Summary of the invention
Fundamental purpose of the present invention is, overcomes the problem that existing business classification exists, and provides a kind of new business model analytical approach and device based on self-defined classifying rules, problem to be solved to be, can be automatically to the business processing of classifying.
Object of the present invention and solve its technical matters and can adopt following technical scheme to realize.
A kind of business model analytical approach based on self-defined classifying rules proposing according to the present invention, comprising: obtain identification text;
Utilize the business model of each business objective that success has been set up to carry out business Classification and Identification to described identification text;
Wherein, the process of establishing of described business model comprises:
For an industry, collect the business objective of the sector;
Determine associated keyword corresponding to each business objective difference;
Utilize described associated keyword, for each business objective, recognition rule is set respectively, to set up the first type of business model of each business objective;
Utilize a plurality of recognition training texts to the just type training of the business model of described each business objective, to improve the recognition rule of each business objective, thereby successfully set up the business model of each business objective.
A kind of business model analytical equipment based on self-defined classifying rules proposing according to the present invention, comprising: acquisition module, identification module and set up module;
Described acquisition module, for obtaining identification text;
Described identification module, for utilizing the business model of each business objective that success has been set up to carry out business Classification and Identification to described identification text;
Wherein, described business model is set up by the described module of setting up, and the described module of setting up comprises:
Collect submodule, for for an industry, collect the business objective of the sector;
Determine submodule, for determining associated keyword corresponding to each business objective difference;
Submodule is set, for utilizing described associated keyword, for each business objective, recognition rule is set respectively, to set up the first type of business model of each business objective;
Training submodule, for utilizing a plurality of recognition training texts to the just type training of the business model of described each business objective, to improve the recognition rule of each business objective, thereby successfully sets up the business model of each business objective.
By technique scheme, business model analytical approach based on self-defined classifying rules of the present invention and device at least have following advantages and beneficial effect: the present invention is by setting up business model for each business objective, and utilize the business model of successfully setting up to carry out business Classification and Identification to identification text, can automatically and fast and accurately carry out business classification to business such as voice calls and process; Thereby improved the automaticity of service management.
Above-mentioned explanation is only the general introduction of technical solution of the present invention, in order to better understand technological means of the present invention, and can be implemented according to the content of instructions, and for above and other object of the present invention, feature and advantage can be become apparent, below especially exemplified by preferred embodiment, and coordinate Figure of description, be described in detail as follows.
Embodiment
For further setting forth the present invention, reach technological means and the effect that predetermined goal of the invention is taked, below in conjunction with accompanying drawing and preferred embodiment, to the business model analytical approach based on self-defined classifying rules proposing according to the present invention and embodiment, structure, feature and the effect thereof of device, be described in detail as follows.
Embodiment mono-, the business model analytical approach based on self-defined classifying rules.The flow process of the method as shown in Figure 1.
In Fig. 1, S100, obtain identification text.
Concrete, this identification text can be the identification text being converted by voice call, as the voice call of current ongoing voice call or recording is carried out to speech recognition, thereby obtains identification text corresponding to this voice call according to voice identification result.The present embodiment can adopt existing speech recognition technology to obtain identification text, at this, no longer speech recognition process is elaborated.
S110, utilize the business model that success has been set up to carry out business Classification and Identification to the above-mentioned identification text getting, to determine the business objective under this identification text.
Concrete, the above-mentioned identification text getting and the business model that success is set up can be carried out to matching operation, and determine the business objective under this identification text according to matching result, thereby realized, the business of identification text is classified.
A concrete example, determine the key word in identification text, and judge that according to the sequencing of business model the key word in identification text meets the degree of the recognition rule of business model, when the present embodiment can reach predetermined extent at the matching degree of key word and certain business model, determine that current business model and identification text match, and determine that this identification text belongs to current business objective corresponding to business model, no longer carry out follow-up deterministic process, thereby realized the business classification to this identification text.In addition, in above-mentioned example, also can judge respectively the matching degree of identification text and all business models, and therefrom choose the best business model of matching degree, the business objective of business objective corresponding to business model that this is selected under identifying text.Certainly, the present invention can also adopt other modes to carry out business Classification and Identification to identification text on the basis of the business model that success is set up, at this, describes in detail no longer one by one.
Described in the process of establishing of business model of the present invention is specific as follows:
S200, for an industry, collect the business objective of the sector.
Concrete, the business objective that different industries comprises also can be different, and an object lesson of business objective is that, for this industry of China Mobile, its business objective can comprise conventionally: 10086; Mend card, change card and standby card; Open an account, transfer and cancellation; Call reminding; GPRS; CRBT etc.
For a concrete industry, all business objectives in the sector can pool together with the form of business objective list.
S210, determine each business objective corresponding associated keyword respectively.
Concrete, each business objective in an industry can be to there being corresponding associated keyword, and associated keyword corresponding to different business target conventionally should be incomplete same.
A concrete example of the associated keyword that business objective is corresponding is that associated keyword that may be relevant to " GPRS " this business objective comprises: GPRS, flow, surfing Internet with cell phone and WAP etc.
S220, utilize above-mentioned associated keyword, for each business objective, recognition rule is set respectively, with the first type of the business model of setting up each business objective.This recognition rule can be called the just recognition rule of type of business model.
Concrete, the recognition rule in the present embodiment mainly comprises two parts content, i.e. logical relation between associated keyword and associated keyword.
Above-mentioned logical relation can comprise: with or, non-and in abutting connection with at least one in logical relation; Wherein:
" with " logical relation is binary logic relation, can represent by service regeulations symbol " & "; " with " priority of logical relation can be set to normal priority; " with " logical relation refers to " must occur " simultaneously, two parts must exist simultaneously and just can think and meet " with " logical relation, for example, " I want to handle broadband access network " meets " online & broadband " rule, and " I want to handle surfing Internet with cell phone " do not meet " online & broadband " rule;
"or" logical relation is binary logic relation, can represent by service regeulations symbol " | "; The priority of "or" logical relation can be set to normal priority; "or" logical relation refers to " as long as occur wherein any one ";
" non-" logical relation is monadic logic relation, can service regeulations symbol "! " represent; The priority of " non-" logical relation can be set to high priority; " non-" logical relation refers to " can not occur ", and the keyword in this logical relation must not exist;
" adjacency " logical relation is binary logic relation, can service regeulations symbol " # " represent; The priority of " adjacency " logical relation can be set to normal priority; " adjacency " logical relation refers to " in succession occurring ", before and after two parts in relation must exist simultaneously and have, order and closer distance just can be considered to meet " adjacency " relation, a concrete example of closer distance is, it is that 5 Chinese characters of two partial distances are with interior (containing 5 Chinese characters), as " I want to do a card " meets " doing # card " rule that closer distance can pre-define.
When recognition rule is set, should make recognition rule meet following regulation:
The symbol of A, characterization logic relation is half-angle character;
B, recognition rule must have and characterize the DBC case starting and characterize the DBC case finishing, and for example, recognition rule finishes with DBC case " (" starts, and with DBC case ") ";
C, in a recognition rule, other logical relations except NOT sum " in abutting connection with logical relation " can be used continuously, for example, recognition rule can be " (Global Link | M-ZONE | walk in the Divine Land) ";
D, consider the complexity of calculating, in a recognition rule, can be mutually nested except other logical relations in abutting connection with logical relation, and can only be by nested in abutting connection with logical relation; For example, article one, having mutually nested recognition rule can be " (Global Link | (dynamic & area) | walk in the Divine Land) ", and one have by nested in abutting connection with logical relation, can be " (Global Link | (dynamic # area) | walk in the Divine Land) ", be similar to " ((open-minded | opened) # (GPRS| surfing Internet with cell phone)) " such recognition rule, consideration based on computation complexity, the present embodiment can be supported;
E, in a recognition rule, in a recognition rule, the symbol by sign priority that the priority of other logical relations except NOT logic relation should be explicit represents; For example, if recognition rule is " (brand turns mutually | (turning # Global Link)) ", " turns # Global Link " and there is the priority higher than " brand turns mutually "; But if this recognition rule is write as " (brand turns mutually | turn # Global Link) ", be the ranked priority of each logical relation of order according to the order of sequence, may cause not meeting service logic;
F, bracket should occur in pairs, and the number of times that " (" and ") " occurs should be identical, otherwise can think grammar mistake;
The symbol should between the adjacent associated keyword of G, two front and back with characterization logic relation, must subsistence logic relational symbol between any two adjacent keywords;
H, some spaces can be set in the front and back of logical relation symbol, to strengthen the readability of recognition rule;
I, can in recognition rule, insert the symbol without actual logic meaning, to strengthen the readability of recognition rule, as can be inserted " (" and ") " without actual logic meaning in recognition rule.
After being provided with recognition rule for each business objective, the business model of each business objective just type has also just successfully been set up.
S230, utilize a plurality of recognition training texts to the just type training of the business model of each business objective, to improve the recognition rule of each business objective, thereby form the business model of each business objective.The process of training is specific as follows:
Training step 1, first utilize single business model just type a plurality of recognition training texts identifys to checking, and according to identifying the result, a plurality of recognition training texts are screened; Utilize a plurality of recognition training texts (as being no less than the recording identification text of 100,000) to the just type training of certain business model, thereby all recognition training text areas can be divided into, meet the recognition training text of the first type of this business model and do not meet the just recognition training text of type of this business model.To training step 2.
Training step 2, the result obtaining according to above-mentioned screening are revised just type of this single business model.To training step 3.
Concrete, if certain recognition training text should meet just type of this business model, and by after above-mentioned training step 1, this recognition training text does not meet just type of this business model, should supplement the just recognition rule of type of this business model according to this recognition training text; If certain recognition training text should not meet just type of this business model, and by after above-mentioned training step 1, this recognition training text meets just type of this business model, should adjust the just recognition rule of type of this business model according to this recognition training text.
Training step 3, repeat above-mentioned training step 1-2, with one by one to all business models just type revise, revise the just recognition rule of type of all business models.To training step 4.
Training step 4, utilize current revised all business models just type a plurality of recognition training texts (as being no less than the recording identification text of 100,000) are identified to checking, to obtain the recognition training text of two types,, a, can not by current revised all business models just type all identify the recognition training text of checking, b, meet the just recognition training text of type of a plurality of current revised business models simultaneously; And continue to revise just type of corresponding business model according to this recognition training text of two types.
Concrete, if certain recognition training text should meet just type of current a certain business model, yet, in the training of this step, this recognition training text does not meet just type of arbitrary business model, it should supplement the just recognition rule of type of corresponding business model according to this recognition training text, so that can meet just type of a certain business model; If certain recognition training text should not meet just type of a plurality of business models simultaneously, and in the training of this step, this recognition training text has met just type of a plurality of business models simultaneously, should adjust the just recognition rule of type of corresponding business model according to this recognition training text, so that it only meets just type of a certain business model.
Training step 5, repeat to call above-mentioned training step 4, until the quantity of the quantity of two kinds of recognition training texts of current acquisition and two kinds of recognition training texts that last time obtained all identical (both do not increased also and do not reduced), thereby using the business model that each business objective is current, just type is as the business model of each business objective, and the recognition rule of the current first type of business model is the recognition rule of business model.
The recognition rule of the business model obtaining by above-mentioned training should comprise can fully describe business objective corresponding to this business model existence certain/keyword sequence of some logical relation, that is to say, the recognition rule of business model has been described keyword and the logical relation thereof that may occur in other the actual call of certain specific transactions target class.
Business objective classification can refer to, expectation is by the attribute under a regular call that can identify of corresponding business model, and this attribute can be Business Name, action type and self-defining thematic attribute etc. (as complained early warning or the preferential activity in campus etc.).
The object lesson of explaining a business objective classification and corresponding recognition rule is as follows:
Information service class->GPRS:((GPRS| flow | surfing Internet with cell phone | WAP)).
Embodiment bis-, the business model analytical equipment based on self-defined classifying rules, the structure of this device as shown in Figure 2.
In Fig. 2, this device mainly comprises: acquisition module 1, identification module 2 and set up module 3.
Acquisition module 1 is mainly used in obtaining identification text.This identification text can be the identification text being converted by voice call, as speech recognition is carried out in the voice call of 11 pairs of current ongoing voice calls of the speech recognition submodule in acquisition module 1 or recording, thereby speech recognition submodule 11 can obtain identification text corresponding to this voice call according to voice identification result.Speech recognition submodule 11 can adopt existing speech recognition technology to obtain identification text, at this, no longer describes in detail.
The identification text that identification module 2 is mainly used in utilizing the business model of each business objective that success has been set up to get above-mentioned acquisition module 1 carries out business Classification and Identification.Identification module 2 can carry out matching operation by the above-mentioned identification text getting and the business model that success is set up, and determines the business objective under this identification text according to matching result, thereby realized, the business of identification text is classified.
A concrete example, identification module 2 is determined the key word in identification text, and judge that according to the sequencing of business model the key word in identification text meets the degree of the recognition rule of business model, when identification module 2 can reach predetermined extent at the matching degree of key word and certain business model, determine that current business model and identification text match, and determine that this identification text belongs to current business objective corresponding to business model, identification module 2 no longer carries out follow-up deterministic process, thereby has realized the business classification to this identification text.In addition, in above-mentioned example, identification module 2 also can judge respectively the matching degree of identification text and all business models, and therefrom chooses the best business model of matching degree, the business objective of business objective corresponding to business model that this is selected under identifying text.Certainly, identification module 2 can also adopt other modes to carry out business Classification and Identification to identification text on the basis of the business model that success is set up, at this, describes in detail no longer one by one
Set up module 3 and be mainly used in setting up business model.By the business model of setting up module 3 success foundation, can be stored in and set up in module 3, also can be stored in identification module 2, can also be stored in and be independent of identification module 2 and set up in the memory module of module 3.
Setting up module 3 mainly comprises: collect submodule 31, determine submodule 32, submodule 33 and training submodule 34 are set.
Collecting submodule 31 is mainly used in, for an industry, collecting the business objective of the sector.Collecting submodule 31 can the form with business objective list pool together all business objectives in the sector.
Determine that submodule 32 is mainly used in determining associated keyword corresponding to each business objective difference.Determine that the associated keyword that submodule 32 arranges for each business objective in the sector conventionally should be incomplete same.
Submodule 33 is set and is mainly used in utilizing above-mentioned associated keyword, for each business objective, recognition rule is set respectively, to set up the first type of business model of each business objective.Regulation that content that recognition rule that submodule 33 arranges mainly comprises and recognition rule should meet etc. is set concrete as the description in above-described embodiment one, is not repeated.The set recognition rule of submodule 33 is set can be specially: information service class->GPRS:((GPRS| flow | surfing Internet with cell phone | WAP)).
Arranging after submodule 33 is provided with recognition rule for each business objective, the business model of each business objective just type has also just successfully been set up.
Training submodule 34 is mainly used in utilizing a plurality of recognition training texts to the just type training of the business model of each business objective, to improve the recognition rule of each business objective, thereby successfully sets up the business model of each business objective.
Above-mentioned training submodule 34 mainly comprises: first module, second unit, Unit the 3rd, Unit the 4th and Unit the 5th.
First module is mainly used in utilizing the first type of single business model to identify checking to a plurality of recognition training texts, so that a plurality of recognition training texts are screened.
Second unit is mainly used according to the first type of this single business model of the selection result correction of first module.
Unit the 3rd is mainly used in repeating to call first module and second unit is carried out corresponding operation, one by one the first type of all business models is revised.
Unit the 4th is mainly used in utilizing the first type of current revised all business models to identify checking to a plurality of recognition training texts, to obtain two kinds of recognition training texts, can not and meet the just recognition training text of type of a plurality of current revised business models by the recognition training text of just type identification of current revised all business models checking simultaneously, and continue to revise just type of corresponding service model according to these two kinds of recognition training texts.
Unit the 5th is mainly used in repeating to call Unit the 4th, until the quantity of two kinds of recognition training texts of current acquisition is identical respectively with the quantity of two kinds of recognition training texts that last time obtained, thereby successfully sets up the business model of each business objective.
The performed concrete operations in first module to the Unit five refer to the description of the training step 1-5 in above-described embodiment one, at this, no longer describe in detail.
Each embodiment in this instructions all adopts the mode of going forward one by one to describe, between each embodiment identical similar part mutually referring to, each embodiment stresses is the difference with other embodiment.Especially, for device embodiment, because it is substantially similar in appearance to embodiment of the method, so describe fairly simplely, relevant part is referring to the part explanation of embodiment of the method.Device embodiment described above is only schematic, wherein said module and unit as separating component explanation can or can not be also physically to separate, the parts that show as unit can be or can not be also physical locations, can be positioned at a place, or also can be distributed in a plurality of network element.Can select according to the actual needs some or all of module wherein to realize the object of the present embodiment scheme.Those of ordinary skills, in the situation that not paying creative work, are appreciated that and implement.
The above is only preferred embodiment of the present invention, not the present invention is done to any pro forma restriction, although the present invention discloses as above with preferred embodiment, yet not in order to limit technology of the present invention, any those skilled in the art are not departing within the scope of technical solution of the present invention, when can utilizing the technology contents of above-mentioned announcement to make a little change or being modified to the equivalent embodiment of equivalent variations, in every case be the content that does not depart from technical solution of the present invention, any simple modification of above embodiment being done according to technical spirit of the present invention, equivalent variations and modification, all still belong in the scope of technical solution of the present invention.