CN100428224C - Dynamic database reordering system - Google Patents

Dynamic database reordering system Download PDF

Info

Publication number
CN100428224C
CN100428224C CNB038007894A CN03800789A CN100428224C CN 100428224 C CN100428224 C CN 100428224C CN B038007894 A CNB038007894 A CN B038007894A CN 03800789 A CN03800789 A CN 03800789A CN 100428224 C CN100428224 C CN 100428224C
Authority
CN
China
Prior art keywords
word
reorders
frequency
database
ordering
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CNB038007894A
Other languages
Chinese (zh)
Other versions
CN1703693A (en
Inventor
伊斯安·布哦福特
大卫·彰·柯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Historic AOL LLC
Original Assignee
America Online Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by America Online Inc filed Critical America Online Inc
Publication of CN1703693A publication Critical patent/CN1703693A/en
Application granted granted Critical
Publication of CN100428224C publication Critical patent/CN100428224C/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Images

Abstract

A dynamic database reordering system provides a linguistics database that contains words that are ordered according to a linguistics model that dictates the order in which words are presented to a user. While a user enters keystrokes on a keypad of a communications device is pressing keys, the invention predicts the words, letters, numbers, or word stubs that the user is trying to enter. The invention reorders the linguistics model order based on the user's usage of the system by tracking the user's word selections. Once a word has been selected as a result of a next key selection (the nexted word), a frequency value is applied to the selected word and the word ordered first by the linguistics model in the linguistics database for that key sequence. The frequency value of the nexted word will become greater than the frequency value of the first displayed word upon repeated nexting to the same word. Subsequent user entries of the key sequence for the nexted word and the first ordered word will result in displaying the nexted word before the word ordered first by the linguistics model.

Description

The equipment of dynamic database reordering and method
Technical field
The present invention relates to ordering to the element that from database, extracts.Particularly, the present invention relates to the ordering of the order of the database element of user's actual selection the display element that is derived from database.
Background technology
Technical progress has made manufacturer produce various small sized personal electronic installations.Some examples are the data input cell of PDA(Personal Digital Assistant), cell phone, Small Form Factor and the communication unit of other Small Form Factor.
Because the size of these miniature electric data sets reduces, make the size of the data entry keyboard that these install dwindle.The scheme that reduces keyboard size is to reduce the quantity of button on the keyboard.The quantity that reduces button has produced many problems.Be apparent that most, cause such as the keyboard on the cell phone and overload.A singly-bound can be represented many characters.When input text was in the device of reduction keyboard, the text that makes the user import reasonable amount became not only slow but also difficult.The overload button needs a plurality of processing to obtain correct character usually.
Keyboard disambiguation system (such as, the disclosed keyboard disambiguation of U.S. Pat 5818437, US5953541, US6011554 and the US6286064 system that the applicant had) by the process user thump and form word and word is presented to the user related, solves text input problem with pressing keys.The word of finishing is presented to the user who starts from being pushed by button the letter of representative.Present with button and push the related one additional thump of input that speech saved the user to spell whole word of itemizing, also saved the time simultaneously.The user only need select first word of these row or roll downwards and select required word.
The word of presenting to the user is stored in the lexical data base.The United States Patent (USP) feature (sequencing feature) that the applicant had will greatly be convenient to the user and save more time, and can make the faster and input text more effectively of user.
It will be favourable that a dynamic database reordering system is provided, and this system can will push the word that is associated with button and be shown to the user sequentially according to the situation of the use word of user's reality.A not dynamic database reordering system of the frequency of storage use information in master data base is provided, and also will be favourable.
Summary of the invention
The invention provides a kind of dynamic database reordering system.The present invention pushes the word that be associated to user's demonstration with button with the order based on user's actual use word (if any).In addition, the present invention is the frequency of storage use information in master data base not, thereby needs minimum storage space.
The preferred embodiments of the present invention provide a kind of language database that comprises according to the language model word ordered, and this language model domination is presented to the ordering of user's word.The user imports thump on the keypad of communicator.In user's pressing keys input, predictive user of the present invention is being attempted word, letter, numeral or the brachymemma word (word stub) imported.The word of finishing dynamically is shown to the user, and those words start from being pushed by button the letter of representative.The user push usually with database in the keystroke sequence that above word is associated.In order to save the space of storing language database,, and then insert in the product at first to the language database presort.
The invention provides based on the rearrangement of system's operating position of user the language model order.If there is more than one word to share identical keystroke sequence, then the word of normal use is selected to be shown in the Display directory as first.If the user does not need this word, but another word that will be associated with keystroke sequence, then the user can be by pushing next button or by pressing/button that descends to roll the Display directory of the word that rolls.In case the user has found required word, the user just activates a selection key, so system is input to required word in user's the text message in the insertion point.
Preferential embodiment of the present invention follows the tracks of user's word and selects.In case a word has been selected as result's (word continues) that next button is selected, and just a frequency value is applied to this selected word, and this word is listed in the first place of the related word of this keystroke sequence by the language model in the language database.
When word for the first time by user's selections (nexted) that continue, the frequency value of the word that continues is usually less than in the Display directory the first frequency value that sorts word.When this word that continues was continued selection once more, with respect to the frequency value of the first ordering word, the frequency value of this word that continues increased.During word of each selection, all can adjust the frequency value.
Another embodiment of the present invention is searched the position of word in Display directory of continuing.Word is positioned in after second word in the Display directory if continue, and the word that then this continued rises to the second word position.The frequency that this makes the uni-gram frequency that continues be increased to the frequency of second word or be higher than second uni-gram frequency, but be lower than the frequency of first word.
When the word that continues continues same word repeatedly, the frequency value of this word that continues will become greater than the frequency value of the first demonstration word.Subsequently, user input be used to the to continue keystroke sequence of the word and the first ordering word can cause, and this word that continues is presented at by before language model first word ordered.
In one embodiment of the invention, the number of times of being selected by continuing when a word is three times (or predetermined number) of the first ordering word when above, and the frequency of this word becomes and has the word of higher frequency greater than first sort word or other.
Other aspects and advantages of the present invention will become clearer by the detailed description below in conjunction with accompanying drawing, and wherein accompanying drawing has illustrated principle of the present invention by embodiment.
Description of drawings
Fig. 1 is the synoptic diagram that has the Portable Communications Unit of simplifying keyboard in accordance with the principles of the present invention;
Fig. 2 is the synoptic diagram of cell phone keyboard in accordance with the principles of the present invention;
Fig. 3 is the block diagram of a get-it-over-and-done-with attitude in accordance with the principles of the present invention;
Fig. 4 is the block diagram of language, manufacturer, customer data base composition in accordance with the principles of the present invention;
Fig. 5 is the synoptic diagram that is stored in the frequency information in the customer data base in accordance with the principles of the present invention.
Embodiment
The present invention is implemented in dynamic database reordering system.System of the present invention shows to the user with the order based on user's actual use word (if any) pushes related word with button.In addition, the present invention is the frequency of storage use information in master data base not, thus the storage space of demand minimum.
A preferred embodiment of the present invention provides a kind of method, and this method can show the result that retrieval obtains from language database to the user, and this result and the user button on keyboard is pushed and is complementary.Give one by the definite frequency value of user's word use pattern with any word that is derived from display result by user's selection.This frequency value preferably is not stored in the host language data base, and only gives a frequency word of the actual use of user.The frequency that can not need appointment with some word of other word collision that the user uses.This method is reduced to required memory space and is 1/7 of the use required storage of usually following the tracks of the user.
Referring to Fig. 1, shown that here has a personal communicator 101 of simplifying keyboard.Keyboard disambiguation system (among U.S. Pat 5818437, US5953541, US6011554 and the US6286064 that has such as the applicant disclosed keyboard disambiguation system) has solved 102 overload of input button and on behalf of the text of several characters, one key can import problem.User's thump is processed and be presented on 103 on keyboard 102.In user's pressing keys, the word 106,107 that the system prediction user is attempting to import, letter 108, numeral 109 or brachymemma word (not shown).The word of finishing dynamically is presented on 105, and those words start from the letter that button is pushed representative.
Present and push the word catalogue that is associated with button and make the user save the additional thump of the whole word of input spelling, and saved the time.The user selects first word in the catalogue, and the catalogue of perhaps rolling is downwards also selected required word.In this embodiment, the user can push space bar 112, to accept first word in the catalogue.Selected word appears in the user version input position 104.The user can also continue pressing keys, with further reduction or weigh the selection of shown word.
The word of presenting to the user is stored in the lexical data base.If the usage frequency by each word is put in order the word in the database, so, identical word always is presented on identical ordering and gives the user on 105.
Of the present invention Language database (LDB) comprises according to the language model word ordered, the order of the word that this language model Zhi Peixiang user presents.The user pushes a sequence of the button related with above word in the database usually.
For save the storage LDB the space, LDB before the product of packing into by pre-sorting.Utilize a language model that word is sorted, this language model measurement is used for the general frequency value of each word of database.The data base manipulation frequency sorts and collects.During the compiling database, the frequency value will not store in the database with word, thereby reduces the space of storage LDB.
For example, Fig. 2 has shown typical cell phone keyboard 201.If language model indication word " in " has the frequency that is higher than " go ", when by 4 buttons 202 on this sequencing selection traditional cellular phone keyboard and 6 buttons 203, " in " appears at " go " before in Display directory so.
The invention provides according to the user operating position of system is reordered to the language model order.
Referring to Fig. 1, if the same key sequence shared in more than one word, then the most frequently used word will be used as first and select to be presented in the Display directory again.If the user does not need this word, but want another word related with keystroke sequence, so the user can by push 0 or the NEXT button or on/button (if any) that time rolls, the Display directory 105 of the word that rolls.In case the user has found required word, then the user can activate a selection key or space bar 112, so system is input to required word in user's the text message 103 in insertion point 104.
The preferred embodiments of the present invention are followed the tracks of user's word and are selected.In case a word is selected as the result that the NEXT button is selected, and just a frequency value is applied to selected word, and this word is come the first place of the word list relevant with this keystroke sequence by the language model of LDB.The word that is applied to the frequency value by language model first ordering word and is selected as NEXT button (word continues) result depends on a plurality of factors.An example factor comprises the use versatility of word with respect to the first ordering word that continue.
Continue the user and to select a word for use first, the frequency value of this word that continues is usually less than the frequency value of the first ordering word in Display directory.This word that continues is continued when selecting once more, and with respect to the frequency value of the first ordering word, the frequency value of this word that continues increases.When word of each selection, just adjust the frequency value.
An alternative embodiment of the invention is searched the position of the word that continues in Display directory.Word is positioned in after second word in the Display directory if this continues, and so the word that continues is risen to the second word position.This frequency that will continue word is increased to the frequency of second word or is higher than the frequency of second word, but is lower than the frequency of first word.Below collision problem will be discussed.
On certain point of the same word that continues in succession, the frequency value of the word that continues will become greater than the frequency value of the first demonstration word.Like this, the subsequent user input for the keystroke sequence of the continue word and the first ordering word will cause: showing this word that continues before by language model first word ordered.
In one embodiment of the invention, when the number of times of selecting a word is three times (or predetermined number) of first ordering word when above, the frequency of this word becomes greater than first the sort word or have other word of higher frequency.
In another embodiment of the present invention, the ordering of word and this word is stored in the language database, and will be stored into as the result's that continues frequency value in the independent database (database that reorders).Reorder in the database and to be higher than the ordering of word in the language database by the priority of the ordering of frequency value sign.When the frequency value was stored in the language database, the frequency that the frequency in the language database is attached most importance in the sorting data storehouse provided initial value.
Still in another embodiment of the present invention, if do not store the frequency value in the language database, then the ordering of the word in the language database is used for synthesizing the initial value of frequency in the database that reorders.
Referring to Fig. 3, there is shown the part of text processor of the present invention.The button of keyboard manager 303 monitoring users is pushed.Each button is pushed and is sent to database manager 301.Database manager 301 is collected each button and is pushed, and carries out the prediction word and handle.
The language database of database manager 301 visits on the memory storage 304 of host apparatus, and from the language database that the coupling button is pushed, extract a n word, form prediction word catalogue.The n value depends on the length of the catalogue that manufacturer or user recommend.If catalogue is oversize, the total amount that the thump quantity of the catalogue that then is used to roll will be pushed greater than the required button of typewriting input whole-word.Database manager 301 is sent to display manager 302 with the word catalogue.Display manager 302 shows the catalogue of word to the user.
When push when rolling button or selection key keyboard manager 303 notice display managers 302 at every turn.The display manager 302 utilization button that rolls is pushed suitable word in the brightness enhancing display catalogue.When the user pushed selection key, display manager 302 was inserted into selected word in user's the text entry field, and notification data librarian 301 has been selected this word.
If a word is just tracked, 301 frequency records of adjusting selected words of database manager.This frequency is provided with by each user's word use pattern.The present invention does not follow the tracks of each word of use.Some words can not bump and not need frequency is counted.The collision ordering that the present invention occurs being equal to each other because of the frequency value.This technical scheme will store usually the required memory space of frequency data reduce to the typical technology scheme required 1/7.
Display manager 301 must limit the frequency of the word of following the tracks of, because if do not limit, counting may surpass the memory capacity of register on the device.The use that system must adapt to the user changes.Aging algorithm of the present invention can be finished the calculating of having write down frequency, and the older usage frequency of deduction.Older use does not reflect user's current custom, therefore is worth low.
This part describes the study user in detail and uses pattern and generate the process that the imitation user uses the display part catalogue of pattern.
Referring to Fig. 4, the preferred embodiment that database of the present invention is implemented comprises: 402 and customer data bases of 401, one manufacturer databases of a language database (LDB) (MDB) (UDB) 403.Language database 401 is fixing preediting databases, contains the word usage frequency word ordered of measuring according to language model.
Manufacturer database 402 is customized databanks that OEM provides.Manufacturer database 402 is chosen wantonly, and is the database that has in language database 401, and OEM provides when needing an additional customized databank.
Customer data base 403 comprises user-defined a plurality of word and the database that reorders (RDB), and this database that reorders is a zone of following the tracks of the customer data base 403 of word that the user selects and these uni-gram frequencies.Fig. 5 has shown the exemplary illustration of some information 501, and these information are included in the customer data base 403 of following the tracks of the use of user's word.This information listed usage frequency, the word 504 of all letters, the word 503 of each word 502 the input keystroke sequence, be included in the object number of word in LDB505 in the language database 401.
Below everyly in text, use in the following manner:
Reorder word-of UDB selects to add to LDB word on the RDB via word.
Effective word-the UDB that has the LDB field of the current LDB of the coupling word that reorders.When following discussion was worn out with garbage collection algorithm, effectively the notion of word was important.
Deletion recently is by frequency-this is the reorder frequency of word of the nearest UDB that utilizes garbage collection algorithm to delete from RDB.
Word-the word that reorders that reorders is its frequency ends frequency more than or equal to deletion recently a word.These words will come the first place of selecting Display directory, and they comprise that UDB adds word and the UDB word that reorders.
UDB adds word-UDB interpolation word and the UDB word has identical meanings; They relate to by the user adds word on the UDB to, this word or directly add by the user, or by the detected new word that constitutes by the user of system.
Database (the RDB)-storage UDB that the reorders UDB part of word that reorders.RDB resides in the storer identical with UDB or in the storage space.
First word and non-first word-first word and non-first word all are the LDB words.With respect to LDB, first word is the first LDB word that shows, but not then not first demonstration of first word.
Database manager 301 of the present invention adds words and each UDB the reorder frequency or the usage count of word, study user's use pattern by keep following the tracks of a plurality of UDB.Use schema update and the usage count (frequency) that each word is set according to word.Usage count is utilized for refuse collection, aging algorithm determines to show the word relative position of selecting in the catalogue.
It is that this LDB word is added among the RDB that maintenance is used the first step of the tracking of pattern to the LDB word.When a word is accepted, kicks open this interpolation and handle.The key-press event of accepting a word is described below.In case a LDB word is added on the RDB, just it is called the UDB word that reorders.Table 1 has been described word has been added to rule on the RDB.In table 1, row relate to the type of selecting first object in the catalogue, and row relates to the object type of current selection.Numbering in the table relates to adds rule, as shown in the table:
Table 1-adds rule
UDB adds word The LDB word The MDB word
UDB adds word Rule _ 0 Rule _ 1 Rule _ 0
The LDB word Rule _ 3 Rule _ 2 Rule _ 3
The MDB word Rule _ 0 Rule _ 0 Rule _ 0
Based on first object type in current alternative type and the demonstration selection catalogue, this table has shown following rule:
● it is that UDB adds the word or the situation of MDB word too that the first couple in rule _ 0 catalogue as if UDB add word or MDB word and selected object.In this case, the selected object and first object can not add among the RDB in the demonstration selection catalogue.
● rule _ 1 the first couple in selecting catalogue likes LDB first word and alternative is that UDB adds under the situation of word, and the present invention adds LDB first word (showing first object of selecting in the catalogue) on the RDB to.
● rule _ 2 these rule application in show selecting catalogue the first couple as if LDB first word and selected to as if non-first word of LDB the time.This moment, these two words all were added among the RDB.
● add selected LDB word among the RDB to when selecting object to be first object selected in the catalogue of LDB word (non-first word of LDB first or LDB) and showing or UDB interpolation word or MDB word rule _ 3.
When accepting effective word, based on adding rule, this word is added to or is not added on the RDB.If this word in RDB/UDB, then promotes the frequency of this word.The incident of accepting word is described below.
● when by switching to a newspeak, registration or not registering MDB or stop database manager when changing the context of setting up the word catalogue.
● when pushing right arrow key.
● when pushing space bar, after space bar or arrow key are pressed, will cut little ice by space bar more subsequently, because this word is accepted.
● when the user imports punctuate or during with the unmatched explicit character of known word.
As mentioned above, when accepting UDB interpolation word, this word is added on the UDB.Initial frequency will be set to the value (can be set to and the irrelevant any value of first word) than high 3 usage counts of first word.When using this word, promote the frequency that UDB adds word.
The present invention impels the ordering of word by the progressive learning that the user is used pattern, uses " not invading " learning principle, and the single use of wherein non-first word will can not force first word to arrive to show the primary importance of selecting in the catalogue." not invading " principle can and be provided with frequency by careful renewal and realize.
● when first word and non-first word being added to RDB and go up, the frequency of first word is set, so that lose for first word and show to select first placement location in the catalogue, the use that three times (or predetermined number) that obtain non-first word are not offseted.The user pushes space bar and selects be considered to once use the first time of non-first word.For example, if just in time added this two words, the above use of the secondary of so non-first word will make this word move to the primary importance that shows the selection catalogue.
● all non-first words start from identical initial frequency.How long their frequency increase or minimizing will be depended on every is used them.Frequency reduces when burin-in process.
If ● from show to select catalogue, select one first word, and this first word then promotes the frequency of this word in RDB.
If ● attempt to add first word to RDB (when selecting non-first word), and first word does not then increase its frequency in RDB.Suppose that this first word has been in the tram.
● the initial frequency that UDB adds word is two usage counts that are higher than first word.Can adjust the δ value for different embodiments.
In some time interval, aging all UDB are added and the UDB word that reorders.The aging frequency that reduces these words by some factor that is meant.Carry out the aging unit that depends on the employed time interval every how long, the unit in the time interval is kept by the data base administrator.The notion because database has no time, the present invention is by safeguarding more interval computing time, New count exploration ground.When pushing space bar and select a word, this more New count just increase progressively 1.When more New count reaches a certain value, just kick open (termination) burin-in process.This process is carried out in the following manner:
● when New count more reaches its maximal value (being 1000 in the present embodiment), burin-in process is ended.In statistics, 1000 are applicable to one fast
Figure C0380078900181
The user, this user can use
Figure C0380078900182
Key in 20wpm (20 words of per minute), in 50 minutes under the state that continues input text.Just be about 50 information, every information has 20 words (~120 characters).
● the frequency of all words is reduced by aging (31/32) factor.Those skilled in the art will easily understand any aging factor can be used for realizing the rate of decay of expecting.For example, if the frequency of non-first word is 54, then it will be reduced to 52 (54 * (31/32)).
For refuse collection, " easy comeing easy go " principle is used for deleting UDB then and adding word from RDB and the UDB storage space deletion UDB word that reorders earlier.Suggestion uses factor 2 to delete the word that reorders.Like this, have in deletion before the UDB interpolation word of X frequency, the present invention at first removes frequency and is less than or equal to 2 *All of X words that reorders.The following describes garbage collection algorithm:
1, removes all UDB that put on delete flag and add word.
If 2 steps 1 are carried out the amount of space of back free space and are greater than or equal to 1/8 of total UDB data space, then carry out refuse collection.
Otherwise 3, increase deletion recently by frequency, and all UDB that all UDB with this frequency of deletion reorder word and have half this frequency add words.It is free space that this processing continues to carry out until 1/8 of UDB space.
Those skilled in the art will readily appreciate that and can adjust 1/8 threshold value, to reach the UDB free space amount of expection.
The new demonstration selects catalogue to be made up of reorder word, UDB, LDB and MDB word (depending on MDB fence (fence)) of first (from the top).The MDB fence is used for being provided with the maximum that can appear at MDB word catalogue LDB word before.This has guaranteed that OEM can show its word.Show to select the quantity of the word that reorders in the catalogue determined by frequency or the initial frequency of non-first word by nearest deletion, its quantity is always less-be referred to as the RDB counting.Its frequency adds words and the UDB word that reorders more than or equal to all UDB of RDB counting, is the word that reorders, and will appear at and show the first place of selecting in the catalogue.Illustrate by different level that below they appear at the ordering that shows in the selection catalogue:
1, its frequency is higher than all words that reorder by frequency.
2, UDB termination word.
3, reach the LDB word of MDB fence.
4, MDB termination word.
5, UDB stem.
6, MDB stem.
Reorder word when preceding 5 positions of catalogue show are selected in competition when what have an identical frequency, can relate to tie breaker.Following application of rules is in interrupting connection:
If 1 two UDB add word and are connected, the up-to-date word that then adds on the UDB is won.
Word is connected if 2 UDB add and UDB reorders, and then UDB adds the word triumph.
Word is connected if 3 two UDB reorder, and the word that then has less LDB object number is won.
Each UDB word that reorders is stored as keystroke sequence with its LDB object number.Had their information and message from which LDB, just can the reconstruct word.This technology has used storer still less to store the RDB word.A character only uses nybble, rather than every character uses one or two bytes.
A UDB reorders word with 8 bytes-two of mean consumption frequency byte, a length byte, and a LDB object number byte, a Language ID byte adds three bytes (average word is long) that are used for six characters word.The 4K byte in RDB space can be held about 512 UDB word that reorders, and the 3K byte can be held about 384 words, and the 2K byte can be held about 256 words, and the 1K byte is held about 128 words.
The user can switch on and off RDB.Its operational motion is described below:
If go back access failure RDB, then connect the RDB functional part.
Disconnect RDB and will have following two effects:
● show and select catalogue to get back to its original ordering.
● before refuse collection, any UDB ordering word among the RDB all will keep.
Those skilled in the art will readily appreciate that although used term " word " hereinbefore, the present invention is equally applicable to other Languages input unit, such as syllable (byte) or phrase (phrase).For example, in Chinese, single syllable can be imported, in Japanese, whole phrase can be imported.
Although with reference to preferred embodiment the present invention has been described, those skilled in the art will readily appreciate that, under the condition that does not deviate from spirit and scope of the invention, also can use replacement application described here with other.Therefore, the present invention is only limited by its claim.

Claims (58)

1, a kind of reordering from the method for clauses and subclauses database retrieval and that be shown to the user may further comprise the steps:
Accept the input of user from keyboard;
A language database is provided;
Wherein said language database comprises uses the language model word ordered;
From the dynamic searching word of described language database, described language database comprises the letter that forms by the input of user's keyboard;
The catalogue that shows described searching word to the user;
Wherein use language database ordering and test sequence dynamic reordering frequency counting that the word in the described catalogue is sorted;
Wherein for the word in the described catalogue, the test sequence dynamic reordering counting has precedence over the language database ordering;
Test sequence dynamic reordering frequency counting is distributed to the word that the user selects from described catalogue.
2, method according to claim 1, further comprising the steps of:
The database that reorders is provided;
Wherein after the described database that reorders is provided, described allocation step is inserted into being associated of the word selected and these words frequency that reorders in the described database that reorders.
3, method according to claim 2, further comprising the steps of:
A customer data base is provided;
Wherein after described customer data base was provided, the word that the user is input in the system especially was stored in the described customer data base; With
Wherein after described customer data base was provided, described searching step was also retrieved from described customer data base and is comprised the word of importing formed letter by user's keyboard.
4, method according to claim 2, wherein, if the user selects non-first an ordering word for the first time, then described allocation step will be inserted in the described database that reorders from the first ordering word and this non-first ordering word of described catalogue, if wherein the first ordering word is not present in the described database that reorders, then insert this first ordering word.
5, method according to claim 4, if wherein the user selects pre-determined number of the non-first ordering word, in the then described catalogue first ordering word loses its position, and this non-first ordering word is assigned with a frequency value that is higher than the first ordering word then.
6, method according to claim 4, all non-first ordering words that wherein are input in the described database that reorders are assigned with the equal frequency that reorders at first.
7, method according to claim 1 when wherein each user selects a word, just increases the frequency that reorders of this word.
8, method according to claim 1, wherein, if the user selects a word in the described catalogue, and this word is positioned under second sorting position, and then described allocation step is assigned to a value that this word is placed described catalogue second sorting position with the frequency that reorders of word.
9, method according to claim 2, further comprising the steps of:
After the described database that reorders is provided, burin-in process is carried out on the frequency period ground that reorders in the described database that reorders; With
Wherein, after the described database that reorders was provided, described aging step was utilized a frequency that reorders in the described database that reorders of predetermined factors minimizing.
10, method according to claim 3, further comprising the steps of:
After the described database that reorders is provided, periodically check the free space of the described database that reorders;
Wherein, after the described database that reorders is provided,, then from the described database that reorders, remove word with the frequency that reorders that is lower than predetermined threshold if the free space of the described database that reorders is lower than a predetermined threshold.
11, processing according to claim 10, wherein said inspection step removes user-defined word from the described database that reorders, this user definition word has the frequency that reorders that is lower than predetermined threshold, and is positioned at after other word with the frequency that reorders that is lower than this predetermined threshold.
12, method according to claim 1, further comprising the steps of:
Use language database ordering and test sequence dynamic reordering frequency counting to the ordering of the word in the described catalogue after and the test sequence dynamic reordering frequency is being counted distributing to the word that the user selects from described catalogue after, if two words have the identical frequency that reorders, then the word by up-to-date selection in these two words that at first sort solves the frequency collision of reordering in the described catalogue.
13, method according to claim 1, further comprising the steps of:
After described language database is provided, use language database ordering and test sequence dynamic reordering frequency counting to the ordering of the word in the described catalogue after and the test sequence dynamic reordering frequency is being counted distributing to the word that the user selects from described catalogue after, if two words have the identical frequency that reorders, then the frequency that reorders that solves in the described catalogue by the word that has higher ranked in the described language database that at first sorts collides.
14, method according to claim 2, further comprising the steps of:
After described language database is provided, use language database ordering and test sequence dynamic reordering frequency counting to the ordering of the word in the described catalogue after and the test sequence dynamic reordering frequency is being counted distributing to the word that the user selects from described catalogue after, if have under the case of collision at the user definition word with from the word of language database, if two words have the identical frequency that reorders, then the frequency that reorders that solves in the described catalogue by the user definition word that at first sorts collides.
15, method according to claim 1, do not possess and other word that wherein user selects are collided possible word and are regardless of counterweight ordering frequency counting.
16, a kind of reordering from the equipment of clauses and subclauses database retrieval and that be shown to the user comprises:
An acceptance is derived from the module of user's input of keyboard;
A language database;
Wherein said language database comprises uses the language model word ordered;
Module from the dynamic searching word of described language database, described language database comprise the letter that forms by the input of user's keyboard;
A module that shows the catalogue of described searching word to the user;
Wherein use language database ordering and test sequence dynamic reordering frequency counting that the word in the described catalogue is sorted;
Wherein for the word in the described catalogue, the test sequence dynamic reordering counting has precedence over the language database ordering;
A module of test sequence dynamic reordering frequency counting being distributed to the word of from described catalogue, selecting by the user.
17, equipment according to claim 16 also comprises:
The database that reorders;
Wherein said distribution module is inserted into word of selecting and the frequency that reorders that is associated with them in the described database that reorders.
18, equipment according to claim 17 also comprises:
A customer data base;
Wherein the word that is input to especially in the system of user is stored in the described customer data base; With
Wherein said retrieval module is also retrieved from described customer data base and is comprised the word of importing formed letter by user's keyboard.
19, equipment according to claim 17, wherein, if the user selects non-first an ordering word for the first time, then described distribution module will be inserted in the described database that reorders from the first ordering word and this non-first ordering word of described catalogue, if wherein the first ordering word is not present in the described database that reorders, then insert this first ordering word.
20, equipment according to claim 19, if wherein the user selects pre-determined number of the non-first ordering word, in the then described catalogue first ordering word loses its position, and this non-first ordering word is assigned with a frequency value that is higher than the first ordering word then.
21, equipment according to claim 19, all non-first ordering words that wherein are input in the described database that reorders are assigned with the equal frequency that reorders at first.
22, equipment according to claim 16 when wherein each user selects a word, just increases the frequency that reorders of this word.
23, equipment according to claim 16, wherein, if the user selects a word in the described catalogue, and this word is positioned under second sorting position, and then described distribution module is assigned to a value that this word is placed described catalogue second sorting position with the frequency that reorders of word.
24, equipment according to claim 17 also comprises:
A module of the frequency period ground that reorders in the described database that reorders being carried out burin-in process; With
Wherein, described aging module is utilized a frequency that reorders in the described database that reorders of predetermined factors minimizing.
25, equipment according to claim 17 also comprises:
The module of the free space of the described database that reorders is checked on one-period ground;
Wherein, if the free space of the described database that reorders is lower than a predetermined threshold, then from the described database that reorders, remove word with the frequency that reorders that is lower than predetermined threshold.
26, equipment according to claim 25, wherein said inspection module removes user-defined word from the described database that reorders, this user definition word has the frequency that reorders that is lower than predetermined threshold, and is positioned at after other word with the frequency that reorders that is lower than this predetermined threshold.
27, equipment according to claim 16 also comprises:
A device that solves the frequency collision of reordering in the described catalogue;
Wherein, if two words have the identical frequency that reorders, then described solution module solves a collision by up-to-date selection word in these two words that at first sort.
28, equipment according to claim 16 also comprises:
A device that solves the frequency collision of reordering in the described catalogue;
Wherein, if two words have the identical frequency that reorders, then described solution module solves a collision by the word that has higher ranked in the described language database that at first sorts.
29, equipment according to claim 16 also comprises:
A device that solves the frequency collision of reordering in the described catalogue;
Wherein, if having under the case of collision at the user definition word with from the word of language database, if two words have the identical frequency that reorders, then described solution module solves a collision by the user definition word that at first sorts.
30, equipment according to claim 16, do not possess and other word that wherein user selects collide possible word and are regardless of counterweight ordering frequency counting.
31, a kind of reordering from the method for clauses and subclauses database retrieval and that be shown to the user may further comprise the steps:
Acceptance is from user's input of keyboard;
A language database is provided;
Wherein said language database comprises the ordering frequency that uses language model and be used for each word and comes word ordered;
From the dynamic searching word of described language database, described language database comprises the letter that forms by the input of user's keyboard;
The catalogue that shows described searching word to the user;
Wherein use language database ordering and test sequence dynamic reordering frequency counting that the word in the described catalogue is sorted;
Wherein for the word in the described catalogue, the test sequence dynamic reordering counting has precedence over the language database ordering;
Test sequence dynamic reordering frequency counting is distributed to the word that the user selects from described catalogue, and the described test sequence dynamic reordering frequency counting that has distributed that will be used for a word is initialised to the ordering frequency value that obtains from the described language database that is used for this word.
32, method according to claim 31, further comprising the steps of:
The database that reorders is provided;
Wherein after the described database that reorders was provided, described allocation step was inserted into word of selecting and the frequency that reorders that is associated with it in described database that reorders.
33, method according to claim 32, further comprising the steps of:
A customer data base is provided;
Wherein after described customer data base was provided, the word that the user is input in the system especially was stored in the described customer data base; With
Wherein after described customer data base was provided, described searching step was also retrieved from described customer data base and is comprised the word of importing formed letter by user's keyboard.
34, method according to claim 33, wherein allocation step is with the equal frequency that reorders of all word initial allocation of the special input of user.
3
Figure C038007890006C1
Method according to claim 32, if wherein the user selects non-first an ordering word for the first time, then described allocation step will be inserted in the described database that reorders from the first ordering word and this non-first ordering word of described catalogue, if wherein the first ordering word is not present in the described database that reorders, then insert this first ordering word.
36, method according to claim 35, if wherein the user selects pre-determined number of the non-first ordering word, in the then described catalogue first ordering word loses its position, and this non-first ordering word is assigned with a frequency value that is higher than the first ordering word then.
37, method according to claim 31 when wherein each user selects a word, just increases the frequency that reorders of this word.
38, method according to claim 31, wherein, if the user selects a word in the described catalogue, and this word is positioned under second sorting position, and then described allocation step is assigned to a value that this word is placed described catalogue second sorting position with the frequency that reorders of word.
39, method according to claim 32, further comprising the steps of:
After the described database that reorders is provided, burin-in process is carried out on the frequency period ground that reorders in the described database that reorders; With
Wherein, after the described database that reorders was provided, described aging step was utilized a frequency that reorders in the described database that reorders of predetermined factors minimizing.
40, method according to claim 32, further comprising the steps of:
After the described database that reorders is provided, periodically check the free space of the described database that reorders;
Wherein, after the described database that reorders is provided,, then from the described database that reorders, remove word with the frequency that reorders that is lower than predetermined threshold if the free space of the described database that reorders is lower than a predetermined threshold.
41, according to the described processing of claim 40, wherein said inspection step removes user-defined word from the described database that reorders, this user definition word has the frequency that reorders that is lower than predetermined threshold, and is positioned at after other word with the frequency that reorders that is lower than this predetermined threshold.
42, method according to claim 31, further comprising the steps of:
Using after language database ordering and test sequence dynamic reordering frequency counting sort to the word in the described catalogue, and be initialised to after the ordering frequency value that the described language database that is used for this word obtains test sequence dynamic reordering frequency counting being distributed to word that the user selects and the described test sequence dynamic reordering frequency counting that has distributed that will be used for a word from described catalogue, if two words have the identical frequency that reorders, then the word by up-to-date selection in these two words that at first sort solves the frequency collision of reordering in the described catalogue.
43, method according to claim 31, further comprising the steps of:
After language database is provided, using after language database ordering and test sequence dynamic reordering frequency counting sort to the word in the described catalogue, and be initialised to after the ordering frequency value that the described language database that is used for this word obtains test sequence dynamic reordering frequency counting being distributed to word that the user selects and the described test sequence dynamic reordering frequency counting that has distributed that will be used for a word from described catalogue, if two words have the identical frequency that reorders, then the frequency that reorders that solves in the described catalogue by the word that has higher ranked in the described language database that at first sorts collides.
44, according to the described method of claim 43, further comprising the steps of:
After language database is provided, using after language database ordering and test sequence dynamic reordering frequency counting sort to the word in the described catalogue, and be initialised to after the ordering frequency value that the described language database that is used for this word obtains test sequence dynamic reordering frequency counting being distributed to word that the user selects and the described test sequence dynamic reordering frequency counting that has distributed that will be used for a word from described catalogue, if have under the case of collision at the user definition word with from the word of language database, if two words have the identical frequency that reorders, then the frequency that reorders that solves in the described catalogue by the user definition word that at first sorts collides.
45, a kind of reordering from the equipment of clauses and subclauses database retrieval and that be shown to the user comprises following:
Acceptance is from the module of user's input of keyboard;
A language database;
Wherein said language database comprises the ordering frequency word ordered of using language model and each word;
Module from the dynamic searching word of described language database, described language database comprise the letter that forms by the input of user's keyboard;
A module that shows the catalogue of described searching word to the user;
Wherein use language database ordering and test sequence dynamic reordering frequency counting that the word in the described catalogue is sorted;
Wherein for the word in the described catalogue, the test sequence dynamic reordering counting has precedence over the language database ordering;
A module of test sequence dynamic reordering frequency counting being distributed to the word of from described catalogue, selecting by the user, the described test sequence dynamic reordering frequency counting that has distributed that this module also will be used for a word is initialised to the ordering frequency value that obtains from the described language database of this word.
46, according to the described equipment of claim 45, also comprise:
The database that reorders;
Wherein said distribution module is inserted into word of selecting and the frequency that reorders that is associated with it in described database that reorders.
47, according to the described equipment of claim 46, also comprise:
A customer data base;
Wherein the word that is input to especially in the system of user is stored in the described customer data base; With
Wherein said retrieval module is also retrieved from described customer data base and is comprised the word of importing formed letter by user's keyboard.
48, according to the described equipment of claim 47, wherein said distribution module is with the equal frequency that reorders of all word initial allocation of the special input of user.
49, according to the described equipment of claim 46, if wherein the user selects non-first an ordering word for the first time, then described distribution module will be inserted in the described database that reorders from the first ordering word and this non-first ordering word of described catalogue, if wherein the first ordering word is not present in the described database that reorders, then insert this first ordering word.
50, according to the described equipment of claim 49, if wherein the user selects pre-determined number of the non-first ordering word, in the then described catalogue first ordering word loses its position, and this non-first ordering word is assigned with a frequency value that is higher than the first ordering word then.
51, according to the described equipment of claim 45, wherein each user selects a word, just increases the frequency that reorders of this word.
52, according to the described equipment of claim 45, wherein, if the user selects a word in the described catalogue, and this word is positioned under second sorting position, and then described distribution module is assigned to a value that this word is placed described catalogue second sorting position with the frequency that reorders of word.
53, according to the described equipment of claim 46, also comprise:
A module of the frequency period ground that reorders in the described database that reorders being carried out burin-in process; With
Wherein, described aging module is utilized a frequency that reorders in the described database that reorders of predetermined factors minimizing.
54, according to the described equipment of claim 46, also comprise:
The module of the free space of the described database that reorders is checked on one-period ground;
Wherein, if the free space of the described database that reorders is lower than a predetermined threshold, then from the described database that reorders, remove word with the frequency that reorders that is lower than predetermined threshold.
55, according to the described equipment of claim 54, wherein said inspection module removes user-defined word from the described database that reorders, this user definition word has the frequency that reorders that is lower than predetermined threshold, and is positioned at after other word with the frequency that reorders that is lower than this predetermined threshold.
56, according to the described equipment of claim 45, also comprise:
A module that solves the frequency collision of reordering in the described catalogue;
Wherein, if two words have the identical frequency that reorders, then described solution module solves a collision by the word of up-to-date selection in these two words that at first sort.
57, according to the described equipment of claim 45, also comprise:
A module that solves the frequency collision of reordering in the described catalogue;
Wherein, if two words have the identical frequency that reorders, then described solution module solves a collision by the word that has higher ranked in the described language database that at first sorts.
58, according to the described equipment of claim 43, also comprise:
A module that solves the frequency collision of reordering in the described catalogue;
Wherein, if having under the case of collision at the user definition word with from the word of language database, if two words have the identical frequency that reorders, then described solution module solves a collision by the user definition word that at first sorts.
CNB038007894A 2002-07-18 2003-07-17 Dynamic database reordering system Expired - Fee Related CN100428224C (en)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US39725302P 2002-07-18 2002-07-18
US60/397,253 2002-07-18
US10/621,864 2003-07-16

Publications (2)

Publication Number Publication Date
CN1703693A CN1703693A (en) 2005-11-30
CN100428224C true CN100428224C (en) 2008-10-22

Family

ID=35632692

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB038007894A Expired - Fee Related CN100428224C (en) 2002-07-18 2003-07-17 Dynamic database reordering system

Country Status (1)

Country Link
CN (1) CN100428224C (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1195409A (en) * 1995-07-26 1998-10-07 蒂吉通信系统公司 Reduced keyboard disambiguating system
US6002390A (en) * 1996-11-25 1999-12-14 Sony Corporation Text input device and method
US6223059B1 (en) * 1999-02-22 2001-04-24 Nokia Mobile Phones Limited Communication terminal having a predictive editor application
US6286064B1 (en) * 1997-01-24 2001-09-04 Tegic Communications, Inc. Reduced keyboard and method for simultaneous ambiguous and unambiguous text input

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1195409A (en) * 1995-07-26 1998-10-07 蒂吉通信系统公司 Reduced keyboard disambiguating system
US6307549B1 (en) * 1995-07-26 2001-10-23 Tegic Communications, Inc. Reduced keyboard disambiguating system
US6002390A (en) * 1996-11-25 1999-12-14 Sony Corporation Text input device and method
US6286064B1 (en) * 1997-01-24 2001-09-04 Tegic Communications, Inc. Reduced keyboard and method for simultaneous ambiguous and unambiguous text input
US6223059B1 (en) * 1999-02-22 2001-04-24 Nokia Mobile Phones Limited Communication terminal having a predictive editor application

Also Published As

Publication number Publication date
CN1703693A (en) 2005-11-30

Similar Documents

Publication Publication Date Title
US7610194B2 (en) Dynamic database reordering system
CN101133388B (en) Multiple index based information retrieval system
Belazzougui et al. Alphabet-independent compressed text indexing
CN100594470C (en) System and method for disambiguating the ambiguous input sequence of user
CN101061479B (en) Presentation of search results based on document structure
CN100334529C (en) Reduced keyboard disambiguating system
CN100589099C (en) Configurable information identification system and method
US4218760A (en) Electronic dictionary with plug-in module intelligence
JP3113814B2 (en) Information search method and information search device
CN100424703C (en) Method for obtaining newly encoded character string, input method system and word stock generation device
US20070016581A1 (en) Category setting support method and apparatus
US7555428B1 (en) System and method for identifying compounds through iterative analysis
KR20060048777A (en) Phrase-based generation of document descriptions
CN101438283A (en) Demographic based classification for local word wheeling/WEB search
CN101715579A (en) Language independent index storage system and retrieval method
CN105930546B (en) File association display method
CN115391495B (en) Method, device and equipment for searching keywords in Chinese context
CN109902090A (en) Field name acquisition methods and device
CN101449318A (en) Alphanumeric data entry apparatus and method using multicharacter keys of a keypad
US20090150375A1 (en) Detecting zero-result search queries
Okamoto et al. Discovering volatile events in your neighborhood: Local-area topic extraction from blog entries
CN100428224C (en) Dynamic database reordering system
JP3360693B2 (en) Customer information search method
CN111666398A (en) Method for searching and matching based on house source information keywords
KR101233052B1 (en) Succession chinese character input method

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20081022

Termination date: 20200717