CN103425664A

CN103425664A - Method and equipment for searching and displaying entity data units

Info

Publication number: CN103425664A
Application number: CN201210152176XA
Authority: CN
Inventors: 李伟健
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2012-05-16
Filing date: 2012-05-16
Publication date: 2013-12-04

Abstract

The invention discloses a method and equipment for searching and displaying entity data units. The method includes: whether keyword information of a searching request contains category information and/or attribute information or not is judged, the searching range is determined according to different judging results to obtain the corresponding searching results, and accordingly searched data amount is reduced, and searching efficiency is increased. In addition, commodity information, whose similarity with the keyword information satisfies a preset threshold, in the entity data units of the searching results is determined, the determined commodity information is screened and displayed according to preset conditions, pertinence of commodity information checking of a user is increased, commodity information acquiring efficiency of the user is increased, and user experience of entity data unit searching and checking by using the Internet is improved.

Description

A kind of method and apparatus that the solid data unit is searched for, shown

Technical field

The application relates to the internet information processing technology field, relates in particular to a kind of method and apparatus that the solid data unit is searched for, shown.

Background technology

Along with the continuous development of Internet technology, increasing user obtains the information oneself needed by the function of search of internet.Usually the mode of search information is: search server, according to the Client-initiated searching request, is determined the information relevant to the content of carrying in the Client-initiated searching request, and by these information pushings to the user, for the user, check.

The concrete steps of searching for information by search server comprise following content:

The first step, search server receives the searching request that includes keyword message that the user initiates by client.

Second step, search server, according to the described keyword message received, carries out comparing of similarity with all information that comprise in search database, filters out the information that similarity is greater than setting threshold.

Particularly, at first search server carries out thorough search according to the keyword message that receives in database, find comprise keyword message all related datas as Search Results; Then according to predefined similarity threshold, the Search Results that contains keyword message is screened, for example: the described keyword message received is " phone ", search server is according to " phone " that receive, with comprise being searched of " phone " in search database, carry out the similarity comparison by searching the Search Results that obtains with " phone ", filter out and include the information that " phone " content and similarity are greater than 50 percent, may obtain like this 100000 Search Results.

As can be seen here, in existing search technique, after search server receives searching request at every turn, according to the keyword message carried in searching request, content to the full text webpage is searched for, determine the information be complementary with keyword message, make like this search server carry out the efficiency of searching for lower, the search data amount is larger.

The 3rd step, search server is sorted the described information filtered out according to a certain or multiple attribute information of information, for example: using access times as a kind of attribute information of information, by the described information that filters out, according to access times, order is from high to low sorted search server.

Again for example: the attribute information of described information comprises access times, update time and evaluating deg, and the method that search server is sorted to the information filtered out according to access times, update time and evaluating deg is:

Access times, update time and the evaluating deg of each described information that at first, the search server confirmation filters out;

Then, for each information, search server carries out computing according to the mode of setting by the access times of this information, update time and evaluating deg, obtains the aggregative index of this information;

Finally, search server is according to the aggregative index of each information of obtaining according to order from big to small, and the information that described aggregative index is corresponding is sorted.

The 4th step, search server pushes to client using the information after sequence as Search Results, and client is shown the Search Results received in order.

After client arrives the Search Results shown, the data message comprised in each Search Results is limited, such as: in each Search Results, generally include: message subject, informative abstract (about 50 characters), chained address and rise time etc., as shown in Figure 1.The user can not finely determine whether this information is oneself needs by the content shown in Search Results, if think further to understand the particular content of this information, need again to search server, to send the request of checking for a certain Search Results, search server according to receive check request after, by chained address, corresponding site information is pushed to client and is shown; If, by checking that detailed content determines that this information is not the information that the user needs, now, the user need to close current page, continues to check next Search Results.If the user has checked 10 Search Results, be not the information that the user needs, but search server link info webs in full for 10 times by 10 secondary response users' the request of checking, 10 times the full text info web is pushed to client.This process, with respect to the user, check that the efficiency of information is lower; With respect to search server, not only wasted Internet resources, and strengthened the hardware loss of search server.

Therefore, aspect the information search of prior art and demonstration, exist search efficiency low, and the search data amount is large, and then causes the problem of searching resource waste; When Search Results returns to the user, the user checks that the information specific aim of Search Results is poor, and then causes the efficiency of user's obtaining information low, to the problem such as the experience of internet information search is poor.

Summary of the invention

The embodiment of the present application provides a kind of method and apparatus that the solid data unit is searched for, and for solving prior art information, aspect search and demonstration, exist search efficiency low, and the search data amount is large, and then causes the problem of searching resource waste.

A kind of method that the solid data unit is searched for, the method comprises:

The first searching request that reception comprises keyword message;

Judge in described keyword message and whether comprise classification information and/or attribute information;

While having comprised classification information in described keyword message, the corresponding class of this classification information solid data unit is now searched for, obtained Search Results;

While in described keyword message, having comprised attribute information, the solid data unit with this attribute information is searched for, obtained Search Results;

While having comprised classification information and attribute information in described keyword message, to the corresponding class of this classification information now and the solid data unit with this attribute information searched for, obtain Search Results.

A kind of equipment that the solid data unit is searched for, this equipment comprises:

Receiver module, the first searching request that comprises keyword message for reception;

Whether judge module, comprise classification information and/or attribute information for judging described keyword message;

Search module, for the judged result according to different, carry out following different search:

While having comprised classification information and attribute information in described keyword message, to the corresponding class of this classification information now and each solid data unit with this attribute information searched for, obtain Search Results.

The application's beneficial effect is as follows:

The embodiment of the present application is by judging whether the keyword message in searching request comprises classification information and/or attribute information, determine the scope of carrying out search according to different judged results, obtain corresponding Search Results, make the data volume of search reduce, and then improved search efficiency.

The embodiment of the present application also provides a kind of method and apparatus that the Search Results of solid data unit is shown, in order to solve when Search Results is shown, the user checks that the information specific aim of Search Results is poor, and then cause the efficiency of user's obtaining information low, to the problem such as the experience of internet information search is poor.

A kind of method that the Search Results of solid data unit is shown, the method comprises:

Determine the merchandise news that meets setting threshold in Search Results in each solid data unit with the similarity of described keyword message;

According to predefined condition, the merchandise news in the solid data unit of determining is screened;

The sign of the solid data unit at the merchandise news that obtains and this merchandise news place after screening is presented in results page.

A kind of equipment that the Search Results of solid data unit is shown, this equipment comprises:

Determination module, for determining the merchandise news that meets setting threshold in each solid data unit of Search Results with the similarity of described keyword message;

The screening module, for being screened the merchandise news of the solid data unit determined according to predefined condition;

Display module, be presented at results page for the sign by the solid data unit at the merchandise news that obtains and this merchandise news place after screening.

The application's beneficial effect is as follows:

The embodiment of the present application is by determining the merchandise news that meets setting threshold in Search Results in each solid data unit with the similarity of described key word, and according to predefined condition, the merchandise news in the solid data unit of determining is screened and shown, make the user check that the specific aim of merchandise news strengthens, and then improved the efficiency that the user obtains merchandise news, promote the experience of user to using internet to carry out the solid data unit searches and check.

The accompanying drawing explanation

The schematic flow sheet of a kind of method that the solid data unit is searched for that Fig. 1 is the embodiment of the present application one;

Fig. 2 determines the method flow diagram of the comprehensive strength index of each solid data unit in Search Results in embodiment mono-;

The schematic flow sheet of a kind of method that the solid data unit is searched for that Fig. 3 is the embodiment of the present application two;

The schematic flow sheet of a kind of method that the solid data unit is searched for that Fig. 4 is the embodiment of the present application three;

The schematic flow sheet of a kind of method that the Search Results of solid data unit is shown that Fig. 5 is the embodiment of the present application four;

The Search Results that Fig. 6 is electronics retail shop shows pictorial diagram;

The method schematic diagram that Fig. 7 is screening merchandise news in embodiment five;

The structural representation of a kind of equipment that the solid data unit is searched for that Fig. 8 is the embodiment of the present application six;

The structural representation of a kind of equipment that the Search Results of solid data unit is shown that Fig. 9 is the embodiment of the present application seven.

Embodiment

In order to realize the application's purpose, the embodiment of the present application provides a kind of method and apparatus that the solid data unit is searched for, shown, by judging whether the keyword message in searching request comprises classification information and/or attribute information, determine the scope of carrying out search according to different judged results, obtain corresponding Search Results; Meet the merchandise news of setting threshold with the similarity of described key word by determining in Search Results in each solid data unit in addition, and according to predefined condition, the merchandise news in the solid data unit of determining is screened and shown.

Compared with prior art, the application, when the solid data unit is searched for, determines hunting zone by classification information and/or attribute information that in the judgement searching request, keyword message comprises, makes the data volume of search reduce, and then has improved search efficiency; When the solid data unit is shown, determine with the similarity of described key word and meet the merchandise news of setting threshold and shown, make the user check that the specific aim of merchandise news strengthens, and then improved the efficiency that the user obtains merchandise news, promote the experience of user to using internet to carry out the solid data unit searches and check.

It should be noted that, in each embodiment of the application, related solid data unit is a set that includes mass data information.For example: for a portal website, music channel wherein, reading channel etc. can be considered the solid data unit; Again for example, for a shopping website, wherein independently seller user or electronics retail shop also can be considered the solid data unit.

The solid data unit is according to the data message difference wherein comprised, it may be a certain classification or a few kind solid data unit now, for example: in the situation that the solid data unit is electronics retail shop, if in this solid data unit, but the merchandise news of merchandising is mobile phone, computer, this solid data unit is mobile phone classification and computer solid data unit now simultaneously, that is to say, the classification information of this solid data unit is mobile phone and computer.

In addition, the solid data unit is according to the data message difference wherein comprised, it may be to have a certain attribute information or the solid data unit of a few kinds of attribute information, the situation that the solid data unit is electronics retail shop of still take is example, and its attribute information had can be one or more the combination in following attribute information: offtake, buyer user are to its favorable comment degree, the credit rating of electronics retail shop, the return of goods rate of commodity etc.

Scheme below in conjunction with Figure of description to the embodiment of the present application is described in detail.

Embodiment mono-:

As shown in Figure 1, the schematic flow sheet of its a kind of method that the solid data unit is searched for that is the embodiment of the present application one specifically comprises:

Step 101: receive the first searching request that comprises keyword message.

In step 101, described searching request for the first time refers to that client is when needs carry out page search for a certain keyword message, the searching request that comprises this keyword message of initiating to search server.

For example: the search box input keyword message that client provides by search server, as " mobile phone ", and the searching request that comprises " mobile phone " this keyword message to the search server submission, after this, search server can " mobile phone " be search key, and each solid data unit in search database is searched for.

For example, if the solid data unit is electronics retail shop, the set that described search database is the electronics retail shop that registers in shopping website.Again for example, if the solid data unit is a channel in portal website, the set that described search database is each channel in this portal website.

Step 102: judge in described keyword message and whether comprise classification information and/or attribute information,

If judgment result is that in described keyword message and comprised classification information, perform step 103;

If judgment result is that in described keyword message and comprised attribute information, perform step 104;

If judgment result is that in described keyword message and comprised classification information and attribute information, perform step 105;

If judgment result is that in described keyword message and do not comprise classification information and attribute information, determine that hunting zone is searched in ，Zai Gai hunting zone, solid data unit all in search database, obtain Search Results.

It should be noted that, in the scheme of the present embodiment, solid data unit all in search database can be defined as to hunting zone, and then the mode of carrying out the solid data unit searches under this hunting zone is referred to as the default search pattern.

Step 103: the corresponding class of classification information each solid data unit now comprised in described keyword message is searched for, obtained Search Results.

In the scheme of the embodiment of the present application, can in advance each solid data unit in search database be carried out to the classification division, determine the classification information of solid data unit, therefore, in this step 103, can find out the solid data unit that classification information is identical with the classification information comprised in keyword message from search database according to the classification information of definite solid data unit, the hunting zone that the hunting zone in this step 103 forms for these solid data unit that find out.

For example: when the classification information comprised in described keyword message is " mobile phone ", but the classification information of each solid data unit in the query search database is therefrom determined the solid data unit that comprises " mobile phone " in classification information.It should be noted that, the classification information of definite solid data unit can be only " mobile phone ", also can in the situation that comprise " mobile phone " this classification information, also comprise other classification information.

Still take the solid data unit as electronics retail shop be example, if in a certain solid data unit, but the merchandise news of merchandising is only mobile phone, the classification information of this solid data unit is " mobile phone ", is the solid data unit in hunting zone in this step 103; If in a certain solid data unit, but the merchandise news of merchandising is mobile phone and computer, the classification information of this solid data unit is " mobile phone " and " computer " these two classification information, is also the solid data unit in hunting zone in this step 103.

Preferably, in determining search database after the classification information of each solid data unit, can set up the index relative between each classification information and the represented class of the type information solid data unit now, so that in this step 103, according to the classification information comprised in keyword message, determine fast the solid data unit that there is index relative with this classification information, as the solid data unit in hunting zone in this step 103.

After determining hunting zone, take described keyword message as basis, search out feature key word and the akin solid data of described keyword message unit in the solid data unit from hunting zone.Still take the solid data unit as electronics retail shop be example, the key word that the feature key word of solid data unit is merchandise news in the solid data unit, in the present embodiment, the feature key word of solid data unit can be one, also can be for a plurality of.For example, if the merchandise news in a certain solid data unit is mobile phone and computer, the feature key word of this solid data unit can be " mobile phone ", " computer ", " smart mobile phone ", " panel computer " etc.

In the scheme of the present embodiment, can will take classification information determine hunting zone as basis, and then the mode of carrying out the solid data unit searches under this hunting zone is referred to as C(Category, classification) search pattern.

Step 104: the solid data unit with the attribute information comprised in described keyword message is searched for, obtained Search Results.

In the scheme of the embodiment of the present application, can pre-determine the attribute information that each solid data unit in search database has, therefore, in this step 104, the attribute information that can have according to definite solid data unit, find out the attribute information solid data unit identical with the attribute information comprised in keyword message had from search database, the hunting zone that the hunting zone in this step 104 forms for these solid data unit that find out.

For example: when the attribute information comprised in described keyword message is " offtake ", but the attribute information that in the query search database, each solid data unit has, therefrom determine the solid data unit that comprises " offtake " in the attribute information breath had.It should be noted that, the attribute information that definite solid data unit has can be only " offtake " this attribute information, also can be in the situation that there is " offtake " this attribute information, also there are other attribute informations, as also there are the attribute informations such as " the favorable comment degree of buyer user to it ", " credit rating of electronics retail shop ".

Preferably, after the attribute information that each solid data unit has in determining search database, can set up each attribute information and there is the index relative between the solid data unit of this attribute information, so that in this step 104, according to the attribute information comprised in keyword message, determine fast the solid data unit that there is index relative with this attribute information, as the solid data unit in hunting zone in this step 104.

Similar with step 103, after determining hunting zone, take described keyword message as basis, in the solid data unit from hunting zone, search out feature key word and the akin solid data of described keyword message unit.

In the scheme of the present embodiment, can will take attribute information determine hunting zone as basis, and then the mode of carrying out the solid data unit searches under this hunting zone is referred to as PV(Property Value, attribute) search pattern.

Step 105: to the corresponding class of the classification information that comprises in described keyword message now and the solid data unit with the attribute information comprised searched for, obtain Search Results.

In the scheme of the embodiment of the present application, can be according to the scheme of step 103, find out from search database the solid data unit that classification information is identical with the classification information comprised keyword message, and according to the scheme of step 104, again find out the attribute information solid data unit identical with the attribute information comprised in keyword message had from the solid data unit found out, the solid data unit finally found out has formed the hunting zone in this step 105.This step 105 also is not limited to first according to the scheme of step 104, be searched, then the mode of being searched according to the scheme of step 103.

Similar with step 103 and step 104, after determining hunting zone, can take described keyword message as basis, in the solid data unit from hunting zone, search out feature key word and the akin solid data of described keyword message unit.

In the scheme of the present embodiment, can will take classification information and attribute information as basic definite hunting zone, and then the mode of carrying out the solid data unit searches under this hunting zone is referred to as CPV(Category-Property-Value, classification and attribute) search pattern.

Scheme by above step 101 ~ step 105 judges whether the key message carried in searching request comprises classification information and/or attribute information, determine that the scope of carrying out search operation obtains Search Results, so just make search server when the data in database are carried out search operation, dwindled hunting zone, improved the efficiency of search, the wasting of resources of also having avoided search server to cause when utilizing prior art to be searched for.

In the scheme of the present embodiment one, Search Results that can be based on arbitrary in above-mentioned steps 103 ~ step 105, effectively sort to this Search Results, and its sortord is as follows:

Step 106: the comprehensive strength index of determining each solid data unit in Search Results.

In this step 106, as shown in Figure 2, determine that the mode of the comprehensive strength index of each solid data unit in Search Results specifically comprises:

The first step: pre-determine N index, and weighted value corresponding to each index, described N is not less than 1 integer.

Described index is the parameter of the reflection solid data unit performance quality determined based on experience value, take the solid data unit as electronics retail shop be example, described index can be a plurality of dimensions such as retail shop's popularity, retail shop's transaction, service level, merchandise control level and retail shop's operation situation.

Can be according to different indexs the extent of reaction value of assigning weight for solid data unit performance quality, as: the index that more can reflect solid data unit performance quality, its corresponding weighted value is larger, weighted value can be determined with the form of percentage, also can determine by other mathematical way, do not limit here.

Particularly, when determining each index for the importance (weighing the degree of its performance quality) of solid data unit, can determine that to the assessment of solid data unit the data message that solid data unit in business weight and reality produces determines objective weight by the expert of the art, and definite business weight and objective weight are obtained to the weighted value of this index according to sixteen rule weightings.

Second step: determine theoretical maximum value and theoretical minimal value when each index means with quantized versions.

Particularly, described theoretical maximum value and theoretical minimal value can be to determine for the actual needs of an index of solid data unit, also can be rear definite for this quantification of targets degree statistics to all solid datas unit in search database.

The 3rd step: determine the index value of the indices that in Search Results, each solid data unit obtains after being quantized for every index, for indices, carry out respectively following operation:

While occurring being greater than the situation of theoretical maximum value in the index value obtained for a certain index, replace by the theoretical maximum value of this index the index value that is greater than theoretical maximum value, and while occurring being less than theoretical minimizing situation in the index value obtained for a certain index, by the theoretical minimal value of this index, replace and be less than theoretical minimizing index value.

In this step, for a certain index, each solid data unit index value in this index in Search Results can be shown the mode by scatter diagram, and the theoretical maximum value of this index and theoretical minimal value are recorded in the configuration information table of this index, while in the index value that scatter diagram shows, the index value that is greater than theoretical maximum value being arranged, assert that this index value, for abnormal, needs to replace by theoretical maximum value; While in the index value that scatter diagram shows, being less than theoretical minimizing index value, assert that this index value, for abnormal, needs to replace by theoretical minimal value.

The 4th step: the distribution situation to the index value of every index determining in the 3rd step is carried out the discretize processing.

Described discretize is treated to: the discrete schematic diagram that definite index value is reflected carries out interval decile, and each interval after definite decile.

In this step, when the number of the index value obtained for a certain index is more, more by stages such as (as 1000) can be divided into, when the number of the index value obtained for a certain index is less, less by stages such as (as 3) can be divided into.

The 5th step: after every index discretize is processed etc. the index value in by stages carry out unified quantification treatment.

For example: in Search Results, the quantity of solid data unit is 6; Index 1 is retail shop's trading volume, and in Search Results, each solid data unit is respectively 1000,1500,1300,1600,1800 and 2000 for the index value obtained after being quantized of index 1; Index 2 is service level, and in Search Results, each solid data unit is respectively 2,4,3,4,3 and 5 for the index value obtained after being quantized of index 2.

6 index values that obtain after index 1 is quantized are divided into the by stages such as 3,1000 and 1300 in the by stages such as, 1500 and 1600 in the by stages such as, 1800 and 2000 in the by stages such as; 6 index values that obtain after index 2 is quantized are divided into the by stages such as 3,2 and 3 in the by stages such as, 3 and 4 in the by stages such as, 4 and 5 in the by stages such as.

Because the index value size after different quantification of targets differs larger on the order of magnitude, therefore, in this step, need to again carry out unified quantification treatment to the index value of different indexs, make the index value size that finally obtains indices on the same order of magnitude.Specific practice is:

(1), for arbitrary index, it is 0 as the X1 value that the index value comprised in the first interval that this index decile is obtained is quantified as minimum value X1().

The good and bad degree of the performance of the solid data unit that the index value comprised in described the first interval reflects is the poorest.For example: if this index is the good and bad degree (be that index value is larger, the performance of solid data unit is more excellent) of performance of positive reflection solid data unit, the index value comprised in the first interval is not more than the index value comprised in other intervals; If this index is the good and bad degree (be that index value is larger, the performance of solid data unit is poorer) of performance of reverse side reflection solid data unit, the index value comprised in the first interval is not less than the index value comprised in other intervals.

(2), for arbitrary index, it is 100 as the X2 value that the index value comprised between the area postrema that this index decile is obtained is quantified as maximal value X2().

The good and bad degree of the performance of the solid data unit that the index value comprised between described area postrema reflects is optimum.

(3), for arbitrary index, this index decile is obtained except between the first interval and area postrema other intervals in index value, carry out unified quantization according to following formula:

S_{k} = (\frac{Σ_{i = 2}^{k} {Cnt}_{i}}{Σ_{i = 2}^{L - 1} {Cnt}_{i}}) * X_{2} * \frac{(L - 2)}{L}

Wherein, S _kFor for a certain index, the numerical value after the index value unified quantization comprised in k interval, this numerical value is between X1 and X2; Cnt _iFor the number of the index value that comprises in interval for i of a certain index, the number of the by stages such as L is division.

It should be noted that,

Expression is for a certain index, and from starting between the Second Region of dividing to interval end the second from the bottom, each interval is the number sum of the corresponding index value comprised respectively; After L-2 means to remove between the first interval and area postrema remaining divide etc. the number of by stages.

The 6th step: the unified quantization numerical value of the indices of each the solid data unit that utilizes the 5th step to obtain and weighted value corresponding to each index obtain the comprehensive strength index of each solid data unit.

Particularly, by the numerical value of the indices of each solid data unit and weighted value corresponding to each index determined be weighted after summation obtain and value, as the comprehensive strength index of each solid data unit.

If described X1 value is 0, X2 value is 100, the interval of described comprehensive strength index is 0 ~ 100.

More preferably, in order to make this equipment of search server, can identify the comprehensive strength index of each solid data unit, the comprehensive strength index that aforesaid way can be obtained is converted to the form that search server can be identified:

The first step: determine the quantity of solid data unit in Search Results, for example: in Search Results, the quantity of solid data unit is 6;

Second step: the comprehensive strength index of each solid data unit is enlarged to Y doubly, and get the comprehensive strength index of the integral part of the rear comprehensive strength index of expansion as the solid data unit, described Y is positive integer.

The purpose of this step is, because the comprehensive strength index of solid data unit may not be positive integer, therefore, if directly the comprehensive strength index is rounded, may have certain error.So, first the comprehensive strength index is enlarged to Y doubly in this step, as enlarge 10 times, can reduce and round brought error.

The 3rd step: the comprehensive strength index of the integer form that obtains is converted into to the character that search server can be identified form.

For example, the comprehensive strength index of a solid data unit is 53.72193, and rounding after enlarging 100 times is 5372, and by the 5372 similarity forms that are converted into respect to character " 1 ", i.e. the comprehensive strength index of search server identification is 5372 " 1 " characters.

Step 107: determine the feature key word of each solid data unit in described Search Results, and described feature key word and described keyword message are carried out to the comparison of similarity, determine the similarity value of each solid data unit with respect to described keyword message.

Similarity computation process in this step 107 can be by being calculated such as similarity algorithms such as Lucene.

Step 108: the comprehensive strength exponential sum similarity value according to each solid data unit, sorted to each solid data unit in the first Search Results obtained.

Particularly, in step 108, after the similarity value that the comprehensive strength exponential sum step 107 that step 106 is obtained obtains is weighted summation, according to being arranged with value order from big to small, the Search Results after being sorted.

Scheme by step 106 ~ step 108 is sorted to the Search Results obtained, the mode of this sequence can so that with searching request in the high solid data unit of the keyword message similarity that comprises and/or the high solid data unit of solid data unit comprehensive potential index come front, as the user when the Search Results to obtaining is checked, can preferably the solid data unit, can improve the efficiency that the user obtains the solid data unit information like this from theory.

Embodiment bis-:

As shown in Figure 3, the schematic flow sheet of its a kind of method that the solid data unit is searched for that is the embodiment of the present application two.

The scheme of the present embodiment two is that on the Search Results basis obtained at embodiment mono-, the process of again being searched for for this Search Results specifically comprises:

Step 201: the scheme by embodiment mono-obtains the first Search Results for the first searching request.

It should be noted that, the scheme of the present embodiment two is referred to as the first searching request by the searching request related in embodiment mono-, the Search Results obtained in embodiment mono-is referred to as to the first Search Results, this be for the present embodiment two in the second searching request of relating to and the second Search Results obtained distinguish.

Step 202: determine the classification information of the affiliated classification in each solid data unit in described the first Search Results and the attribute information had.

Due to the content difference comprised in the first searching request in embodiment mono-, therefore the first Search Results finally obtained is not identical yet, though but comprise which solid data unit in the first Search Results obtained, all can determine in the following manner the classification information that each solid data unit of the first Search Results is affiliated and the attribute information had:

The first step: the classification information of classification under each solid data unit in the first Search Results is added up, determine the corresponding solid data of each classification information unit, set up the corresponding relation between each classification information and solid data unit.For example: classification information A correspondent entity data cell 1 and solid data unit 2, classification information B correspondent entity data cell 1 and solid data unit 3, classification information C correspondent entity data cell 1, solid data unit 2 and solid data unit 3, now, the classification information of presentation-entity data cell 1 is classification information A, classification information B and classification information C, the classification information of solid data unit 2 is classification information A and classification information C, and the classification information of solid data unit 3 is classification information B and classification information C.

Second step: the attribute information that in the first Search Results, each solid data unit has is added up, determined the corresponding solid data of each attribute information unit, set up the corresponding relation between each attribute information and each solid data unit.For example: attribute information a correspondent entity data cell 1 and solid data unit 2, attribute information b correspondent entity data cell 1 and solid data unit 3, attribute information c correspondent entity data cell 1, solid data unit 2 and solid data unit 3, now, the attribute information that presentation-entity data cell 1 has is attribute information a, attribute information b and attribute information c, the attribute information that solid data unit 2 has is attribute information a and attribute information c, and the attribute information that solid data unit 3 has is attribute information b and attribute information c.

Particularly, take the mode of step 103 obtains in embodiment mono-the first Search Results describes as example:

The first step: determine all classification information of each solid data unit in the first Search Results and the attribute information had.

For example: still take the solid data unit as electronics retail shop is example, the classification information comprised in described keyword message is when " mobile phone ", and to " mobile phone " corresponding class, solid data unit is now searched for, and obtains the first Search Results.

Suppose that the first Search Results about " mobile phone " obtained determined 3 electronics retail shops, the classification information of electronics retail shop 1 is " mobile phone ", and the attribute information had is offtake and the buyer user favorable comment degree to it; The classification information of electronics retail shop 2 also has " computer " this classification information except " mobile phone ", and the attribute information had be electronics retail shop credit rating and offtake; The classification information of electronics retail shop 3 also has " computer " and " TV " these two classification information except " mobile phone ", and the attribute information had be electronics retail shop credit rating, offtake and buyer user is to its favorable comment degree.

Second step: carry out anti-traversing operation for each solid data unit in the first Search Results of determining, add up solid data unit corresponding to each classification information, and add up the solid data unit that each attribute information is corresponding.

At first, determine that the classification information that in the first Search Results, each solid data unit relates to has " mobile phone ", " computer " and " TV ", and the attribute information related to there are " credit rating of electronics retail shop ", " offtake " and " the favorable comment degree of buyer user to it ".

Secondly, determine the solid data unit that each classification information is corresponding.

Still take the solid data unit as electronics retail shop be example, " mobile phone " this classification information correspondent entity data cell is: electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3; " computer " this classification information correspondent entity data cell is: electronics retail shop 2 and electronics retail shop 3; " TV " this classification information correspondent entity data cell is: electronics retail shop 3.

Finally, determine the solid data unit that each attribute information is corresponding.

Still take the solid data unit as electronics retail shop be example, the solid data unit that " offtake " this attribute information is corresponding is: electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3; " the favorable comment degree of buyer user to it " solid data unit that this attribute information is corresponding is: electronics retail shop 1 and electronics retail shop 3; Solid data unit corresponding to " credit rating of electronics retail shop " this attribute information is: electronics retail shop 2 and electronics retail shop 3.

The first Search Results that this step 202 is determined in the first Search Results the classification information of classification under each solid data unit and the mode of the attribute information that has also can obtain for the mode of the mode of step 104 in embodiment mono-or step 105 is carried out, and also can carry out for the first Search Results after sequence in embodiment mono-.

Step 203: will show to the user as navigation information for definite classification information and the attribute information in each solid data unit in described the first Search Results.

In step 203, described navigation information is for informing the user: the first Search Results obtained is solid data unit now in which class, and the solid data unit with which attribute information.Particularly, the form of expression of described navigation information can comprise following two parts content:

Part content is that its form of expression is as shown in table 1 for the classification information of solid data unit in the first Search Results:

Classification information	The quantity of corresponding solid data unit	Corresponding solid data unit specifying information
			Mobile phone	（3）	Electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3
Computer	（2）	Electronics retail shop 2 and electronics retail shop 3
			TV	（1）	Electronics retail shop 3

Table 1

Another part content is for the attribute information that in the first Search Results, the solid data unit has, and its form of expression is as shown in table 2:

Attribute information	The specifying information of corresponding solid data unit
		Offtake	Electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3
The favorable comment degree of buyer user to it	Electronics retail shop 1 and electronics retail shop 3
		The credit rating of electronics retail shop	Electronics retail shop 2 and electronics retail shop 3

Table 2

Step 204: receive the second searching request, carry definite classification information and/or attribute information in described the second searching request.

Described the second searching request can be the navigation information of user according to demonstration, the searching request of again initiating to search server for the search need of user's reality.

For example: if the search need of user's reality is wish to obtain the information relevant to mobile phone, can be in the second searching request this classification information of carrying mobile phone; Perhaps, if the search need of user's reality is to wish to obtain the electronics retail shop with favorable comment, can in the second searching request, carry this attribute information of favorable comment degree.

Step 205: while in described the second searching request, having carried classification information, perform step 206;

While in described the second searching request, having carried attribute information, perform step 207;

While in described the second searching request, having carried classification information and attribute information, perform step 208.

Step 206: searched for having comprised the corresponding class of the classification information solid data unit now carried in the second searching request in described the first Search Results, obtain the second Search Results.

Due to the corresponding relation of having set up in step 202 between each classification information and solid data unit, therefore, in this step 206, can obtain solid data unit corresponding to classification information carried in the second searching request by this corresponding relation, the solid data unit that these obtain is the second Search Results.

Take in step 202 situation that is example for the solid data unit as electronics retail shop, when the classification information of carrying in described the second searching request is " computer ", can directly determine that this classification information correspondent entity data cell of computer is: electronics retail shop 2 and electronics retail shop 3, therefore, electronics retail shop 2 and electronics retail shop 3 are the second Search Results.

Step 207: searched for having comprised the solid data unit with the attribute information carried in the second searching request in described the first Search Results, obtain the second Search Results.

Due to the corresponding relation of having set up in step 202 between each attribute information and solid data unit, therefore, in this step 206, can obtain solid data unit corresponding to attribute information carried in the second searching request by this corresponding relation, the solid data unit that these obtain is the second Search Results.

Take in step 202 situation that is example for the solid data unit as electronics retail shop, when the attribute information carried in described the second searching request is " offtake ", can directly determine that this attribute information correspondent entity data cell of offtake is: electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3, therefore, electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3 are the second Search Results.

Step 208: searched for having comprised the classification information of carrying in the second searching request and the solid data unit of attribute information in described the first Search Results, obtain the second Search Results.

This step 208 can integrating step 206 and the mode of step 207 obtain the second Search Results.

Scheme by above step 201 ~ step 208, after obtaining the first Search Results, initiate the second searching request, and determine to meet the hunting zone of the second searching request on the basis of the first Search Results obtained, and then obtain the second Search Results, dwindle so further the information content of the solid data unit that meets user's needs, accelerated the search efficiency of search server, improved the speed that the user determines the solid data unit needed simultaneously.

In the scheme of the present embodiment two, Search Results that can be based on arbitrary in above-mentioned steps 206 ~ step 208, effectively sort to this Search Results, and its sortord is as follows:

Step 209: while comprising classification information in the second searching request, determine the quantity of the classification information of each solid data unit in the second Search Results, the order ascending according to the quantity of classification information sorted to each solid data unit.

In step 209, comprising classification information in described the second searching request is a classification information of refering in particular to; In described definite the second Search Results, the quantity of the classification information of each solid data unit refers in the second Search Results obtained for step 206, the classification information of each solid data unit (may be one, may be also a plurality of, but must comprise a specific classification information of carrying in the second searching request) quantity.

It should be noted that, because the second Search Results has been the result of being searched for for the classification information of carrying in the second searching request, therefore, there is the classification information of carrying in the second searching request any one solid data unit in the second Search Results, in the case, if in the second Search Results, the classification information of solid data unit is fewer, the specific aim of classification that means the classification information indication that carries in this solid data unit and the second searching request is stronger, the possibility of the solid data unit that this solid data unit need to be checked for the user so is higher.

Still take the solid data unit as electronics retail shop be example, suppose that the second Search Results obtained is electronics retail shop 2 and electronics retail shop 3 in step 206, wherein, the classification information of electronics retail shop 2 is " mobile phone " and " computer " (2 classification information), the classification information of electronics retail shop 3 is " mobile phone ", " computer " and " TV " (3 classification information), the ascending order of quantity according to the classification information of solid data unit is arranged, and the ranking results obtained is: electronics retail shop 2 will come the front of electronics retail shop 3.

Step 210: while in the second searching request, comprising attribute information, determine in the second Search Results that each solid data unit is for the property value of this attribute information, according to the size of described property value, each solid data unit is sorted.

In step 210, because attribute information is the parameter information that characterizes the performance quality of solid data unit, therefore when the property value of the attribute information according to each solid data unit is sorted, need to consider that definite attribute information is to characterize solid data unit forward performance or reverse performance.If characterize the attribute information of solid data unit forward performance, according to the property value size of the attribute information of each solid data unit according to being arranged to bad order from good; If characterize the attribute information of the reverse performance in solid data unit, according to the property value size of the attribute information of each solid data unit according to being arranged to good order from bad.

Still take the solid data unit as electronics retail shop be example, wherein, attribute information " credit rating of electronics retail shop " characterizes the credit of electronics retail shop, belongs to the forward performance of characterization attributes information, the property value of credit rating that is electronics retail shop is higher, and the credit rating of corresponding electronics retail shop is better; Attribute information " return of goods rates of commodity " characterizes the quality of the commodity of electronics retail shop, belongs to the reverse performance of characterization attributes information, and the property value of the return of goods rate of the commodity of electronics retail shop is lower, and commodity corresponding to the merchandise news of corresponding electronics retail shop are better.

For example: in the scheme of utilizing step 207, obtain under the prerequisite of the second Search Results, suppose that the Search Results obtained is: electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3, the attribute information carried in the second searching request is " offtake ", and the method specifically the second Search Results obtained sorted is:

At first, determine the property value 100 of electronics retail shop 1 for " offtake " this attribute information;

And electronics retail shop 2 is for the property value 1000 of " offtake " this attribute information;

And electronics retail shop 3 is for the property value 10 of " offtake " this attribute information;

Secondly, according to the size of the property value of " offtake ", sorted from high to low, the ranking results obtained is: electronics retail shop 2, electronics retail shop 1 and electronics retail shop 3.

Again for example: in the scheme of utilizing step 207, obtain under the prerequisite of the second Search Results, suppose that the Search Results obtained is: electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3, the attribute information carried in the second searching request is " return of goods rates of commodity ", and the method specifically the second Search Results obtained sorted is:

At first, determine the property value 80% of electronics retail shop 1 for " return of goods rates of commodity " this attribute information;

And electronics retail shop 2 is for the property value 50% of " return of goods rates of commodity " this attribute information;

And electronics retail shop 3 is for the property value 10% of " return of goods rates of commodity " this attribute information;

Secondly, according to the size of the property value of " offtake ", sorted from high to low, the ranking results obtained is: electronics retail shop 3, electronics retail shop 2 and electronics retail shop 1.

It should be noted that, when determining property value corresponding to attribute information, attribute information is carried out obtaining after quantification treatment, the standard that attribute information is quantized is not done restriction here.

Distinguishingly, for some electronics retail shop, in the test operation stage, may complete a transaction with attempting, the favorable comment degree obtained is 100%, if do not relate to " offtake " this attribute information in the second searching request now received, and only relate to " the favorable comment degree of buyer user to it " this attribute information, this electronics retail shop probably sorts in forward position, but in fact, this sales volume is low, the electronics retail shop that not necessarily the actual hope of user is checked of test operation electronics retail shop that buyer user is high to its favorable comment degree, therefore, introduce fast-selling filtering algorithm in the scheme of the present embodiment.

Described fast-selling filtering algorithm is: using electronics retail shop within recently a period of time " offtake " this attribute information of (as 30 days) all the time as sort by, and, in conjunction with the attribute information carried in the second searching request, jointly electronics retail shop is sorted.

For example: if the attribute information carried in the second searching request does not comprise " offtake ", according to the weighted value of the attribute information carried in the second searching request and " offtake " these attribute informations, the comprehensive quality of determining electronics retail shop, sorted.

If the attribute information carried in the second searching request comprises " offtake ", can comprehensively determine the quality of electronics retail shop directly according to the weighted value of each attribute information carried in the second searching request, sorted.

Step 211: while comprising classification information and attribute information in the second searching request, according to the quantity of the classification information of each solid data unit in the second Search Results with for the property value of this attribute information, each solid data unit is sorted.

This step 211 comprehensively mode of step 209 and step 210 is sorted, and particularly, can comprise following several mode:

The first sortord: first the quantity according to classification information sorts, then according to the property value size of attribute information, ranking results is adjusted.

The second sortord: first the property value according to attribute information sorts, then according to the quantity of classification information, ranking results is adjusted.

The third sortord: for classification information and attribute information distribute different weights, utilize the quantity of classification information and property value and each self-corresponding weighted value of attribute information to be weighted summation, according to each solid data unit being sorted with the value size of obtaining.

It should be noted that, the division of weight can be determined respective value according to actual needs, also can carry out statistical calculation to a large amount of data and obtain.

Scheme by step 209 ~ step 211 is sorted to the second Search Results obtained, as much as possible the solid data unit of meeting consumers' demand is arranged in to front, facilitate the solid data unit needed that finds of user's fast accurate, improved speed and accuracy that the user obtains the solid data unit information.

Embodiment tri-:

The present embodiment three is to take electronics retail shop is searched for as example, describes the detailed process of embodiment mono-and embodiment bis-each steps in detail.As shown in Figure 4, specifically comprise the following steps:

Step 301: receive and comprise first searching request of keyword message for " mobile phone of selling fast ".

In step 301, the user can send by retail shop's search engine the first searching request that comprises keyword message to search server.Described retail shop search engine refers in platform-type e-commerce website for other search engine of retail shop's level.This search engine, according to the keyword message of user's input, is searched for significant information of retail shop such as retail shop's title, the manager of retail shop title, main management classification, main management brand, product features keywords, and then is searched out the electronics retail shop that meets user's query demand.

Step 302: determine in described keyword message and comprise classification information and attribute information.

In step 302, through classification information and attribute information pre-stored in the keyword message in step 301 and search database are compared, determine in the first searching request and comprised classification information " mobile phone " and attribute information " offtake ".

Step 303: under the corresponding classification of this classification information (" mobile phone "), the electronics retail shop information with this attribute information (" offtake ") is searched for, obtain the first Search Results.

In step 303, can be according to the scheme of step 103 in embodiment mono-, find out from search database the electronics retail shop information that classification information is identical with the classification information " mobile phone " comprised keyword message, again according to the scheme of step 104 in embodiment mono-, again find out the electronics retail shop information that the attribute information had and the attribute information comprised in keyword message " offtake " are identical from the electronics retail shop found out, the electronics retail shop finally found out has formed the hunting zone in this step 303, because the keyword message in the present embodiment three has only comprised classification information and attribute information, therefore, electronics retail shop in definite hunting zone is the first Search Results herein.

Suppose: the first Search Results obtained is electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3.

Step 304: determine the classification information under each electronics retail shop information and the attribute information had in described the first Search Results.

For example: the classification information of determining electronics retail shop 1 is " mobile phone " and " computer ", and attribute information is " offtake (100) " and " the favorable comment degree (3.5) of buyer user to it ";

The classification information of determining electronics retail shop 2 is " mobile phone ", " computer " and " TV ", and attribute information is " offtake (1000) " and " the favorable comment degree (4.5) of buyer user to it ";

The classification information of determining electronics retail shop 3 is " mobile phone " and " TV ", and attribute information is " offtake (500) " and " the favorable comment degree (3) of buyer user to it ".

Step 305: will determine that in described the first Search Results, the classification information under each electronics retail shop information is set up corresponding relation with the attribute information had with corresponding electronics retail shop information, and added up, the navigation information that generation comprises classification information and attribute information, and show.

For example: classification information: " mobile phone " respective amount is 3, i.e. electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3;

" computer " respective amount is 2, i.e. electronics retail shop 1 and electronics retail shop 2;

" TV " respective amount is 2, i.e. electronics retail shop 2 and electronics retail shop 3;

Attribute information: the electronics retail shop that " offtake " is corresponding is: electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3;

The electronics retail shop that " the favorable comment degree of buyer user to it " is corresponding is: electronics retail shop 1, electronics retail shop 2 and electronics retail shop 3.

Step 306: the comprehensive strength index of determining each electronics retail shop information in the first Search Results.

In step 306, the mode that can describe according to step 106 in embodiment mono-is determined.

It should be noted that, in determining the first Search Results, before the comprehensive strength index of each electronics retail shop information, can pre-determine the index that characterizes retail shop's informix potential index.

For this particular entity data cell of electronics retail shop information, generally from retail shop's popularity, retail shop's transaction, service level, merchandise control level and five dimensions of retail shop's operation situation, build index system, and segment the detailed index under each dimension, form the hierarchy chart with hierarchical logic relation.Wherein, retail shop's popularity dimension index can be: average daily merchandise news details page pageview, search represent clicking rate, browse merchandise news details page pageview etc. per capita; Retail shop's transaction dimension index can be: buying conversion ratio (is that the user enters the page at retail shop's information place by variety of way, and carry out to click the merchandise news in collection retail shop's information or collection retail shop information or add the access times of shopping cart be that purchase intention UV(is called for short PIUV), be exactly total buy the UV of a certain merchandise news in conversion ratio PIUV/ retail shop information), average daily turnover, average daily objective unit price, the average daily stroke count etc. that strikes a bargain; Service level dimension index can be: service quality evaluation mark, reimbursement rate, the rate of complaints, frequent visitor lead etc.; The horizontal latitude index of merchandise control can be: best-selling product rate, unseasonable goods rate, new commodity rate etc.; Retail shop's operation latitude index can be: credit grade, credit appraisal positive rating, rule are observed to such an extent that grade.

Step 307: the feature key word of determining each electronics retail shop in described the first Search Results, and the keyword message comprised in described feature key word and the first searching request is carried out to the comparison of similarity, determine the similarity value of each electronics retail shop with respect to keyword message.

Step 308: the comprehensive strength exponential sum similarity value according to each electronics retail shop, sorted to each electronics retail shop in the first Search Results obtained.

In this step 308, can be according to the weight that comprehensive strength exponential sum similarity value is set respectively, after the comprehensive strength exponential sum similarity value of each electronics retail shop is weighted to summation, according to the descending order sequence of value.

Suppose that the ranking results obtained is electronics retail shop 2 after step 306 ~ step 308, electronics retail shop 1 and electronics retail shop 3.

Step 309: receive the second searching request, carried definite described classification information and/or the attribute information of classification information " computer " in described the second searching request.

The second searching request in this step 309 can be that the user initiates according to the navigation information shown.

Step 310: determine in described the second searching request and carried classification information.

Step 311: to being searched for for computer electronics retail shop information now in described the first Search Results, obtain the second Search Results.

Classification information in the first Search Results of determining according to step 304 under each electronics retail shop information, the known electronics retail shop that meets " computer " this classification information is electronics retail shop 1 and electronics retail shop 2, is the second Search Results.

Step 312: determine the quantity of the classification information of each electronics retail shop in the second Search Results, the order ascending according to the quantity of classification information sorted to each electronics retail shop.

Because the classification information content of electronics retail shop 1 is 2, the classification information content of electronics retail shop 2 is 3, and therefore, the ranking results obtained is that electronics retail shop 1 is before electronics retail shop 2.

Embodiment tetra-:

As shown in Figure 5, be the schematic flow sheet of a kind of method that the Search Results of solid data unit is shown of the embodiment of the present application four.Embodiment tetra-realizes on the basis of embodiment mono-, embodiment bis-or embodiment tri-, specifically comprises:

Step 401: determine the merchandise news that meets setting threshold in Search Results in each solid data unit with the similarity of described keyword message.

In step 401, described Search Results can be the Search Results obtained for the first searching request in embodiment mono-, can be also the Search Results obtained for the second searching request in embodiment bis-.

Particularly, at first, key word and the described keyword message of merchandise news in each solid data unit in the Search Results obtained are carried out to comparing of similarity, for example: the result that obtains the similarity comparison by Lucene; Secondly, the result of comparison and setting threshold are compared, determine the merchandise news collection that meets described keyword message.

Described setting threshold can be determined based on experience value, when the accuracy requirement to Search Results is higher, can set higher threshold value, otherwise, can set lower threshold value.

When the result by comparison and setting threshold compare, can be greater than the merchandise news that setting threshold is corresponding using being greater than setting threshold as condition, determining the result that meets described comparison, as the merchandise news collection that meets described keyword message.

For example: suppose that setting threshold is 0.8, when in the Search Results by obtaining, in each solid data unit, the key word of merchandise news and described keyword message carry out comparing of similarity, when the result of comparison is greater than 0.8, determine that the commodity that in Search Results, in the solid data unit, the key word of merchandise news is corresponding are the merchandise news that meets the keyword message comprised in searching request.

Step 402: the merchandise news in the solid data unit of determining is screened according to predefined condition.

In step 402, described predefined condition is the K level Rule of judgment with priority orders.

For the merchandise news in the solid data unit, carry out respectively following operation:

The first step: determine and meet the merchandise news of described K level Rule of judgment and the quantity of merchandise news;

Second step: judge whether the quantity of described merchandise news is not less than the demonstration quantity of setting, if, using definite merchandise news as the merchandise news filtered out for showing; Otherwise, carry out the 3rd step;

The demonstration quantity of described setting refers to the number of the merchandise news that can be shown, generally is set as 5.

The 3rd step: when the quantity of merchandise news is less than the demonstration quantity of setting, the minimum one-level Rule of judgment of shielding K level Rule of judgment medium priority, for remaining K-1 Rule of judgment, continue to carry out the operation of the first step to the three steps, until all conductively-closeds of all Rule of judgment.

For example, if described at least one-level Rule of judgment is 2 grades of Rule of judgment, at first determine the merchandise news of satisfied 2 grades of Rule of judgment and the quantity of merchandise news; When the quantity of the merchandise news that meets 2 grades of Rule of judgment is less than the demonstration quantity of setting, shield 2 grades of one-level Rule of judgment that the Rule of judgment medium priority is minimum, continue for remaining 1 grade of Rule of judgment, determine and meet the merchandise news of this 1 grade of Rule of judgment and the quantity of merchandise news.When the quantity of the merchandise news that meets 1 grade of Rule of judgment is not less than the demonstration quantity of setting, to meet the merchandise news of 1 grade of Rule of judgment as the merchandise news filtered out for showing, otherwise, shield 1 grade of Rule of judgment, now, the all conductively-closeds of all Rule of judgment, will meet the merchandise news of 1 grade of Rule of judgment as the merchandise news for showing, simultaneously, filter out the merchandise news of some for showing from the remaining merchandise news in solid data unit, make finally and equal for the merchandise news quantity shown the demonstration quantity of setting.

Step 403: the sign of the solid data unit at the merchandise news that obtains and this merchandise news place after screening is presented in results page.

In step 403, search server is pushed to client by the merchandise news that obtains after screening and the sign of the solid data unit at this merchandise news place, the results page demonstration in the display interface of client, as shown in Figure 6.

Scheme by embodiment tetra-, the merchandise news comprised in solid data unit in Search Results is screened, determine the most approaching with described keyword message and meet the merchandise news imposed a condition, the content finally shown in client has not only comprised the sign of solid data unit, and comprised in the solid data unit the most approaching with described keyword message and met the merchandise news imposed a condition, improved like this efficiency that the user obtains merchandise news, the internet hunt that has promoted the user is experienced.

Embodiment five:

The present embodiment five is detailed descriptions of each step in embodiment tetra-, specifically comprises:

The first step: determine the merchandise news that meets setting threshold in Search Results in each electronics retail shop information with the similarity of described keyword message.

Second step: the merchandise news in the electronics retail shop determined is screened according to predefined condition.

In this step, the condition of described setting can be set according to the characteristics of merchandise news in electronics retail shop information.Such as: be whether best buy information, whether be detail view commodity, the commodity sales volume of the nearest 30 days sales volume threshold value that whether is greater than setting, whether be whether the real scene shooting picture of merchandise news and purchase intention conversion ratio are greater than 0 etc.

Described best buy information is merchandise news indices merchandise news or the guaranteed merchandise news of service quality preferably.Described detail view sheet commodity information is that the picture during the merchandise news that authenticates of website official is described has the merchandise news of local detail picture clearly, and it is clear detailed that this merchandise news is described, and confidence level is higher.The real scene shooting picture of described commodity be retail shop's information according to merchandise news describe picture be whether in kind the shooting and in merchandise news is described record.

Wherein, above-mentioned condition can be set according to different priority orders.

Whether usually, can set the highest condition of priority is best buy information; Priority next whether be the detail view commodity; Priority the 3rd be the sales volume threshold value whether the commodity sales volume of nearest 30 days is greater than setting; Priority the 4th whether be the real scene shooting picture of merchandise news; Whether what priority was minimum is greater than 0 for the purchase intention conversion ratio.

More preferably, for the merchandise news in electronics retail shop in search database is set different Rule of judgment, and combined with regard to the Rule of judgment of setting, obtained a plurality of Rule of judgment collection, as shown in Figure 7:

The first Rule of judgment collection is: merchandise news is best buy information and is the Rule of judgment collection that detail view commodity and the commodity sales volume of nearest 30 days are greater than the sales volume threshold value of setting;

The second Rule of judgment collection is: merchandise news be best buy information but be not detail view commodity, the commodity sales volume of the nearest 30 days sales volume threshold value that is greater than setting, be the Rule of judgment collection of the real scene shooting picture of merchandise news;

The 3rd Rule of judgment collection is: merchandise news is best buy information but is not that detail view commodity, the commodity sales volume of nearest 30 days are greater than the sales volume threshold value of setting and are not the Rule of judgment collection of the real scene shooting picture of merchandise news;

The 4th Rule of judgment collection is: merchandise news is best buy information, is that detail view commodity, the commodity sales volume of nearest 30 days are less than the Rule of judgment collection that is greater than 0 in sales volume threshold value, the purchase intention conversion ratio set;

The 5th Rule of judgment collection is: merchandise news is best buy information, is that detail view commodity, the commodity sales volume of nearest 30 days are less than the Rule of judgment collection that equals 0 in sales volume threshold value, the purchase intention conversion ratio set;

The 6th Rule of judgment collection is: merchandise news is best buy information but is not that detail view commodity, the commodity sales volume of nearest 30 days are less than in the sales volume threshold value of setting, are the Rule of judgment collection that real scene shooting picture, the purchase intention conversion ratio of merchandise news is greater than 0;

The 7th Rule of judgment collection is: merchandise news is best buy information but is not that detail view commodity, the commodity sales volume of nearest 30 days are less than in the sales volume threshold value of setting and are not the Rule of judgment collection that real scene shooting picture, the purchase intention conversion ratio of merchandise news is greater than 0;

The 8th Rule of judgment collection is: merchandise news is best buy information but is not that detail view commodity, the commodity sales volume of nearest 30 days are less than in the sales volume threshold value of setting, are the Rule of judgment collection that real scene shooting picture, the purchase intention conversion ratio of merchandise news equals 0;

The 9th Rule of judgment collection is: merchandise news is best buy information but is not that detail view commodity, the commodity sales volume of nearest 30 days are less than in the sales volume threshold value of setting and are not the Rule of judgment collection that real scene shooting picture, the purchase intention conversion ratio of merchandise news equals 0;

The tenth Rule of judgment collection is: merchandise news is not best buy information, is that detail view commodity, the commodity sales volume of nearest 30 days are less than in the Rule of judgment collection of the sales volume threshold value of setting;

The 11 Rule of judgment collection is: merchandise news is not best buy information, is not or not detail view commodity, the commodity sales volume of nearest 30 days are greater than in the sales volume threshold value of setting, are the Rule of judgment collection of the real scene shooting picture of merchandise news;

The 12 Rule of judgment collection is: merchandise news is not best buy information, is not or not detail view commodity, the commodity sales volume of nearest 30 days are greater than in the sales volume threshold value of setting and are not the Rule of judgment collection of the real scene shooting picture of merchandise news;

The 13 Rule of judgment collection is: merchandise news is not best buy information, be that detail view commodity, the commodity sales volume of nearest 30 days are less than the Rule of judgment collection that is greater than 0 condition in sales volume threshold value, the purchase intention conversion ratio set;

The 14 Rule of judgment collection is: merchandise news is not best buy information, be that detail view commodity, the commodity sales volume of nearest 30 days are less than the Rule of judgment collection that equals 0 condition in sales volume threshold value, the purchase intention conversion ratio set;

The 15 Rule of judgment collection is: merchandise news is not best buy information, is not or not detail view commodity, the commodity sales volume of nearest 30 days are less than in the sales volume threshold value of setting, are the Rule of judgment collection that real scene shooting picture, the purchase intention conversion ratio of merchandise news is greater than 0;

The 16 Rule of judgment collection is: merchandise news is not best buy information, be not or not detail view commodity, the commodity sales volume of nearest 30 days be less than be not in the sales volume threshold value of setting, the Rule of judgment collection that real scene shooting picture, the purchase intention conversion ratio of merchandise news is greater than 0;

The 17 Rule of judgment collection is: merchandise news is not best buy information, is not or not detail view commodity, the commodity sales volume of nearest 30 days are less than in the sales volume threshold value of setting, are the Rule of judgment collection that real scene shooting picture, the purchase intention conversion ratio of merchandise news equals 0;

The 18 Rule of judgment collection is: merchandise news is not best buy information, be not or not detail view commodity, the commodity sales volume of nearest 30 days be less than be not in the sales volume threshold value of setting, the Rule of judgment collection that real scene shooting picture, the purchase intention conversion ratio of merchandise news equals 0.

When the merchandise news in the electronics retail shop to determining is screened, can concentrate and select a Rule of judgment collection from above-mentioned 18 Rule of judgment, and utilize the Rule of judgment of selecting to concentrate a plurality of Rule of judgment with priority sequencing that comprise, carry out the screening of merchandise news according to the mode of step 402.

More preferably, in the scheme of this step, can the merchandise news that meet Rule of judgment in electronics retail shop be shown to the user according to the mode of step 402, make the user can electron gain retail shop in the good merchandise news of comprehensive state; On the other hand, for the merchandise news that the user is obtained reflects user's actual search demand more exactly, can in electronics retail shop with the second searching request in the merchandise news of classification information matches, carry out the screening of merchandise news according to the mode of step 402, if the quantity of the merchandise news of the classification information matches in all and the second searching request is less than while showing quantity, can, for other classes merchandise news now in electronics retail shop, carry out the screening of merchandise news according to the mode of step 402.

The 3rd step: the sign of retail shop's information at the merchandise news that obtains and this merchandise news place after screening is presented in results page.

Particularly, in this step, the mode by the merchandise news that obtains after screening by the show window position is shown.

Described show window position refers in search result list, for each retail shop's information arranges a window with page-turning function.

By the number of show window in the show window position is set, the merchandise news obtained after screening is placed in corresponding show window and is shown.

More preferably, the adjustment button is set in the show window position, can, according to the number of visits of the merchandise news to retail shop's information and displaying, adjusts in real time the order of the merchandise news shown in each show window in the show window position.

By show window position Indivitual display merchandise news, realized Recommendations information targetedly, improve the degree of understanding of user to merchandise news, and then improved the efficiency that the user obtains merchandise news.

Embodiment six:

As shown in Figure 8, the structural representation for a kind of equipment that the solid data unit is searched for of the embodiment of the present application six, specifically comprise: receiver module 61, judge module 62 and search module 63.Wherein

Receiver module 61, the first searching request that comprises keyword message for reception;

Whether judge module 62, comprise classification information and/or attribute information for judging described keyword message;

Search module 63, for the judged result according to different, carry out following different search:

Described equipment also comprises: determine information module 64.Wherein,

Determine information module 64, for the classification information of determining each solid data unit of described Search Results and the attribute information had.

Particularly, described receiver module 61, for receiving the second searching request, carry definite described classification information and/or attribute information in described the second searching request.

Described search module 63, the information difference of carrying for the second searching request received according to receiver module, carry out following different search:

While in described the second searching request, carrying definite classification information, the corresponding class of this classification information solid data unit now in described Search Results is searched for, obtained new Search Results;

While in described the second searching request, carrying definite attribute information, the solid data unit that has this attribute information in described Search Results is searched for, obtained new Search Results;

While in described the second searching request, carrying definite classification information and attribute information, to the corresponding class of this classification information in described Search Results now and the solid data unit with this attribute information searched for, obtain new Search Results.

Described equipment also comprises: determine comprehensive strength index module 65, determine similarity value module 66 and the first order module 67.Wherein,

Determine comprehensive strength index module 65, for determining the comprehensive strength index of each solid data unit of Search Results;

Determine similarity value module 66, for determining the feature key word of each solid data unit of Search Results, and described feature key word and described keyword message are carried out to the comparison of similarity, determine the similarity value of each solid data unit with respect to described keyword message;

The first order module 67, the comprehensive strength exponential sum similarity value for according to each solid data unit, sorted to the solid data unit in the Search Results obtained.

Particularly, described definite comprehensive strength index module 65, specifically for basis, it is the weighted value of the default indices in solid data unit and each index, by after the numerical value of this solid data unit indices and corresponding weighted value weighted sum, obtain the comprehensive strength index of this solid data unit.

Described equipment also comprises: classification information sorting module 68, attribute information order module 69 and classification combinations of attributes order module 70.Wherein,

Classification information sorting module 68, for when the second searching request comprises classification information, determine the quantity of the classification information of each solid data unit in the Search Results based on the second searching request, the order ascending according to the quantity of classification information sorted to each solid data unit.

Attribute information order module 69, for when the second searching request comprises attribute information, determine in the Search Results based on the second searching request that each solid data unit is for the property value of this attribute information, according to the size of described property value, each solid data unit is sorted.

Classification combinations of attributes order module 70, for when the second searching request comprises classification information and attribute information, according to the quantity of the classification information of each solid data unit in the Search Results based on the second searching request with for the property value of this attribute information, each solid data unit is sorted.

Embodiment seven:

As shown in Figure 9, the structural representation for a kind of equipment that the Search Results of solid data unit is shown of the embodiment of the present application seven specifically comprises: determine merchandise news module 71, screening module 72 and display module 73.Wherein,

Determine merchandise news module 71, for determining the merchandise news that meets setting threshold in each solid data unit of Search Results with the similarity of described keyword message;

Screening module 72, for being screened the merchandise news of the solid data unit determined according to predefined condition;

Display module 73, be presented at results page for the sign by the solid data unit at the merchandise news that obtains and this merchandise news place after screening.

Particularly, described screening module 72, specifically for according to presetting the multistage Rule of judgment with priority orders, to the merchandise news in the solid data unit, carry out following operation:

Determine and meet the merchandise news of described multistage Rule of judgment and the quantity of merchandise news, and the demonstration quantity of the quantity of definite merchandise news and setting is compared, when the quantity of definite merchandise news is not less than the demonstration quantity of setting, using definite merchandise news as the merchandise news filtered out for showing;

When the quantity of merchandise news is less than the demonstration quantity of setting, shield the one-level Rule of judgment that multistage Rule of judgment medium priority is minimum, repeat aforesaid operations for remaining Rule of judgment, until all conductively-closeds of all Rule of judgment.

Obviously, those skilled in the art can carry out various changes and modification and not break away from the spirit and scope of the present invention the present invention.Like this, if within of the present invention these are revised and modification belongs to the scope of the claims in the present invention and equivalent technologies thereof, the present invention also is intended to comprise these changes and modification interior.

Claims

1. the method that the solid data unit is searched for, is characterized in that, the method comprises:

The first searching request that reception comprises keyword message;

2. the method for claim 1, is characterized in that, described method also comprises:

Determine the classification information of each solid data unit in described Search Results and the attribute information had;

Receive the second searching request, carry definite described classification information and/or attribute information in described the second searching request;

3. method as claimed in claim 2, is characterized in that, described method also comprises:

Determine the comprehensive strength index of each solid data unit in Search Results;

Determine the feature key word of each solid data unit in Search Results, and described feature key word and described keyword message are carried out to the comparison of similarity, determine the similarity value of each solid data unit with respect to described keyword message;

Comprehensive strength exponential sum similarity value according to each solid data unit, sorted to the solid data unit in the Search Results obtained.

4. method as claimed in claim 3, is characterized in that, it is characterized in that, described method also comprises:

While comprising classification information in the second searching request, determine the quantity of the classification information of each solid data unit in the Search Results based on the second searching request, the order ascending according to the quantity of classification information sorted to each solid data unit;

While in the second searching request, comprising attribute information, determine in the Search Results based on the second searching request that each solid data unit is for the property value of this attribute information, according to the size of described property value, each solid data unit is sorted;

While comprising classification information and attribute information in the second searching request, according to the quantity of the classification information of each solid data unit in the Search Results based on the second searching request with for the property value of this attribute information, each solid data unit is sorted.

5. method as claimed in claim 3, is characterized in that, determines the comprehensive strength index of each solid data unit, specifically comprises:

According to the weighted value that is the default indices in solid data unit and each index, by after the numerical value of this solid data unit indices and corresponding weighted value weighted sum, obtain the comprehensive strength index of this solid data unit.

6. the method that the arbitrary Search Results obtained of claim 1 ~ 5 is shown, is characterized in that, the method comprises:

7. method as claimed in claim 6, is characterized in that, according to predefined condition, the merchandise news in the solid data unit of determining screened, and specifically comprises:

Preset the multistage Rule of judgment with priority orders;

For the merchandise news in the solid data unit, carry out following operation:

8. the equipment that the solid data unit is searched for, is characterized in that, this equipment comprises:

9. equipment as claimed in claim 8, is characterized in that, described equipment also comprises:

Determine information module, for the classification information of determining each solid data unit of described Search Results and the attribute information had;

Described receiver module, also for receiving the second searching request, carry definite described classification information and/or attribute information in described the second searching request;

Described search module, the information difference of also carrying for the second searching request received according to receiver module, carry out following different search:

10. the equipment that the Search Results that the search equipment arbitrary based on claim 8 ~ 9 obtained is shown, is characterized in that, this equipment comprises:

The merchandise news determination module, for determining the merchandise news that meets setting threshold in each solid data unit of Search Results with the similarity of described keyword message;