (2)
| Title |
|---|
| MINGJUN XIAO等: "Homingspread: community home-based Multi-copy Routing in Mobile Social Networks", 《IEEE》 * |
| 宫月: "基于节点兴趣的P2P信息搜索机制研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106534481A (en) * | 2016-09-28 | 2017-03-22 | 努比亚技术有限公司 | Image or video sharing system and method |
| CN109658277A (en) * | 2018-11-30 | 2019-04-19 | 华南师范大学 | A kind of science social networks friend recommendation method, system and storage medium |
| CN109658277B (en) * | 2018-11-30 | 2022-12-27 | 华南师范大学 | Academic social network friend recommendation method, system and storage medium |
| CN111694919A (en) * | 2020-06-12 | 2020-09-22 | 北京百度网讯科技有限公司 | Method and device for generating information, electronic equipment and computer readable storage medium |
| CN111694919B (en) * | 2020-06-12 | 2023-07-25 | 北京百度网讯科技有限公司 | Method, device, electronic equipment and computer readable storage medium for generating information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105447188B (en) | 2018-11-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | Leveraging social networks for P2P content-based file sharing in disconnected MANETs | |
| US9237199B2 (en) | Contiguous location-based user networks | |
| JP2012533797A (en) | Method and apparatus for efficiently processing multiple keyword queries over a distributed network | |
| Xiaoqiang et al. | An in-network caching scheme based on betweenness and content popularity prediction in content-centric networking | |
| Palazzi et al. | An overview of opportunistic ad hoc communication in urban scenarios | |
| CN105447188A (en) | Knowledge learning based peer-to-peer social network document retrieval method | |
| Ciobanu et al. | Data dissemination in opportunistic networks | |
| CN102377826B (en) | Method for optimal placement of unpopular resource indexes in peer-to-peer network | |
| CN108810139B (en) | Monte Carlo tree search-assisted wireless caching method | |
| CN103209207B (en) | A kind of ordinary node of peer-to-peer network system is found and maintenance of neighbor nodal method | |
| Hemmati et al. | A comparison of name-based content routing protocols | |
| Li et al. | An efficient mobile social network for enhancing contents sharing over mobile ad-hoc networks | |
| Yeferny et al. | An efficient peer-to-peer semantic overlay network for learning query routing | |
| Chang et al. | Social VoD: A social feature-based P2P system | |
| Bhargavi et al. | A hybrid secure routing scheme for MANETS | |
| Girolami et al. | Discovery of services in smart cities of mobile social users | |
| KR101613688B1 (en) | Method for providing peer to peer social networking service (sns) using triangle relationship among nodes | |
| Khatibi et al. | Dynamic multilevel feedback-based searching strategy in unstructured peer-to-peer systems | |
| Chen et al. | Efficient file search in delay tolerant networks with social content and contact awareness | |
| Mala et al. | Route discovery in cellular networks using soft computing techniques | |
| Sharifkhani et al. | A new metric for comparison of p2p search algorithms | |
| Xu et al. | A search strategy for social resource in decentralized social networks | |
| Nakashima et al. | Scalable tree-based consistency maintenance in heterogeneous P2P file sharing systems | |
| Liu et al. | Double‐layer P2P networks supporting semantic search and keeping scalability | |
| Deng et al. | A social similarity-aware multicast routing protocol in delay tolerant networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |