The content of the invention
It is to provide a kind of file fine system and method that one of viewpoint of the invention is exactly, can be referring concurrently to file
With the correlation of keyword, line search is entered in the frequent behavior with user, to provide more preferably search result.
It is to provide a kind of file fine system and method that another viewpoint of the invention is exactly, and adds user to file
The behavior record of operation and analysis, to provide the search result closer to user's demand.
It is to provide a kind of file fine system and method that another viewpoint of the invention is exactly, and forgets for user and closes
Key word or when cannot provide accurate keyword, stills provide search result, and favourable user's amendment keyword, carries out more
It is accurate to search.
Above-mentioned viewpoint of the invention, proposes a kind of document searching method, is applied to one and media are locally stored, and locally deposits
Storage media storage has multiple files so that a user can enter line search to file.Document searching method includes:By each text
The content of part produces a file full-text index feature of each file of correspondence for the relation of multiple keywords, and stores respectively
Database is searched in one.At least two files in user is for file carry out an operation, or to described these files
When one of them carries out at least two operations, this operation and this corresponding number of times of operation are stored in search database, and produce respectively
A linked character between each file of life, and it is stored in search database.When keyword is searched in user's input one, according to
The file full-text index feature and the relation for searching keyword in database are searched, one first search result is produced, and according to searching
The linked character sought in database and the first search result, produce one second search result.Integrate the first search result and second
Search result, to produce correspondence to search a sequence of the partial document of keyword.
Some embodiments of the invention, multiple files include word shelves, mail shelves, the temporary shelves of browser, Yi Jiwen
Part shelves.And operation of the user to file includes:Open, store, switching, searching, replicating, pasting and link.Linked character bag
Operation of the user to this two file is included, with for this operation, the temporal correlation between two files is constituted.And the time closes
Connection property includes:Time sequencing and time interval.
Some embodiments of the invention, producing the method for the file full-text index feature of each file of correspondence includes:
Using a word frequency-reverse document frequency algorithm, the relation of the content for multiple keywords of each file is calculated.
Some embodiments of the invention, wherein produce the linked character of each file, by record user for
All operations that multiple files are carried out, are obtained with carrying out a correlation analysis.And correlation analysis for periodically, irregularly or
Immediately carry out.
Some embodiments of the invention, wherein the first search result, except the file in the search database
Outside the relation of full-text index feature and search keyword, also including the power according to search keyword in file full-text index feature
Produce again.And the step of integration the first search result and the second search result, further include a weight set-up procedure.
Above-mentioned viewpoint of the invention, proposes a kind of file fine system, including:One is locally stored media, an input
Module, a transcript analysis module, an association analysis module, one search database, one first search module, one second search mould
Block, one integrate module and an output module.Media storage is locally stored multiple files, and input module is suitable to reception one and makes
One search condition of user's input.Transcript analysis module, is used to the relation for multiple keywords by the content of each file, produces
One file full-text index feature of the raw each file of correspondence.Association analysis module be used to when user in multiple files extremely
When few two files carry out an operation or one of described these files are carried out with least two operations, record this operation and
This corresponding number of times of operation, and a linked character of each file is produced respectively.Database purchase is searched in media are locally stored
In, and including:The file full-text index feature of each file, the relative recording and linked character of this operation.First search module
It is suitable to according to search condition, searches the relation of the file full-text index feature and search condition in database, produces one first to search
Seek result;And the second search module, it is suitable to according to the linked character and the first search result searched in database, produce one second
Search result.Module integrated first search result and the second search result are integrated, to produce the partial document of correspondence search condition
A sequence, and output module is used to show this sequence.
Some embodiments of the invention, wherein transcript analysis module produce the file full-text index of each file of correspondence
The method of feature includes:Using a word frequency-reverse document frequency algorithm, the content of each file is calculated for multiple keywords
Relation.
Some embodiments of the invention, wherein association analysis module is further included:Module is collected in one user's behavior, is used
To record user for all operations that file is carried out;And user's behavioural analysis module, for user's behavior
The content that collection module is recorded carries out a correlation analysis, and produces linked character.And correlation analysis is for periodically, irregularly
Or carry out immediately.
Some embodiments of the invention, wherein the first search module, except the file in the search database
Outside the relation of full-text index feature and search condition, also including the weight according to search condition in file full-text index feature,
And produce the first search result.
Some embodiments of the invention, search wherein integrating module and also being used to adjust first including a weight adjusting module
The weight sought between result and the second search result.
By file fine system and method for the invention, operation of the user for file is recorded, to set up each text
Linked character between part, can be referring concurrently to the correlation of file and keyword during search, and linked character, to provide more preferably
Search result.
By file fine system and method for the invention, for user is to the behavior record of file operation and divides
Analysis, considers the relevance between file in search result sequence, can provide the search result closer to user's demand.
By file fine system and method for the invention, even if user forgets keyword or cannot provide accurate
Keyword when, by the search of linked character, still provide search result, and provide the search knot closest to actual key
Infructescence is arranged, and favourable user corrects keyword accordingly, then carries out more accurate search.
Specific embodiment
In order to advantages of the present invention, spirit can be easier and be expressly understood that with feature, subsequently will be with embodiment simultaneously
It is described in detail and discussion with reference to institute's accompanying drawings.It is worth noting that, these embodiments are only representational embodiment of the invention, its
Middle illustrated ad hoc approach, device, condition and material etc. are not limited to of the invention or corresponding embodiment.
It is to illustrate an embodiment of the invention referring to Fig. 1 and Fig. 2, Fig. 1, a kind of side of file fine system
Block Figure 100;Fig. 2 is to illustrate an embodiment of the invention, a kind of flow chart 200 of document searching method.Text of the invention
Part hunting system and method are established in local computing (local computer) or media (local storage are locally stored
Medium in), user can be provided and enters line search for these local files.Local computing includes a PC or service
Device etc., and wherein at least is locally stored media 110 comprising one;Media 110 are locally stored such as to be hard disk (hard disk), consolidates
State hard disk (SSD) or disk array (RAID) etc..User can store all multifiles 112 in media 110 are locally stored, than
Such as by installing software, foundation/editing files, file is replicated from other storage medias or downloading file from world-wide web,
All multifiles 112 will be produced and stored or be temporarily stored into and be locally stored in media 110.These files 112 include Email, text
Word shelves, Office document files and the temporary shelves of browser etc..
Document searching method of the invention was broadly divided into for two megastages, and one is the study stage 210, and one is the search stage 220.
Mainly carried out by this paper analysis modules 124 and association analysis module 126 in the study stage 210.User is usually for text
The keyword search of part 112, can be carried out by an input module 122, input module 122 be such as keyboard, contact panel or
Voice input module etc., such as be input into keyword " Inventec Appliances ", to search Inventec Appliances associated documents.The meeting of this paper analysis modules 124
Keyword in these search activities is coupled with file 112, and a file full-text index is set up for each file 112
Feature 120, and the search database 114 being locally stored in media 110 is stored in, that is, the step 212 in Fig. 2.But file
Content recorded in full-text index feature 120 includes but is not limited to the relation of keyword and file content, and key in file
Frequency that the weight of word, such as keyword occur hereof etc..For example, the file full-text index of certain file 112 is special
Levying 120 may have recorded " Inventec Appliances " and " Inventec " two keywords, and the number of times that this two keywords occur.Although but
Technology of the invention is directed to the search of local file, but in the stage 210 is learnt file full-text index feature 120 foundation,
Can also be for user when world-wide web search is carried out, the keyword being input into is recorded and analyzed.
In certain embodiments of the present invention, produce the method for file full-text index feature 120 can be using word frequency-reverse
Document frequency algorithm (term frequency-inverse document frequency, TF-IDF), as file full text rope
Draw the parameter in feature 120.Wherein word frequency refers to the frequency that special key words occur hereof, i.e. keyword occurrence number
With the ratio of total number of word.Reverse text file frequency is the measurement of keyword general importance, is divided by comprising this pass by total files
The number of files of key word, the business of gained takes the logarithm value and obtains.Above-mentioned word frequency is multiplied with reverse document frequency, i.e., as file
One of parameter in full-text index feature.
Association analysis module 126, the relation mainly set up between each file, because user grasps for file
When making, it will usually have certain train of thought or logic, and these operations are the relevances between file.For example, user exists
When writing of report file shelves is carried out, then data that may be in search for networks download the data on network, or even replicate
Partial content, therefore this part of document file, will have certain to associate with the file downloaded, and may include therebetween same or similar
Keyword.So, these users can exist certain for the operation of file, the file association for being formed when searching
Relation.In certain embodiments of the present invention, association analysis module 126 collects module 128 and user comprising user's behavior
Behavioural analysis module 130.Module 128 is collected in user's behavior, is used to record user for all behaviour that file 112 is carried out
Make, these operation include but is not limited to open, storage, switching, search, replicate, paste and link etc..All user's behaviors
Operation note 116 can be stored in search database 114.User's behavioural analysis module 130, collects for user's behavior
The content that module 128 is recorded, i.e., operation note 116 carries out a correlation analysis, and produces the linked character of each file 112
118, and be stored in search database 114, that is, step 214 in Fig. 2.And the content of linked character 118 at least includes two
Operation note between file, the frequency/frequency of same operation, for the time correlation between certain two file of operation
Property, such as time sequencing or time interval, or to two of a certain file operations etc..But user's behavioural analysis module
130 correlation analysis can be regular, irregularly or immediately carry out, and update the linked character searched in database 114
118。
For example, we with user for file opened (open), storage (save), switching (goto), search
Seek as a example by the operation of (search).For the file of keyword search result, it should which the linked character of reference includes:
The file that (latest) is opened, be most recently stored, switching recently or searching recently recently.
When a certain laggard line search of file of unlatching.
Another file is switched to after opening a certain file.
The occurrence frequency opened, store, switch, search.
Two files are opened, store, switch, searched simultaneously.
Above-mentioned these linked characters should all influence the sequence between keyword search result file.For example, if made
User search keyword A, and want the file looked for be he when mail is write, once with reference to the correlation of keyword A in certain file
Data.So search in addition to for keyword A, it should with reference to open with keyword A mail shelves after, while open or
Later on or switching file, this linked character should influence the sequence of search result.If keyword A search results have text
Part B, C, D and E, and wherein B is mail shelves, and be directed to the linked character with B, it is found that D is relative high with the frequency that B is opened simultaneously,
Then search result should be preferential with D.Certainly, if user can be when searching, except input key condition, can also be defeated
Enter some linked character conditions, system of the invention is further searched it is of course possible to be directed to the linked character.However, making in practice
User is often fuzzy or when cannot determine for linked character, system of the invention be then directed to linked character appearance time
Number or frequency go to adjust search result.For example, if having five files of F, G, H, I and J for keyword search result, so
And be directed to the linked character searched in database and find that I files are especially more with the linked character record of other four files, then can
The search result sequence of I files is adjusted forward.
Therefore, file fine System and method for of the invention can also be divided into two portions in second stage searches the stage 220
Point.The step of user is by input module 122, and input search condition is such as keyword and/or linked character, i.e. Fig. 2
222.In some embodiments of the invention, the first search module 132, can be special according to file full-text index in searching database 114
120 are levied, line search is entered relative to the keyword of input, and produce the first search result (Fig. 2 steps 224).Then, second search
Module 134, according to the search result of linked character 118 and first searched in database 114, produces one second search result (Fig. 2
Step 226).If user has input linked character, line search is entered directly against linked character, and produce second to search knot
Really.If user is not input into linked character, can be according to the first search result in, each file association feature 118 time
Number or frequency, produce the second search result accordingly.And it is then whole with the second search result by the first search result to integrate module 136
Close, and produce the file sequence (Fig. 2 steps 228) of search result, and be supplied to user to refer to by output module 138.
Output module 138 is such as display or voice output module etc..However, when integrate module 136 then by the first search result with
When second search result is integrated, that is, keyword search result and linked character search result are integrated, can be logical according to demand
Cross the weight of weight adjusting module 140, the first search result of adjustment and the second search result.Certainly, in weight adjusting module 140
Weight distribution, can by user's behavioural analysis obtain.For example, each user of monitoring statisticss is opened after searching
The linked character type of file, to adjust the weight of the type.
For example, in certain embodiments of the present invention, according to the search condition of keyword a, the first search module
132, the file sequence produced according to file full-text index feature 120 is A, B, C, D and E (the first search result), and second searches
It is B, C, A, E, D, F and G that module 134 is sought according to the file sequence that linked character 118 is produced, and wherein file F and G is and file C
There is linked character relatively high with A, and after integrating module 136 and integrating, file sequence may be B, A, C, E, D, F and G
(linked character weight is relatively low).In this embodiment it can be found that by file fine System and method for of the invention, can look for
Go out and do not have keyword, but relevance file F and G very high.When the keyword that user is input into is inaccurate or fuzzy, can
It is present in file F and G with finding possible keyword by this mode, user can adjust search condition accordingly.
It is worth noting that, in certain embodiments of the present invention, the first search module 132 and the second search module 134
Operation can independently carry out, integrated again in module 136 to being integrated.In addition, the generation of linked character be not limited to it is right
In the associated record of the associated record, or multiple files or multiple operation rooms of two files or two operation rooms.It is familiar with
This those skilled in the art, it should be understood that it is above-mentioned by user's operation behavior produce file between linked character, to adjust search result, wherein associating
Feature can have other many changes, combinations, so should all be contained in spirit and scope of the invention.
In sum, by file fine system and method for the invention, operation of the user for file is recorded, with
The linked character set up between each file, can be referring concurrently to the correlation of file and keyword during search, and linked character, with
More preferably search result is provided.File fine system and method for the invention, the behavior for user to file operation is remembered
Record and analysis, consider the relevance between file and operation in search result sequence, can provide closer to user's demand
Search result.And by file fine system and method for the invention, even if user forgets keyword or cannot carry
During for accurate keyword, by the search of linked character, relevance search result high is stilld provide, and provide closest in fact
The search result sequence of border keyword, favourable user corrects keyword accordingly, then carries out more accurate search.
By the above detailed description of preferred embodiments, it would be desirable to more clearly describe feature of the invention and spirit, and
Scope of the invention is any limitation as with above-mentioned disclosed preferred embodiment not.On the contrary, the purpose is to wish
Being arranged in the category of the scope of the claims to be applied of the invention for various changes and tool equality can be covered.Although the present invention is
Disclosed above with implementation method, so it is not limited to the present invention, and any those skilled in the art are not departing from this creation
In spirit and scope, when can be used for a variety of modifications and variations, therefore protection scope of the present invention is when regarding appended claims institute
That defines is defined.