CN103064887A

CN103064887A - Method and device for recommending information

Info

Publication number: CN103064887A
Application number: CN2012105280796A
Authority: CN
Inventors: 廖莎; 王磊
Original assignee: Huawei Technologies Co Ltd
Current assignee: Hefei Zhonghuan Technology Information Consulting Co ltd
Priority date: 2012-12-10
Filing date: 2012-12-10
Publication date: 2013-04-24
Anticipated expiration: 2032-12-10
Also published as: CN103064887B

Abstract

The invention discloses a method and device for recommending information and belongs to the filed of internet communication. The method includes: receiving search information transmitted by a terminal, obtaining the initial searching results according to the searching information, and obtaining a first feature vector of the initial searching result according to the searching information and the initial searching result; performing parallel computation on the first feature vector of the initial searching result to obtain the hash value corresponding to the first feature vector of the initial searching result; obtaining a first sample vector of the initial searching result according to the harsh value corresponding to the first feature vector of the initial searching result; and obtaining the information to be recommended according to the first sample vector of the initial searching result, and recommending the information to be recommended. By means of the method and device for recommending information, the hash value corresponding to the first feature vector of the initial searching result can be calculated in parallel mode through a plurality of hash functions, the computing time is reduced, and the information recommending efficiency is improved.

Description

A kind of method and apparatus of recommendation information

Technical field

The present invention relates to field of Internet communication, particularly a kind of method and apparatus of recommendation information.

Background technology

Fast development along with Internet technology, increasing user obtains the information that needs by the internet, when canned data quantity is more in the internet, the user obtains the difficulty of effective information and understands increasing from the internet, therefore, the method by the recommended engine recommendation information has been subject to paying close attention to widely.

At present, detailed process by the recommended engine recommendation information is: the preliminary Search Results of search acquisition of information that recommended engine sends according to terminal, extract the feature that this preliminary Search Results comprises, the feature that this preliminary Search Results comprises is carried out repeatedly cluster computing, obtain the classification that feature that this preliminary Search Results comprises belongs to, the classification that belongs to according to this feature is obtained sample characteristics corresponding to this classification from the corresponding relation of the classification of having stored and sample characteristics, the information that the sample characteristics that obtains is corresponding is recommended information to be recommended as information to be recommended.

In realizing process of the present invention, the inventor finds that there is following problem at least in prior art:

Prior art is carried out repeatedly cluster computing to the feature that preliminary Search Results comprises, cause the complexity of computing higher, and when the volumes of searches of information was larger, recommended engine may run into performance bottleneck and information to be recommended can not be provided in real time, causes the recommendation inefficiency of information.

Summary of the invention

In order to solve the problem of prior art, the embodiment of the invention provides a kind of method and apparatus of recommendation information.Described technical scheme is as follows:

First aspect, a kind of method of recommendation information, described method comprises:

The search information that receiving terminal sends according to the preliminary Search Results of described search acquisition of information, and is obtained the first eigenvector of described preliminary Search Results according to described search information and described preliminary Search Results;

First eigenvector to described preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of described preliminary Search Results;

The cryptographic hash corresponding according to the first eigenvector of described preliminary Search Results obtained the first sample vector of described preliminary Search Results;

The first sample vector according to described preliminary Search Results obtains information to be recommended, recommends described information to be recommended.

In the possible implementation of the first of above-mentioned first aspect, the search information that described receiving terminal sends, according to the preliminary Search Results of described search acquisition of information, and obtain the first eigenvector of described preliminary Search Results according to described search information and described preliminary Search Results, comprising:

According to search acquisition of information preliminary Search Results, and obtain the first primitive attribute value of the First Characteristic that the First Characteristic set of described preliminary Search Results comprises and the second primitive attribute value of the Second Characteristic that described search information comprises;

First Characteristic according to described preliminary Search Results is gathered the first primitive attribute value of the First Characteristic that comprises and the second primitive attribute value of the Second Characteristic that described search information comprises, the First Characteristic of described preliminary Search Results is gathered the First Characteristic that comprises filter, obtain the Second Characteristic set of described preliminary Search Results;

The first primitive attribute value of the First Characteristic that set comprises according to the Second Characteristic of described preliminary Search Results consists of the first eigenvector of described preliminary Search Results.

In the possible implementation of the second of above-mentioned first aspect, described First Characteristic according to described preliminary Search Results is gathered the first primitive attribute value of the First Characteristic that comprises and the second primitive attribute value of the Second Characteristic that described search information comprises, the First Characteristic of described preliminary Search Results is gathered the First Characteristic that comprises to be filtered, obtain the Second Characteristic set of described preliminary Search Results, comprising:

Calculate the difference between the second primitive attribute value of the first primitive attribute value of the First Characteristic that the First Characteristic set of described preliminary Search Results comprises and the Second Characteristic that described search information comprises;

In the set of the First Characteristic of described preliminary Search Results, filter out and the second primitive attribute value of the Second Characteristic that described search information comprises between difference greater than the First Characteristic of the first default value, obtain the Second Characteristic set of described preliminary Search Results.

In the third possible implementation of above-mentioned first aspect, the first primitive attribute value of the First Characteristic that described Second Characteristic set according to described preliminary Search Results comprises consists of the first eigenvector of described preliminary Search Results, comprising:

Determine the interval that the first primitive attribute value of the First Characteristic that the Second Characteristic set of described preliminary Search Results comprises belongs to;

The interval that the first primitive attribute value of the First Characteristic that set comprises according to the Second Characteristic of described preliminary Search Results belongs to obtains the First Eigenvalue corresponding to First Characteristic that the Second Characteristic set of described preliminary Search Results comprises;

Second Characteristic according to described preliminary Search Results is gathered the First Eigenvalue corresponding to First Characteristic that comprises, consists of the first eigenvector of described preliminary Search Results.

In the 4th kind of possible implementation of above-mentioned first aspect, described first eigenvector to described preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of described preliminary Search Results, comprising:

The First Eigenvalue that the first eigenvector of described preliminary Search Results comprises is encoded concurrently, obtain one-dimensional coding corresponding to described preliminary Search Results;

The one dimension corresponding to described preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of described preliminary Search Results.

In the 5th kind of possible implementation of above-mentioned first aspect, the described the First Eigenvalue that the first eigenvector of described preliminary Search Results is comprised is encoded concurrently, obtains one-dimensional coding corresponding to described preliminary Search Results, comprising:

The First Eigenvalue that the first eigenvector of described preliminary Search Results comprises is encoded concurrently, obtain the coding of the First Eigenvalue that the first eigenvector of described preliminary Search Results comprises;

The coding of the First Eigenvalue that the first eigenvector of described preliminary Search Results is comprised is connected, and obtains one-dimensional coding corresponding to described preliminary Search Results.

In the 6th kind of possible implementation of above-mentioned first aspect, the described cryptographic hash corresponding according to the first eigenvector of described preliminary Search Results obtained the first sample vector of described preliminary Search Results, comprising:

The cryptographic hash corresponding according to the first eigenvector of described preliminary Search Results obtained the second sample vector from the corresponding relation of the cryptographic hash of having stored and the second sample vector;

Calculate the first eigenvector of described preliminary Search Results and the Hamming distance between described the second sample vector;

Select individual second sample vector of the second default value of the Hamming distance minimum between the first eigenvector with described preliminary Search Results, and the second sample vector of described selection is defined as the first sample vector of described preliminary Search Results.

In the 7th kind of possible implementation of above-mentioned first aspect, the described information to be recommended of described recommendation comprises:

The second primitive attribute value of the Second Characteristic that comprises according to described search information is obtained Second Eigenvalue corresponding to described Second Characteristic, and the Second Eigenvalue of the Second Characteristic that comprises according to described search information consists of the Second Characteristic vector of described search information;

Calculate the distance between first sample vector of the described preliminary Search Results of Second Characteristic vector sum of described search information;

According to described distance described information to be recommended is sorted, and according to the described information to be recommended of Sequential output after the ordering.

Second aspect, a kind of equipment of recommendation information, described equipment comprises:

The first acquisition module is used for the search information that receiving terminal sends, and according to the preliminary Search Results of described search acquisition of information, and obtains the first eigenvector of described preliminary Search Results according to described search information and described preliminary Search Results;

Computing module, the first eigenvector that is used for described preliminary Search Results that described the first acquisition module is obtained carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of described preliminary Search Results;

The second acquisition module for the cryptographic hash corresponding to first eigenvector of the described preliminary Search Results that obtains according to described computing module, obtains the first sample vector of described preliminary Search Results;

Recommending module, the first sample vector that is used for the described preliminary Search Results that obtains according to described the second acquisition module obtains information to be recommended, recommends described information to be recommended.

In the possible implementation of the first of above-mentioned second aspect, described the first acquisition module comprises:

The first acquiring unit is used for according to the preliminary Search Results of search acquisition of information, and obtains the first primitive attribute value of the First Characteristic that the First Characteristic set of described preliminary Search Results comprises and the second primitive attribute value of the Second Characteristic that described search information comprises;

Filter element, First Characteristic for the described preliminary Search Results that obtains according to described the first acquiring unit is gathered the first primitive attribute value of the First Characteristic that comprises and the second primitive attribute value of the Second Characteristic that described search information comprises, the First Characteristic of described preliminary Search Results is gathered the First Characteristic that comprises filter, obtain the Second Characteristic set of described preliminary Search Results;

Component units for the first primitive attribute value of gathering the First Characteristic that comprises according to the Second Characteristic of described preliminary Search Results, consists of the first eigenvector of described preliminary Search Results.

In the possible implementation of the second of above-mentioned second aspect, described filter element comprises:

Computation subunit is used for calculating the difference between the second primitive attribute value of the first primitive attribute value of the First Characteristic that the First Characteristic set of described preliminary Search Results comprises and the Second Characteristic that described search information comprises;

Filter subelement, be used for that First Characteristic set at described preliminary Search Results filters out and the second primitive attribute value of the Second Characteristic that described search information comprises between difference greater than the First Characteristic of the first default value, obtain the Second Characteristic set of described preliminary Search Results.

In the third possible implementation of above-mentioned second aspect, described Component units comprises:

Determine subelement, be used for to determine the interval that the first primitive attribute value of the First Characteristic that the Second Characteristic set of described preliminary Search Results comprises belongs to;

Obtain subelement, be used for the First Eigenvalue corresponding to First Characteristic that Second Characteristic set that interval that the first primitive attribute value of the First Characteristic that the Second Characteristic set of the described preliminary Search Results determined according to described definite subelement comprises belongs to obtains described preliminary Search Results comprises;

Consist of subelement, be used for the First Eigenvalue corresponding to First Characteristic that comprise according to described Second Characteristic set of obtaining the described preliminary Search Results that subelement obtains, consist of the first eigenvector of described preliminary Search Results.

In the 4th kind of possible implementation of above-mentioned second aspect, described computing module comprises:

Coding unit is encoded concurrently for the First Eigenvalue that the first eigenvector to described preliminary Search Results comprises, obtains one-dimensional coding corresponding to described preliminary Search Results;

Arithmetic element, one-dimensional coding corresponding to described preliminary Search Results that is used for described coding unit is obtained carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of described preliminary Search Results.

In the 5th kind of possible implementation of above-mentioned second aspect, described coding unit comprises:

The coding subelement is used for the First Eigenvalue that the first eigenvector to described preliminary Search Results comprises and encodes concurrently, obtains the coding of the First Eigenvalue that the first eigenvector of described preliminary Search Results comprises;

The sub-series unit, the coding of the First Eigenvalue that the first eigenvector that is used for described coding subelement is obtained described preliminary Search Results comprises is connected, and obtains one-dimensional coding corresponding to described preliminary Search Results.

In the 6th kind of possible implementation of above-mentioned second aspect, described the second acquisition module comprises:

Second acquisition unit is used for the cryptographic hash corresponding according to the first eigenvector of described preliminary Search Results, obtains the second sample vector from the corresponding relation of the cryptographic hash of having stored and the second sample vector;

The first computing unit is used for calculating the first eigenvector of described preliminary Search Results and the Hamming distance between described the second sample vector;

Selected cell, be used for to select individual second sample vector of the second default value of the Hamming distance minimum between the first eigenvector with described preliminary Search Results, and the second sample vector of described selection is defined as the first sample vector of described preliminary Search Results.

In the 7th kind of possible implementation of above-mentioned second aspect, described recommending module comprises:

The 3rd acquiring unit, the first sample vector that is used for the described preliminary Search Results that obtains according to described the second acquisition module obtains information to be recommended;

The 4th acquiring unit, the second primitive attribute value for the Second Characteristic that comprises according to described search information, obtain Second Eigenvalue corresponding to described Second Characteristic, the Second Eigenvalue of the Second Characteristic that comprises according to described search information consists of the Second Characteristic vector of described search information;

The second computing unit is for the distance between the first sample vector of the described preliminary Search Results of Second Characteristic vector sum that calculates described search information;

Sequencing unit is used for according to the described distance that described the second computing unit calculates described information to be recommended being sorted, and according to the described information to be recommended of Sequential output after the ordering.

In embodiments of the present invention, the search information that receiving terminal sends is searched for the preliminary Search Results of acquisition of information according to this, and obtains the first eigenvector of this Search Results according to search information and preliminary Search Results; First eigenvector to this preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of this preliminary Search Results; The cryptographic hash corresponding according to the first eigenvector of this preliminary Search Results obtained the first sample vector of this preliminary Search Results; The first sample vector according to this preliminary Search Results that obtains obtains information to be recommended, recommends information to be recommended.Wherein, only the first eigenvector of this preliminary Search Results carried out a concurrent operation, obtain the cryptographic hash corresponding to first eigenvector of this preliminary Search Results, so the computational complexity of the recommendation information method that the embodiment of the invention provides carries out the repeatedly complexity of cluster computing less than the feature that this preliminary Search Results is comprised, reduce the complexity of computing and the speed of accelerating recommendation information, thereby improved the efficient of recommendation information.

Description of drawings

Fig. 1 is the method flow diagram of a kind of recommendation information of providing of the embodiment of the invention;

Fig. 2 is the method flow diagram of the another kind of recommendation information that provides of the embodiment of the invention;

Fig. 3 is the device structure synoptic diagram of a kind of recommendation information of providing of the embodiment of the invention;

Fig. 4 is the device structure synoptic diagram of the another kind of recommendation information that provides of the embodiment of the invention.

Embodiment

For making the purpose, technical solutions and advantages of the present invention clearer, embodiment of the present invention is described further in detail below in conjunction with accompanying drawing.

The embodiment of the invention provides a kind of method of recommendation information, and referring to Fig. 1, the method comprises:

Step 101: the search information that receiving terminal sends, search for the preliminary Search Results of acquisition of information according to this, and obtain the first eigenvector of this preliminary Search Results according to search information and preliminary Search Results;

Step 102: the first eigenvector to this preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of this preliminary Search Results;

Step 103: the cryptographic hash corresponding according to the first eigenvector of this preliminary Search Results, obtain the first sample vector of this preliminary Search Results;

Step 104: the first sample vector according to this preliminary Search Results obtains information to be recommended, recommends information to be recommended.

In embodiments of the present invention, the search information that receiving terminal sends is searched for the preliminary Search Results of acquisition of information according to this, and obtains the first eigenvector of this Search Results according to search information and preliminary Search Results; First eigenvector to this preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of this preliminary Search Results; The cryptographic hash corresponding according to the first eigenvector of this preliminary Search Results obtained the first sample vector of this preliminary Search Results; The first sample vector according to this preliminary Search Results that obtains obtains information to be recommended, recommends information to be recommended.Wherein, only the first eigenvector of this preliminary Search Results carried out a concurrent operation, obtain the cryptographic hash corresponding to first eigenvector of this preliminary Search Results, so the computational complexity of the method for the recommendation information that the embodiment of the invention provides carries out the repeatedly complexity of cluster computing less than the feature that this preliminary Search Results is comprised, reduce the complexity of computing and the speed of accelerating recommendation information, thereby improved the efficient of recommendation information.

Another embodiment of the present invention provides a kind of method of recommendation information, executive agent is the hardware chips such as field programmable gate array (Field-Programmable Gate Array, FPGA) or special IC (Application Specific Integrated Circuit:ASIC) among this embodiment.Referring to Fig. 2, the method comprises:

Step 201: the search information that receiving terminal sends, according to the preliminary Search Results of this search acquisition of information, and the First Characteristic set of obtaining this preliminary Search Results according to this search information and the preliminary Search Results that obtains, comprise the First Characteristic that this preliminary Search Results comprises in the First Characteristic set of this preliminary Search Results, and extract the Second Characteristic that search information comprises;

Particularly, terminal receives the search information of user's input, and should search information send to cloud server, the preliminary Search Results of search acquisition of information that cloud server sends according to terminal, extract the First Characteristic that this preliminary Search Results comprises, the First Characteristic that comprises according to this preliminary Search Results that extracts consists of the First Characteristic set of this preliminary Search Results, and extracts the Second Characteristic that this search information comprises from the search information that terminal sends.

Wherein, extracting the First Characteristic that this preliminary Search Results comprises is: extract the keyword in this preliminary Search Results, the keyword that extracts is defined as the First Characteristic that this preliminary Search Results comprises.

Wherein, extracting the First Characteristic that this search information comprises is: extract the keyword of search information, the keyword that extracts is defined as the Second Characteristic that this search information comprises.

Step 202: the Second Characteristic that comprises according to search information and the First Characteristic of this preliminary Search Results are gathered, and obtain the first eigenvector of this preliminary Search Results;

Particularly, the Second Characteristic that comprises according to search information obtains the second primitive attribute value of the Second Characteristic that this search information comprises, and the First Characteristic that set comprises according to the First Characteristic of this preliminary Search Results obtains the first primitive attribute value of First Characteristic, calculate the difference of the second primitive attribute value and the first primitive attribute value, in the set of the First Characteristic of this preliminary Search Results, filter out and the second primitive attribute value between difference greater than the First Characteristic of the first default value, obtain the Second Characteristic set of this preliminary Search Results, determine the interval that the first primitive attribute value of the First Characteristic that the Second Characteristic set of this preliminary Search Results comprises belongs to, according to this interval, obtain the Second Characteristic of this preliminary Search Results and gather the First Eigenvalue corresponding to First Characteristic that comprises, and the First Eigenvalue corresponding to First Characteristic that set comprises according to the Second Characteristic of this preliminary Search Results, consist of the first eigenvector of this preliminary Search Results.

Wherein, the second primitive attribute value that the Second Characteristic that comprises according to search information obtains the Second Characteristic that this search information comprises can be specially: the Second Characteristic that comprises according to this search information, judge the classification that this Second Characteristic belongs to, the classification that belongs to according to this Second Characteristic is obtained the second primitive attribute value of this Second Characteristic from the corresponding relation of the classification of having stored and primitive attribute value.

Wherein, the first primitive attribute value that the First Characteristic that set comprises according to the First Characteristic of this preliminary Search Results obtains First Characteristic is specially: the First Characteristic that comprises according to this search information, judge the classification that this First Characteristic belongs to, the classification that belongs to according to this First Characteristic is obtained the first primitive attribute value of this First Characteristic from the corresponding relation of the classification of having stored and primitive attribute value.

Wherein, comprise the First Eigenvalue corresponding to First Characteristic that the Second Characteristic set of this preliminary Search Results comprises in the first eigenvector of preliminary Search Results, can from the corresponding relation of the cryptographic hash of having stored and the second sample vector, obtain according to the first eigenvector of this preliminary Search Results the second sample vector of this preliminary Search Results.

Wherein, divide in advance a plurality of intervals, an interval corresponding the First Eigenvalue, between the memory block and the corresponding relation of the First Eigenvalue, and the First Eigenvalue of adjacent interval to differ be 1,2 or 3 etc.

Wherein, the process of determining the interval that the first primitive attribute value of the First Characteristic that the Second Characteristic set of this preliminary Search Results comprises belongs to is specially: the first primitive attribute value of the First Characteristic that the Second Characteristic set of preliminary Search Results is comprised by comparer and the interval of dividing in advance compare, and the Second Characteristic of determining this preliminary Search Results is gathered the interval that the first primitive attribute value of the First Characteristic that comprises belongs to.

Wherein, adopt the first primitive attribute value of a plurality of First Characteristics that Second Characteristic set that a plurality of comparers concurrently will this preliminary Search Results comprises and the interval of in advance division to compare, determine the interval that the first primitive attribute value of a plurality of First Characteristics that the Second Characteristic set of this preliminary Search Results comprises belongs to, so, the first primitive attribute value of the Second Characteristic of this preliminary Search Results being gathered a plurality of First Characteristics that comprise compares simultaneously, can reduce the time of comparison.

Wherein, the interval that the first primitive attribute value of the First Characteristic that set comprises according to the Second Characteristic of this preliminary Search Results belongs to, the process of obtaining the First Eigenvalue corresponding to First Characteristic that the Second Characteristic set of this preliminary Search Results comprises is specially: the interval that the first primitive attribute value of the First Characteristic that comprises according to the Second Characteristic set of this preliminary Search Results belongs to, obtain the First Eigenvalue corresponding to First Characteristic that the Second Characteristic set of this preliminary Search Results comprises from the corresponding relation of the interval of having stored and the First Eigenvalue.

For example, the interval of dividing in advance is (3,7,29), the interval that the corresponding relation of the First Eigenvalue that each interval is corresponding with this interval is as shown in table 1 below and the corresponding relation of the First Eigenvalue; If the set of the Second Characteristic of preliminary Search Results 1 comprises First Characteristic 1, First Characteristic 2 and First Characteristic 3, the first primitive attribute value of First Characteristic 1 is 5, and the first primitive attribute value of First Characteristic 2 is 30, and the first primitive attribute value of First Characteristic 3 is 10; Determine that the first primitive attribute value 5 of First Characteristic 1 belongs to the interval of 3＜T≤7, then the First Eigenvalue of First Characteristic 1 correspondence is 1; The first primitive attribute value 30 of determining First Characteristic 2 belongs to T〉29 interval, then the First Eigenvalue of First Characteristic 2 correspondences is 3; Determine that the first primitive attribute value 10 of First Characteristic 3 belongs to the interval of 7＜T≤29, then the First Eigenvalue of First Characteristic 3 correspondences is 2, according to the First Eigenvalue 1 of First Characteristic 1 correspondence, the First Eigenvalue 3 of First Characteristic 2 correspondences and the First Eigenvalue 2 of First Characteristic 3 correspondences, consist of the first eigenvector (1 of preliminary Search Results 1 correspondence, 3,2).

Table 1

Interval	The First Eigenvalue
		T≤3	0
3<T≤7	1
		7<T≤29	2
T>29	3

Step 203: the First Eigenvalue that the first eigenvector of this preliminary Search Results comprises is encoded concurrently, obtain one-dimensional coding corresponding to this preliminary Search Results;

Particularly, the First Eigenvalue that the first eigenvector of this preliminary Search Results comprises is encoded concurrently, obtain the coding of the First Eigenvalue that the first eigenvector of this preliminary Search Results comprises, the coding of the First Eigenvalue that obtains is connected, obtain one-dimensional coding corresponding to this preliminary Search Results.

Wherein, the first eigenvector of this preliminary Search Results is a multi-C vector.

Wherein, the First Eigenvalue that in embodiments of the present invention first eigenvector of this preliminary Search Results is comprised is encoded, and makes Hamming distance and the difference this two the First Eigenvalues between of two the First Eigenvalues after coding equal.

For example, obtain 4 preliminary Search Results according to search information, be respectively preliminary Search Results 1, preliminary Search Results 2, preliminary Search Results 3 and preliminary Search Results 4, the first eigenvector of the three-dimensional of preliminary Search Results 1 correspondence is P1(1,3,2), the first eigenvector of the three-dimensional of preliminary Search Results 2 correspondences is P2(1,2,4), the first eigenvector of the three-dimensional of preliminary Search Results 3 correspondences is P3(2,1,3) and the first eigenvector of the three-dimensional of preliminary Search Results 4 correspondences be P4(4,2,3), if the maximal value of the First Eigenvalue is 4 in this tri-vector space, then with first eigenvector P1(1,3,0) the First Eigenvalue 1 in is encoded to 1000, and the First Eigenvalue 3 is encoded to 1110, the First Eigenvalue 0 is encoded to 0000, the First Eigenvalue 2 is encoded to 1100, and the coding of the First Eigenvalue that first eigenvector P1 is comprised is connected, and the one-dimensional coding that obtains preliminary Search Results 1 correspondence is 100011101100; In like manner with first eigenvector P2(1,2,4) the First Eigenvalue 1 in is encoded to 1000, the First Eigenvalue 2 is encoded to 1100, the First Eigenvalue 4 is encoded to 1111, the coding of the First Eigenvalue that first eigenvector P2 is comprised is connected, and the one-dimensional coding that obtains preliminary Search Results 2 correspondences is 100011001111; In like manner with first eigenvector P3(2,1,3) the First Eigenvalue 2 in is encoded to 1100, the First Eigenvalue 1 is encoded to 1000, the First Eigenvalue 3 is encoded to 1110, the coding of the First Eigenvalue that first eigenvector P3 is comprised is connected, and the one-dimensional coding that obtains preliminary Search Results 3 correspondences is 110010001110; In like manner with first eigenvector P4(4,2,3) the First Eigenvalue 4 in is encoded to 1111, the First Eigenvalue 2 is encoded to 1100, the First Eigenvalue 3 is encoded to 1110, the coding of the First Eigenvalue that first eigenvector P4 is comprised is connected, and the one-dimensional coding that obtains preliminary Search Results 4 correspondences is 111111001110.

Step 204: the one-dimensional coding corresponding according to this preliminary Search Results, calculate the cryptographic hash corresponding to first eigenvector of this preliminary Search Results;

Particularly, the one-dimensional coding that this preliminary Search Results is corresponding carries out computing by default hash function, obtain the cryptographic hash corresponding to first eigenvector of this preliminary Search Results, default hash function comprises location index at least, and the maximal value of location index is less than or equal to the length of one-dimensional coding corresponding to this preliminary Search Results.

Wherein, if default hash function has N, the one-dimensional coding that this preliminary Search Results is corresponding carries out respectively computing by N default hash function, obtains N the cryptographic hash corresponding to first eigenvector of this preliminary Search Results.

Wherein, the one-dimensional coding that this preliminary Search Results is corresponding by the detailed process that hash function carries out computing is: the location index that comprises according to default hash function, obtain the bit value of one-dimensional coding correspondence position corresponding to this preliminary Search Results, the bit value that obtains is connected, obtain the cryptographic hash corresponding to first eigenvector of this preliminary Search Results.

Wherein, in the present embodiment, a plurality of hash functions are set in advance, when cryptographic hash corresponding to the first eigenvector that calculates preliminary Search Results, make a plurality of hash functions of prior setting carry out simultaneously computing, reduced the time of computing, and then improved the efficient of recommendation information.

For example, 4 hash functions are set in advance, the location index that hash function 1 comprises is H1(1,5,7,8), the location index that hash function 2 comprises is H2(1,3,5,9), the location index that hash function 3 comprises is H3(2,5,9,10), the location index that hash function 4 comprises is H4(2,3,6,8), for preliminary Search Results 1, be 1110 with the one-dimensional coding 100011101100 of preliminary Search Results 1 correspondence by the first cryptographic hash corresponding to first eigenvector that location index H1 obtains preliminary Search Results 1, be 1011 with the one-dimensional coding 100011101100 of preliminary Search Results 1 correspondence by the second cryptographic hash corresponding to first eigenvector that location index H2 obtains preliminary Search Results 1, being 0111 with the one-dimensional coding 100011101100 of preliminary Search Results 1 correspondence by the 3rd cryptographic hash corresponding to first eigenvector that location index H3 obtains preliminary Search Results 1, is 0010 with the one-dimensional coding 100011101100 of preliminary Search Results 1 correspondence by the 4th cryptographic hash corresponding to first eigenvector that location index H4 obtains preliminary Search Results 1.

Step 205: the cryptographic hash corresponding according to the first eigenvector of this preliminary Search Results, from the corresponding relation of the cryptographic hash of having stored and the second sample vector, obtain the second sample vector of this preliminary Search Results;

Wherein, all Search Results characteristics of correspondence to be searched that in advance extraction has been stored, feature characteristic of correspondence value according to this Search Results consists of the second sample vector, this second sample vector is carried out Hash operation, obtain cryptographic hash corresponding to this second sample vector, and the cryptographic hash that this second sample vector is corresponding is stored in the corresponding relation of cryptographic hash and the second sample vector with this second sample vector.

Wherein, a storage array is set in advance, the corresponding relation of cryptographic hash and the second sample vector is stored in this storage array, when obtaining the second sample vector of this preliminary Search Results, can be according to the first eigenvector of this preliminary Search Results corresponding a plurality of cryptographic hash, this storage array is carried out concurrent access, thereby obtain the second sample vector of this preliminary Search Results, so, this storage array is carried out concurrent access can reduce access time to storage array, improve the efficient of recommendation information.

Step 206: calculate respectively the first eigenvector of this preliminary Search Results and the Hamming distance between the second sample vector;

For example, first cryptographic hash 1110 corresponding according to the first eigenvector P1 of preliminary Search Results 1, the second sample vector of the preliminary Search Results 1 that obtains is A1(1,3,2), A2(1,3,3) and A3(1,3,1), calculate first eigenvector P1(1,3,2) and the second sample vector A1(1,3,2) Hamming distance between is 0, first eigenvector P1(1,3,2) with the second sample vector A2(1,3,4) Hamming distance between is 1, first eigenvector P1(1,3,2) with the second sample vector A3(1,3,1) between Hamming distance be 1; Second cryptographic hash 1011 corresponding according to the first eigenvector P1 of preliminary Search Results 1, the second sample vector of the preliminary Search Results 1 that obtains is B1(1,3,4), B2(2,3,2) and B3(1,3,4), calculate first eigenvector P1(1,3,2) with the second sample vector B1(1,3,4) between Hamming distance be 2, first eigenvector P1(1,3,2) with the second sample vector B2(2,3,2) Hamming distance between is 1, first eigenvector P1(1,3,2) and the second sample vector B3(1,3,4) Hamming distance between is 2; Three cryptographic hash 0111 corresponding according to the first eigenvector P1 of preliminary Search Results 1, the second sample vector of the preliminary Search Results 1 that obtains is C1(1,3,4), C2(2,3,3) and C3(2,4,1), calculate first eigenvector P1(1,3,2) with the second sample vector C1(1,2,3) between Hamming distance be 2, first eigenvector P1(1,3,2) with the second sample vector C2(1,4,3) Hamming distance between is 2, first eigenvector P1(1,3,2) and the second sample vector C3(1,1,2) Hamming distance between is 2; Four cryptographic hash 0010 corresponding according to the first eigenvector P1 of preliminary Search Results 1, the second sample vector of the preliminary Search Results 1 that obtains is D1(1,2,4), D2(1,3,3) and D3(1,4,1), calculate first eigenvector P1(1,3,2) with the second sample vector D1(1,2,4) between Hamming distance be 3, first eigenvector P1(1,3,2) with the second sample vector D2(1,3,3) Hamming distance between is 1, first eigenvector P1(1,3,2) and the second sample vector D3(1,4,1) Hamming distance between is 2.

Step 207: select individual second sample vector of the second default value of the Hamming distance minimum between the first eigenvector with this preliminary Search Results, the second sample vector of selection is defined as the first sample vector of this preliminary Search Results;

For example, if the second default value is 5, select the first eigenvector P1(1 with preliminary Search Results 1,3,5 the second sample vectors of the Hamming distance minimum 2) are A1, A2, A3, B2 and D2, and the second sample vector A1, A2, A3, B2 and the D2 that then will select is defined as the first sample vector of this preliminary Search Results 1.

Step 208: the first sample vector according to this preliminary Search Results obtains information to be recommended, recommends this information to be recommended.

Particularly, according to the first sample vector of this preliminary Search Results, from the corresponding relation of the sample vector stored and information, obtain corresponding information, the information of obtaining is defined as information to be recommended, and recommends information to be recommended.

Alternatively, the first sample vector of this preliminary Search Results and the first eigenvector of this preliminary Search Results are compared, information corresponding to proper vector that will be different from this preliminary Search Results is defined as information to be recommended.

Preferably, obtain after the information to be recommended according to the first sample vector of this preliminary Search Results, distance between Second Characteristic vector sum first sample vector of calculating search information, according to and the Second Characteristic vector of search information between the order of distance, information to be recommended corresponding to the first sample properties vector is sorted, according to sequence of information output information to be recommended to be recommended.

Preferably, when information to be recommended is sorted, can according to the distance order from small to large between the Second Characteristic vector of the first sample vector corresponding to information to be recommended and search information, information to be recommended be sorted.

In the embodiment of the invention, the search information that receiving terminal sends is searched for the preliminary Search Results of acquisition of information according to this, and obtains the first eigenvector of this Search Results according to search information and preliminary Search Results; First eigenvector to this preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of this preliminary Search Results; The cryptographic hash corresponding according to the first eigenvector of this preliminary Search Results obtained the first sample vector of this preliminary Search Results; The first sample vector according to this preliminary Search Results that obtains obtains information to be recommended, recommends information to be recommended.Wherein, only the first eigenvector of this preliminary Search Results carried out a concurrent operation, obtain the cryptographic hash corresponding to first eigenvector of this preliminary Search Results, so the computational complexity of the recommendation information method that the embodiment of the invention provides carries out the repeatedly complexity of cluster computing less than the feature that this preliminary Search Results is comprised, reduce the complexity of computing and the speed of accelerating recommendation information, thereby improved the efficient of recommendation information.

The embodiment of the invention provides a kind of equipment of recommendation information, and the equipment of this recommendation information can be FPGA or asic chip specifically, and referring to Fig. 3, this equipment comprises:

The first acquisition module 301 is used for the search information that receiving terminal sends, and searches for the preliminary Search Results of acquisition of information according to this, and obtains the first eigenvector of this preliminary Search Results according to search information and the preliminary Search Results that obtains;

Computing module 302, the first eigenvector that is used for the preliminary Search Results that obtains according to the first acquisition module 301 obtains the cryptographic hash corresponding to first eigenvector of this preliminary Search Results;

The second acquisition module 303 for the cryptographic hash corresponding to first eigenvector of this preliminary Search Results that obtains according to computing module 302, obtains the first sample vector of this preliminary Search Results;

Recommending module 304, the first sample vector that is used for this preliminary Search Results of obtaining according to the second acquisition module 303 obtains information to be recommended, recommends information to be recommended.

Wherein, the first acquisition module 301 comprises:

The first acquiring unit is used for according to the preliminary Search Results of search acquisition of information, and obtains the first primitive attribute value of the First Characteristic that the First Characteristic set of this preliminary Search Results comprises and the second primitive attribute value of the Second Characteristic that this search information comprises;

Filter element, gather the first primitive attribute value of the First Characteristic that comprises and the second primitive attribute value of the Second Characteristic that this search information comprises for the First Characteristic of this preliminary Search Results that obtains according to the first acquiring unit, the First Characteristic of this preliminary Search Results is gathered the First Characteristic that comprises filter, obtain the Second Characteristic set of this preliminary Search Results;

Component units, the Second Characteristic that is used for this preliminary Search Results of basis is gathered the first primitive attribute value of the First Characteristic that comprises, consists of the first eigenvector of this preliminary Search Results.

Further, filter element comprises:

Computation subunit be used for to be calculated the difference between the second primitive attribute value of the first primitive attribute value of the First Characteristic that the First Characteristic set of this preliminary Search Results comprises and the Second Characteristic that this search information comprises;

Filter subelement, be used for that First Characteristic set at this preliminary Search Results filters out and the second primitive attribute value of the Second Characteristic that this search information comprises between difference greater than the First Characteristic of the first default value, obtain the Second Characteristic set of this preliminary Search Results.

Further, Component units comprises:

Determine subelement, be used for to determine the interval that the first primitive attribute value of the First Characteristic that the Second Characteristic set of this preliminary Search Results comprises belongs to;

Obtain subelement, be used for the Second Characteristic that interval that the first primitive attribute value of the First Characteristic that comprises according to the Second Characteristic set of determining this preliminary Search Results that subelement is determined belongs to obtains this preliminary Search Results and gather the First Eigenvalue corresponding to First Characteristic that comprises;

Consist of subelement, be used for gathering the First Eigenvalue corresponding to First Characteristic that comprises according to the Second Characteristic that obtains this preliminary Search Results that subelement obtains, consist of the first eigenvector of this preliminary Search Results.

Wherein, computing module 302 comprises:

Coding unit is encoded concurrently for the First Eigenvalue that the first eigenvector to this preliminary Search Results comprises, obtains one-dimensional coding corresponding to this preliminary Search Results;

Arithmetic element, one-dimensional coding corresponding to this preliminary Search Results that is used for coding unit is obtained carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of this preliminary Search Results.

Further, coding unit comprises:

The coding subelement is used for the First Eigenvalue that the first eigenvector to this preliminary Search Results comprises and encodes concurrently, obtains the coding of the First Eigenvalue that the first eigenvector of this preliminary Search Results comprises;

Sub-series unit, the subelement that is used for encoding obtain the coding of the First Eigenvalue that the first eigenvector of this preliminary Search Results comprises and connect, and obtain one-dimensional coding corresponding to this preliminary Search Results.

Wherein, the second acquisition module 303 comprises:

Second acquisition unit is used for cryptographic hash corresponding to first eigenvector according to this preliminary Search Results, obtains the second sample vector from the corresponding relation of the cryptographic hash of having stored and the second sample vector;

The first computing unit is used for calculating the first eigenvector of this preliminary Search Results and the Hamming distance between the second sample vector;

Selected cell, for the second default value second sample vector of the Hamming distance minimum between the first eigenvector of selection and this preliminary Search Results, and the second sample vector that will select is defined as the first sample vector of this preliminary Search Results.

Wherein, recommending module 304 comprises:

The 4th acquiring unit is used for the second primitive attribute value of the Second Characteristic that comprises according to search information, obtains Second Eigenvalue corresponding to this Second Characteristic, and it is vectorial that the Second Eigenvalue of the Second Characteristic that comprises according to search information consists of the Second Characteristic of this search information;

The second computing unit is for the distance between the first sample vector of this preliminary Search Results of Second Characteristic vector sum that calculates this search information;

Sequencing unit is used for according to the distance that the second computing unit calculates information to be recommended being sorted, and according to the Sequential output information to be recommended after the ordering.

The embodiment of the invention provides a kind of equipment of recommendation information, and the equipment of this recommendation information can be FPGA or asic chip specifically, and referring to Fig. 4, this equipment comprises: storer 401 and processor 402.

Described storer 401 is used for storing cryptographic hash corresponding to first eigenvector, the described first eigenvector of the described preliminary Search Results that described processor obtains and the first sample vector of described preliminary Search Results.

Described processor 402 is used for the search information that receiving terminal sends, and according to the preliminary Search Results of described search acquisition of information, and obtains the first eigenvector of described preliminary Search Results according to described search information and described preliminary Search Results; First eigenvector to described preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of described preliminary Search Results; The cryptographic hash corresponding according to the first eigenvector of described preliminary Search Results obtained the first sample vector of described preliminary Search Results; The first sample vector according to described preliminary Search Results obtains information to be recommended, recommends described information to be recommended.

Wherein, the search information that described receiving terminal sends according to the preliminary Search Results of described search acquisition of information, and is obtained the first eigenvector of described preliminary Search Results according to described search information and described preliminary Search Results, comprising:

Wherein, described First Characteristic according to described preliminary Search Results is gathered the first primitive attribute value of the First Characteristic that comprises and the second primitive attribute value of the Second Characteristic that described search information comprises, the First Characteristic of described preliminary Search Results is gathered the First Characteristic that comprises to be filtered, obtain the Second Characteristic set of described preliminary Search Results, comprising:

Wherein, the first primitive attribute value of the First Characteristic that described Second Characteristic set according to described preliminary Search Results comprises consists of the first eigenvector of described preliminary Search Results, comprising:

Wherein, described first eigenvector to described preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of described preliminary Search Results, comprising:

The one-dimensional coding corresponding to described preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of described preliminary Search Results.

Wherein, the described the First Eigenvalue that the first eigenvector of described preliminary Search Results is comprised is encoded concurrently, obtains one-dimensional coding corresponding to described preliminary Search Results, comprising:

Wherein, the described cryptographic hash corresponding according to the first eigenvector of described preliminary Search Results obtained the first sample vector of described preliminary Search Results, comprising:

Wherein, the described information to be recommended of described recommendation comprises:

The all or part of step that one of ordinary skill in the art will appreciate that realization above-described embodiment can be finished by hardware, also can come the relevant hardware of instruction to finish by program, described program can be stored in a kind of computer-readable recording medium, the above-mentioned storage medium of mentioning can be ROM (read-only memory), disk or CD etc.

The above only is preferred embodiment of the present invention, and is in order to limit the present invention, within the spirit and principles in the present invention not all, any modification of doing, is equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims

1. the method for a recommendation information is characterized in that, described method comprises:

2. method according to claim 1, it is characterized in that the search information that described receiving terminal sends is according to the preliminary Search Results of described search acquisition of information, and obtain the first eigenvector of described preliminary Search Results according to described search information and described preliminary Search Results, comprising:

3. method according to claim 2, it is characterized in that, described First Characteristic according to described preliminary Search Results is gathered the first primitive attribute value of the First Characteristic that comprises and the second primitive attribute value of the Second Characteristic that described search information comprises, the First Characteristic of described preliminary Search Results is gathered the First Characteristic that comprises to be filtered, obtain the Second Characteristic set of described preliminary Search Results, comprising:

4. method according to claim 2 is characterized in that, the first primitive attribute value of the First Characteristic that described Second Characteristic set according to described preliminary Search Results comprises consists of the first eigenvector of described preliminary Search Results, comprising:

5. method according to claim 1 is characterized in that, described first eigenvector to described preliminary Search Results carries out concurrent operation, obtains the cryptographic hash corresponding to first eigenvector of described preliminary Search Results, comprising:

6. method according to claim 5 is characterized in that, the described the First Eigenvalue that the first eigenvector of described preliminary Search Results is comprised is encoded concurrently, obtains one-dimensional coding corresponding to described preliminary Search Results, comprising:

7. method according to claim 1 is characterized in that, the described cryptographic hash corresponding according to the first eigenvector of described preliminary Search Results obtained the first sample vector of described preliminary Search Results, comprising:

8. according to claim 1 to the described method of 7 arbitrary claims, it is characterized in that the described information to be recommended of described recommendation comprises:

9. the equipment of a recommendation information is characterized in that, described equipment comprises:

10. equipment according to claim 9 is characterized in that, described the first acquisition module comprises:

11. equipment according to claim 10 is characterized in that, described filter element comprises:

12. equipment according to claim 10 is characterized in that, described Component units comprises:

13. equipment according to claim 9 is characterized in that, described computing module comprises:

14. equipment according to claim 13 is characterized in that, described coding unit comprises:

15. equipment according to claim 9 is characterized in that, described the second acquisition module comprises:

16. to the described equipment of 15 arbitrary claims, it is characterized in that described recommending module comprises according to claim 9: