CN108280089A - Identify the method and apparatus sent a telegram here extremely - Google Patents
Identify the method and apparatus sent a telegram here extremely Download PDFInfo
- Publication number
- CN108280089A CN108280089A CN201710010573.6A CN201710010573A CN108280089A CN 108280089 A CN108280089 A CN 108280089A CN 201710010573 A CN201710010573 A CN 201710010573A CN 108280089 A CN108280089 A CN 108280089A
- Authority
- CN
- China
- Prior art keywords
- value
- key feature
- contrast characteristic
- current call
- feature
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/90335—Query processing
- G06F16/90344—Query processing by using string matching techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/42—Systems providing special services or facilities to subscribers
- H04M3/436—Arrangements for screening incoming calls, i.e. evaluating the characteristics of a call before deciding whether to answer it
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mobile Radio Communication Systems (AREA)
- Telephonic Communication Services (AREA)
Abstract
The purpose of the application is to provide a kind of method and apparatus for identifying and sending a telegram here extremely, to solve the problems, such as that the reliability identified in the prior art needs to rely on perfect database.Scheme provided by the present application is without establishing presetting database, by the identification feature value for obtaining the key feature of current call, with by search engine retrieving to contrast characteristic's value of the key feature be compared, finally according to the comparing result of the identification feature value and contrast characteristic's value of the key feature, it may be determined that whether the current call is abnormal incoming call.Since contrast characteristic's value for being compared is directly to retrieve acquisition by search engine, it is not rely on preset database, therefore in Non-precondition database or not perfect presetting database, higher accuracy can also be reached.
Description
Technical field
This application involves information technology field more particularly to a kind of schemes that identification is sent a telegram here extremely.
Background technology
In the big data epoch, for some reason, for example account lacks safeguard protection or leakage of information, causes commonly to disappear
The information of expense person is easy to be obtained by criminal, to carry out the target of phone fraud as criminal.
By taking the reservation platform of certain show ticket as an example, criminal may pretend to be this after getting the information of consumer
The staff of platform pretends to claim to be upgraded to VIP due to the account of the maloperation client of staff, and induction consumer carries out silver
Not so row card unbundlings operation can be associated with by Unionpay and withhold.
Many consumers were harassed by such fraudulent call.In view of the radix of customer count is huge, it more difficult to exempt from meeting
Someone has dust thrown into the eyes so that criminal succeeds repeatedly.For avoid it is such happen, prior art scheme includes mainly at present
It is several below:Dependency of incoming call, incoming number similitude ratio of the Application on Voiceprint Recognition to realize accurately authentication, show number
To the similitude of, dialog context than equity.
It in scheme of the Application on Voiceprint Recognition to realize accurately authentication, needs to establish database to the vocal print of caller, profit
Accurate sound-groove model is established with unique physiological characteristic, may be implemented the identification of fast accurate, but each sound
Line only corresponds to a caller, needs to establish huge database, is possible to realization and accurately identifies.
The scheme for showing the dependency of incoming call of number, can realize its ownership place so that answer according to the number of caller
Person can do a simple judgement according to its ownership place to the number.For the unreasonable number of ownership place, those who answer can recognize
May be fraudulent call for current call.But the mode of such judgement is more rough, and presently, there are some to change
The technology of incoming number can evade such scheme completely, therefore reliability is not high.
In the scheme that incoming number similitude compares, after a user has connect incoming call, the number of the incoming call is marked a
The label of property, such as " take-away ", " express delivery " of normal label, abnormal label such as " cheat ", " swindle ".Other users are again
Necessary prompting can be provided when being connected to the incoming call of identical number, detailed process is:The terminal device of other users is listening to
When calling event, inquire whether the corresponding calling number of the calling event is labeled abnormal label to server, in root
When determining that the calling number is labeled abnormal label (such as swindle) according to query result, pass through the side of voice or picture and text
Formula provides prompt message on the terminal device.Such scheme equally relies on a perfect database, labeled in database
Number is more, and the accuracy of identification is also higher.But in actual scene, criminal often replaces incoming number, for not
The new digit that method molecule uses, such mode can not be identified effectively.
It in the scheme that the similitude of dialog context compares, is parsed by the call voice to incoming call, extraction is relevant
Keyword it is compared with the content in preset database, to determine whether abnormal telephone fraud.For example, meter
The likelihood for calculating keyword and corresponding contents in database, judges whether it is greater than or equal to predetermined threshold, if the judgment is Yes,
It is fraudulent call then to remind user's current talking.Such scheme also relies on the degree of perfection of preset database, due to difference
Keyword involved by the swindle way of type is different, and exhaustive realization difficulty is larger, can not be to using new fraudulent side
The fraudulent call of formula is identified.
It follows that the scheme of existing identification fraudulent call, the reliability of identification depends on perfect database, only
It just may be implemented accurately to identify when database more improves.
Apply for content
The purpose of the application is to provide a kind of method and apparatus for identifying and sending a telegram here extremely, to solve in the prior art
The reliability of identification needs the problem of relying on perfect database.
To achieve the above object, the method sent a telegram here extremely is identified this application provides a kind of, the method includes:
Obtain the identification feature value of the key feature of current call;
Pass through contrast characteristic's value of key feature described in search engine retrieving;
According to the comparing result of the identification feature value and contrast characteristic's value of the key feature, determine that the current call is
It is no to send a telegram here to be abnormal.
Another aspect based on the application, additionally provides a kind of equipment for identifying and sending a telegram here extremely, and the equipment includes:
Identification module, the identification feature value of the key feature for obtaining current call;
Module is retrieved, for contrast characteristic's value by key feature described in search engine retrieving;
Matching module is used for the comparing result of the identification feature value and contrast characteristic's value according to the key feature, determines
Whether the current call is abnormal incoming call.
In addition, the embodiment of the present application also provides the equipment that another identification is sent a telegram here extremely, the equipment includes:
Processor;And
It is arranged to the memory of storage computer executable instructions, the executable instruction makes the place when executed
Manage device:The identification feature value for obtaining the key feature of current call, by key feature described in search engine retrieving to bit
Value indicative, and identification feature value and contrast characteristic's value according to the key feature comparing result, determine the current call
Whether it is abnormal incoming call.
Compared with prior art, scheme provided by the present application is without establishing presetting database, by obtaining current call
The identification feature value of key feature, with by search engine retrieving to contrast characteristic's value of the key feature be compared,
Finally according to the comparing result of the identification feature value and contrast characteristic's value of the key feature, it may be determined that the current call is
It is no to send a telegram here to be abnormal.Since contrast characteristic's value for being compared is directly to retrieve acquisition by search engine, it is not relying on
In preset database, therefore in Non-precondition database or not perfect presetting database, higher standard can also be reached
True property.
Description of the drawings
By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, the application's is other
Feature, objects and advantages will become more apparent upon:
Fig. 1 is a kind of process chart identifying the method sent a telegram here extremely provided by the embodiments of the present application;
Fig. 2 is a kind of structural schematic diagram identifying the equipment sent a telegram here extremely provided by the embodiments of the present application;
Fig. 3 is the interaction signal between each equipment when identifying that exception is sent a telegram here using scheme provided by the embodiments of the present application
Figure;
Fig. 4 is that the equipment that the identification of the embodiment of the present application is sent a telegram here extremely is identifying handling principle figure when sending a telegram here extremely;
Fig. 5 is the structural schematic diagram for the equipment that another identification provided by the embodiments of the present application is sent a telegram here extremely;
Same or analogous reference numeral represents same or analogous component in attached drawing.
Specific implementation mode
The application is described in further detail below in conjunction with the accompanying drawings.
In a typical configuration of this application, terminal, the equipment of service network include one or more processors
(CPU), input/output interface, network interface and memory.
Memory may include computer-readable medium in volatile memory, random access memory (RAM) and/or
The forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable medium
Example.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media, can be by any side
Method or technology realize information storage.Information can be computer-readable instruction, data structure, the module of program or other numbers
According to.The example of the storage medium of computer includes, but are not limited to phase transition internal memory (PRAM), static RAM
(SRAM), dynamic random access memory (DRAM), other kinds of random access memory (RAM), read-only memory
(ROM), electrically erasable programmable read-only memory (EEPROM), fast flash memory bank or other memory techniques, CD-ROM (CD-
ROM), digital versatile disc (DVD) or other optical storages, magnetic tape cassette, magnetic tape disk storage or other magnetic storages
Equipment or any other non-transmission medium can be used for storage and can be accessed by a computing device information.
The embodiment of the present application provides a kind of method for identifying and sending a telegram here extremely, and the process flow of this method is as shown in Figure 1, packet
Include following steps:
Step S101 obtains the identification feature value of the key feature of current call.Key feature can be used for sentencing
Whether disconnected current call is the information sent a telegram here extremely, for example, the number of current call, caller inform identity, address, department
Etc. information, identification feature value be refer to the aforementioned key feature identified by all kinds of identification methods value.For example, to Mr. Yu
The identification feature value of the key feature got of once sending a telegram here may include:Incoming number is that+87095555, identity is AA silver
Row, address are the cities the BB roads CC 100, department is DD offices.
In actual scene, the identification feature value of most of key feature can be got by way of speech recognition,
Speech recognition after electrically connecting, is being proceeded by, is obtaining voice content of the caller in communication process, in these voices
The identification feature value of key feature is extracted in appearance.And the information of incoming number, since caller will not generally converse
It says, then can be got in phone incoming call in journey.As a result, in one embodiment of the application, in terminal devices such as mobile phones
When the upper realization above method, calling event monitoring can be carried out on the terminal device, when detecting incoming call, start corresponding work(
Can unit be identified, such as start incoming call recognition unit to extract incoming number, and start sound identification module extraction and come
The information such as electric person's identity, address, department.
Step S102 passes through contrast characteristic's value of search engine retrieving key feature.Search used in the present embodiment
Engine can be several search engines commonly used in the trade, such as Google, refreshing horse etc., by calling search engine to key feature into
Row retrieval, the retrieval result got is contrast characteristic's value of key feature.
During actual retrieval, corresponding retrieval type can be built according to key feature, to obtain corresponding retrieval
As a result.Still by taking aforementioned several key features as an example, if identity is as main key feature, contrast characteristic's value is directly using knowledge
The identification feature value not obtained, i.e. " AA banks ".And when retrieving other key features, build corresponding retrieval in conjunction with identity
Word, such as " addresses AA banks and ", " AA banks and phones " and " and departments of AA banks " etc..Thus, it is possible to get as
Under contrast characteristic's value be identity be AA banks, the incoming number of AA banks is 95555, the address of AA banks is the cities BB CC123
Road 500, AA banks subordinate department include DD offices.
In order to ensure the real-time of identification judgement, search engine retrieving can be passed through in the communication process of current call
Contrast characteristic's value of key feature.For example, in communication process, if passing through the knowledge that speech recognition to identity is AA banks and address
When other characteristic value, you can contrast characteristic's value of structure retrieval type inquiry AA banks address, rather than complete to converse, recognize institute
It is uniformly handled after the identification feature value for having key feature.
In addition, to ensure that contrast characteristic's value that retrieval obtains can be consistent with real information, retrieval result can be carried out
Screening obtains data as reliable as possible, such as is screened according to the source of retrieval result, preferentially adopts the letter from official website
Breath, only adopts the information of unofficial website in feature, for example, can not obtain information from official website or
The credit rating of unofficial website is higher etc..
Step S103 determines current call according to the comparing result of the identification feature value and contrast characteristic's value of key feature
Whether it is abnormal incoming call.
By taking several key features involved in the present embodiment as an example, in comparing result, the incoming number of current call,
The comparing result of this two key features of location is inconsistent, and Ling Liangxiang key features department is consistent with identity.According to comparing result,
Current call can be judged, specific judgment mode can be specifically arranged according to different application scene.
Such as under actual scene, if normal incoming call, the information that caller provides should be consistent with truth completely,
The identification feature value with contrast characteristic's value of all key features should be completely the same at this time, and the fraudulent call of criminal
In, it generally has one or more key feature and is not inconsistent with actual conditions.Therefore, in a kind of embodiment of the application, into
If row (such as mail returned on ground of incorrect address, is not present there are the identification feature value of a key feature is different from contrast characteristic's value when judging
The department etc. that electric person is provided) when, it is determined that current call is abnormal sends a telegram here.
In another embodiment of the application, it can also be judged in the following way:
First, the similarity of the identification feature value and contrast characteristic's value of key feature is determined;Then, according to key feature
Similarity determines that current call is abnormal sends a telegram here.
Similarity refers to the degree of closeness of the identification feature value and contrast characteristic's value of key feature, if the two is completely the same,
Then similarity highest can calculate the similarity of the two if the two is distinct according to corresponding calculation.For example,
Under type such as may be used and calculate similarity:
Wherein, a, b indicate that the identification feature value and contrast characteristic's value of key feature, s (a, b) indicate similarity respectively.It is logical
Crossing aforesaid way more can easily calculate similarity, i.e. identification feature value is identical as contrast characteristic's value, then similarity is 1,
Identification feature value is differed with contrast characteristic's value, then similarity is 0.
In actual scene, for differing identification feature value and contrast characteristic's value, the degree of its difference is if desired determined,
In order to obtain with accurate similarity, then following mode may be used and calculate similarity:
S (a, b)=1-L(|a|,|b|)(a,b)/Max(|a|,|b|)
Wherein, a, b indicate the identification feature value and contrast characteristic's value of key feature respectively, | a |, | b | indicate the character of a, b
String length, s (a, b) indicate similarity, L(|a|,|b|)(a, b) indicates the Levenshtein distances (editing distance) between a, b,
Max (x, y) indicates that the maximum value in x and y, Min (x, y) expressions is taken to take minimum value in x and y.Wherein, Levenshtein distances
L(|a|,|b|)(a, b) refers between the character string of two determinant attribute contents of a, b, and the minimum volume needed for another is changed by one
Number of operations is collected, character string can be Chinese character string, can also be the character string of other words such as English, Levenshtein
The specific formula for calculation of distance is as follows:
The calculation formula is iterative formula, L(i,j)(a, b) indicates the preceding i character of a and the preceding j character of b
The maximum value of Levenshtein distances, i is the string length of a | a |, the maximum value of j is the string length of b | b |,
aiIndicate i-th of character of a, biIndicate i-th of character of b.
Here, those skilled in the art should be understood that the mode of above-mentioned calculating similarity is only for example, other are existing
Or the mode of other calculating similarities that is likely to occur from now on be such as applicable to the application, should also be included in the application and protect model
Within enclosing, and it is incorporated herein by reference herein.
When determining that current call is abnormal incoming call according to the similarity of key feature, a threshold value can be set, according to key
The similarity calculation of feature goes out a fiducial value, and fiducial value and threshold value are compared, if being less than threshold value, then it represents that key feature
Identification feature value and truth difference it is larger, therefore, it is determined that current call be it is abnormal send a telegram here, that is, be likely to be swindle electricity
Words;If being less than threshold value, then it represents that the identification feature value and truth difference of key feature are little, therefore, it is determined that current call
Normally to send a telegram here, i.e., current call is that the possibility of fraudulent call is smaller.
For fiducial value, specific calculation can directly calculate the arithmetic average of every key feature similarity,
Can also be after weight is set for every key feature, to calculate weighted average etc..Its specific calculation is in this application
It does not limit, suitable mode can be used according to concrete application scene.
In another implementation of the application, need to ensure in the communication process of current call, according to the knowledge of key feature
The comparing result of other characteristic value and contrast characteristic's value determines whether current call is abnormal incoming call.Thereby it is ensured that recognition result
Real-time avoids those who answer's fraudulent call when just knowing this call after having answered phone, causes unnecessary loss.
After determining recognition result according to similarity, since the case where being judged according to similarity, is based on setting based on experience value
Fixed threshold value, therefore recognition result is it is difficult to ensure that complete correct.Thus, it is possible to determining that current call is that exception sends a telegram here it
Afterwards, it is obtained and is identified according to the source of the similarity and/or contrast characteristic's value of the identification feature value of key feature and contrast characteristic's value
As a result confidence level.Wherein, confidence level indicates the current correct probability of recognition result, and final result can be identification as a result,
As a result it is abnormal incoming call, confidence level 90%.
For the confidence level of recognition result, relative factor includes at least similarity and the contrast characteristic of key feature
The source of value can be calculated in the embodiment of the present application based on a single one factor, can also combine two kinds of factors one
It rises and calculates.For example, by taking the similarity of key feature as an example, similarity is higher, and confidence level also can be higher.In conjunction with aforementioned comparison's value
Calculation, the similarity of every key feature is higher, and the reduced value being calculated accordingly can improve therewith, it is possible thereby to root
Confidence level is calculated according to the difference of reduced value and threshold value, if reduced value is more than that threshold value is bigger, and confidence level is higher, otherwise confidence level is lower.
About the source of contrast characteristic's value, since the source of contrast characteristic's value in the retrieval result of search engine includes a variety of
Possibility, such as come from official website or other all kinds of websites, the reliability of contrast characteristic's value of various separate sources not phase
Together, under normal circumstances, the reliability in official website source can be higher than other websites.Therefore, in contrast characteristic's value of every key feature
The quantity for coming from official website is more, then the confidence level of final recognition result is also higher.
It is the number sent a telegram here extremely for recognition result, it can be reported, such as can is to concrete mode:To
Specific number either address (such as operator, police etc.) sent in a manner of voice, short message, mail etc. report information or
By the label personalized to the number of incoming call mark, it is sent to specific database so that can when being connected to identical number again
To directly give necessary prompt etc..
In another embodiment of the application, in the communication process of current call, at least one of following letter can be shown
Breath so that those who answer can check these information, the case where to understand current call in real time.These information include:Key feature
Identification feature value, contrast characteristic's value of key feature, the comparing result of key feature, the recognition result of current call and knowledge
The confidence level etc. of other result.By taking mobile phone as an example, in communication process, the screen of mobile phone directly displays any one in above- mentioned information
Kind or it is several, thus those who answer can at any time check in communication process, and to current call the case where carries out certain
Solution.For example, can be used by the form of text, the identification feature value and contrast characteristic's value of real-time display key feature in mobile phone
Family by sliding screen selectivity can check, be judged in conjunction with actual conditions oneself, and not depend on sentencing for mobile phone
Disconnected result.
Based on same inventive concept, the equipment that identification is sent a telegram here extremely is additionally provided in the embodiment of the present application, which corresponds to
Method be method that the identification in previous embodiment is sent a telegram here extremely, and its principle solved the problems, such as is similar to this method.
A kind of structure identifying the equipment sent a telegram here extremely provided by the embodiments of the present application is as shown in Fig. 2, the equipment can be
The terminal devices such as mobile phone include at least identification module 210, retrieval module 220 and matching module 230.Wherein, identification module
The identification feature value of 210 key feature for obtaining current call.Module 220 is retrieved to be used for through search engine retrieving key
Contrast characteristic's value of feature.Matching module 230 is used for the comparison knot of the identification feature value and contrast characteristic's value according to key feature
Fruit determines whether current call is abnormal incoming call.
Key feature can be used for judging whether current call is the information sent a telegram here extremely, such as current call
The information such as identity, address, department that number, caller inform, identification feature value refer to being identified by all kinds of identification methods
The value of the aforementioned key feature arrived.For example, may include for the identification feature value of certain key feature got of once sending a telegram here:
Incoming number be+87095555, identity be AA banks, address is the cities the BB roads CC 100, department is DD offices.
In actual scene, the identification feature value of most of key feature can be got by way of speech recognition,
Speech recognition after electrically connecting, is being proceeded by, is obtaining voice content of the caller in communication process, in these voices
The identification feature value of key feature is extracted in appearance.And the information of incoming number, since caller will not generally converse
It says, then can be got in phone incoming call in journey.As a result, in one embodiment of the application, in terminal devices such as mobile phones
When the upper realization above method, calling event monitoring can be carried out on the terminal device, when detecting incoming call, start corresponding work(
Can unit be identified, such as start incoming call recognition unit to extract incoming number, and start sound identification module extraction and come
The information such as electric person's identity, address, department.
Search engine used in the present embodiment can be several search engines commonly used in the trade, such as Google, refreshing horse
Deng being retrieved to key feature by calling search engine, the retrieval result got is the contrast characteristic of key feature
Value.
During actual retrieval, corresponding retrieval type can be built according to key feature, to obtain corresponding retrieval
As a result.Still by taking aforementioned several key features as an example, if identity is as main key feature, contrast characteristic's value is directly using knowledge
The identification feature value not obtained, i.e. " AA banks ".And when retrieving other key features, build corresponding retrieval in conjunction with identity
Word, such as " addresses AA banks and ", " AA banks and phones " and " and departments of AA banks " etc..Thus, it is possible to get as
Under contrast characteristic's value be identity be AA banks, the incoming number of AA banks is 95555, the address of AA banks is the cities BB CC123
Road 500, AA banks subordinate department include DD offices.
In order to ensure the real-time of identification judgement, retrieval module 220 can be in the communication process of current call, by searching
Index holds up contrast characteristic's value of retrieval key feature.For example, in communication process, if being AA banks by speech recognition to identity
When with the identification feature value of address, you can contrast characteristic's value of structure retrieval type inquiry AA banks address, rather than complete it is logical
Words, the identification feature value for recognizing all key features are uniformly handled later.
In addition, to ensure that contrast characteristic's value that retrieval obtains can be consistent with real information, retrieval result can be carried out
Screening obtains data as reliable as possible, such as is screened according to the source of retrieval result, preferentially adopts the letter from official website
Breath, only adopts the information of unofficial website in feature, for example, can not obtain information from official website or
The credit rating of unofficial website is higher etc..
By taking several key features involved in the present embodiment as an example, in comparing result, the incoming number of current call,
The comparing result of this two key features of location is inconsistent, and Ling Liangxiang key features department is consistent with identity.According to comparing result,
Current call can be judged, specific judgment mode can be specifically arranged according to different application scene.
Such as under actual scene, if normal incoming call, the information that caller provides should be consistent with truth completely,
The identification feature value with contrast characteristic's value of all key features should be completely the same at this time, and the fraudulent call of criminal
In, it generally has one or more key feature and is not inconsistent with actual conditions.Therefore, in a kind of embodiment of the application, matching
If there are the identification feature value of a key feature, different from contrast characteristic's value (such as address is not when being judged for module 230
Symbol, there is no the departments etc. that caller is provided) when, it is determined that current call is abnormal sends a telegram here.
In another embodiment of the application, matching module 230 can also be judged in the following way:
First, the similarity of the identification feature value and contrast characteristic's value of key feature is determined;Then, according to key feature
Similarity determines that current call is abnormal sends a telegram here.
Similarity refers to the degree of closeness of the identification feature value and contrast characteristic's value of key feature, if the two is completely the same,
Then similarity highest can calculate the similarity of the two if the two is distinct according to corresponding calculation.For example,
Under type such as may be used and calculate similarity:
Wherein, a, b indicate that the identification feature value and contrast characteristic's value of key feature, s (a, b) indicate similarity respectively.It is logical
Crossing aforesaid way more can easily calculate similarity, i.e. identification feature value is identical as contrast characteristic's value, then similarity is 1,
Identification feature value is differed with contrast characteristic's value, then similarity is 0.
In actual scene, for differing identification feature value and contrast characteristic's value, the degree of its difference is if desired determined,
In order to obtain with accurate similarity, then following mode may be used and calculate similarity:
S (a, b)=1-L(|a|,|b|)(a,b)/Max(|a|,|b|)
Wherein, a, b indicate the identification feature value and contrast characteristic's value of key feature respectively, | a |, | b | indicate the character of a, b
String length, s (a, b) indicate similarity, L(|a|,|b|)(a, b) indicates the Levenshtein distances (editing distance) between a, b,
Max (x, y) indicates that the maximum value in x and y, Min (x, y) expressions is taken to take minimum value in x and y.Wherein, Levenshtein distances
L(|a|,|b|)(a, b) refers between the character string of two determinant attribute contents of a, b, and the minimum volume needed for another is changed by one
Number of operations is collected, character string can be Chinese character string, can also be the character string of other words such as English, Levenshtein
The specific formula for calculation of distance is as follows:
The calculation formula is iterative formula, L(i,j)(a, b) indicates the preceding i character of a and the preceding j character of b
The maximum value of Levenshtein distances, i is the string length of a | a |, the maximum value of j is the string length of b | b |,
aiIndicate i-th of character of a, biIndicate i-th of character of b.
Here, those skilled in the art should be understood that the mode of above-mentioned calculating similarity is only for example, other are existing
Or the mode of other calculating similarities that is likely to occur from now on be such as applicable to the application, should also be included in the application and protect model
Within enclosing, and it is incorporated herein by reference herein.
When matching module 230 determines that current call is abnormal incoming call according to the similarity of key feature, a threshold can be set
Value, goes out a fiducial value according to the similarity calculation of key feature, fiducial value and threshold value is compared, if being less than threshold value,
It indicates that identification feature value and the truth difference of key feature are larger, therefore, it is determined that current call is abnormal incoming call, i.e., has very much
It may be fraudulent call;If being less than threshold value, then it represents that the identification feature value and truth difference of key feature are little, therefore
Judge that current call is normal incoming call, i.e., current call is that the possibility of fraudulent call is smaller.
For fiducial value, specific calculation can directly calculate the arithmetic average of every key feature similarity,
Can also be after weight is set for every key feature, to calculate weighted average etc..Its specific calculation is in this application
It does not limit, suitable mode can be used according to concrete application scene.
In another implementation of the application, the needs of matching module 230 ensure in the communication process of current call, according to pass
The comparing result of the identification feature value and contrast characteristic's value of key feature determines whether current call is abnormal incoming call.Thereby it is ensured that
The real-time of recognition result avoids those who answer's fraudulent call when just knowing this call after having answered phone, and causing need not
The loss wanted.
After determining recognition result according to similarity, since the case where matching module 230 is judged according to similarity, is based on
The threshold value set based on experience value, therefore recognition result is it is difficult to ensure that complete correct.Thus, it is possible to determining that current call is
After abnormal incoming call, according to the similarity and/or contrast characteristic's value of the identification feature value of key feature and contrast characteristic's value come
Source obtains the confidence level of recognition result.Wherein, confidence level indicates the current correct probability of recognition result, as a result, final result
Can be that recognition result is sent a telegram here to be abnormal, confidence level 90%.
For the confidence level of recognition result, relative factor includes at least similarity and the contrast characteristic of key feature
The source of value can be calculated in the embodiment of the present application based on a single one factor, can also combine two kinds of factors one
It rises and calculates.For example, by taking the similarity of key feature as an example, similarity is higher, and confidence level also can be higher.In conjunction with aforementioned comparison's value
Calculation, the similarity of every key feature is higher, and the reduced value being calculated accordingly can improve therewith, it is possible thereby to root
Confidence level is calculated according to the difference of reduced value and threshold value, if reduced value is more than that threshold value is bigger, and confidence level is higher, otherwise confidence level is lower.
About the source of contrast characteristic's value, since the source of contrast characteristic's value in the retrieval result of search engine includes a variety of
Possibility, such as come from official website or other all kinds of websites, the reliability of contrast characteristic's value of various separate sources not phase
Together, under normal circumstances, the reliability in official website source can be higher than other websites.Therefore, in contrast characteristic's value of every key feature
The quantity for coming from official website is more, then the confidence level of final recognition result is also higher.
In addition, in one embodiment of the application, equipment further includes a report module.It is abnormal send a telegram here for recognition result
Number, report module can report it, such as can be to concrete mode:To specific number or address (example
Such as operator, the police) report information is sent in a manner of voice, short message, mail etc., or it is a by being marked to the number of the incoming call
Property label, be sent to specific database so that can be directly given when being connected to identical number again it is necessary prompt etc..
In another embodiment of the application, equipment further includes a display module.In the communication process of current call, show
Show that module can show at least one of following information so that those who answer can check these information, to understand current call in real time
The case where.These information include:The comparison of the identification feature value of key feature, the contrast characteristic's value, key feature of key feature
As a result, the confidence level etc. of the recognition result of current call and recognition result.
By taking mobile phone as an example, display module can be the display screen and associated circuit components of mobile phone.In communication process,
The screen of mobile phone directly display in above- mentioned information any one or it is several, thus those who answer can be in communication process at any time
It is checked, to current call the case where is centainly understood.For example, the form of text, real-time display can be passed through in mobile phone
The identification feature value and contrast characteristic's value of key feature, user can be by sliding checking for screen selectivity, in conjunction with reality
Border situation is judged oneself, and does not depend on the judging result of mobile phone.
Fig. 4 shows handling principle of the equipment provided by the embodiments of the present application in the abnormal incoming call of identification, below in conjunction with Fig. 3
Shown in scene, entire interactive process is described in detail.The abnormal incoming call of identification is realized based on mobile phone in scene shown in Fig. 3
Scheme, wherein mobile phone A is the mobile phone of caller, and mobile phone B be the mobile phone being called, and server C is to provide the network of retrieval service to set
It is standby.Whole process includes the following steps:
Step S301, caller are dialed the number of mobile phone B using mobile phone A, are called.
Module is guarded in step S302, the monitoring in mobile phone B, can carry out the monitoring of calling event on backstage, when listening to
When incoming call (i.e. the incoming call from mobile phone A), start relevant identification module.
The identification module of step S303, mobile phone B carry out incoming call identification first, that is, the number of current call are obtained, as it
In a key feature identification feature value.
Step S304, when the user (i.e. those who answer) of mobile phone B picks up phone, identification module is right in real time in communication process
Voice content is analyzed, and the identification feature value of other key features is extracted.For example, the identity of caller, address, affiliated portion
Door etc., can specifically be converted into textual form, judge in order to compare.
Step S305, the key feature that mobile phone B is inquired as needed send retrieval request to server C, are drawn by search
Hold up the contrast characteristic's value for searching key feature.For example, by the identity according to caller in key feature, obtained by its official website
With the relevant information of key feature, as its contrast characteristic's value, the authenticity for examining identification feature value.
After step S306, server C retrieve contrast characteristic's value of key feature, it is sent to mobile phone B.
The identification feature value of key feature and contrast characteristic's value are compared by step S307, mobile phone B, obtain comparison knot
Fruit.
Step S308, if determining the recognition result of this incoming call in comparing result.For example, in this incoming call, multinomial key
The identification feature value of feature is different with contrast characteristic's value, then can determine that the possibility that this incoming call belongs to fraudulent call is larger,
It is determined as abnormal incoming call.Meanwhile the intervention means such as report can be taken the incoming number.
Step S309, mobile phone B show recognition result on the screen.Such as can show that this incoming call is abnormal incoming call, this
Outside, for recognition result, it is the authenticity probability (i.e. confidence level) sent a telegram here extremely that can calculate it, is shown simultaneously with recognition result
Show, to provide the user with more rational reference information.For other average informations in processing procedure, such as key feature
Identification feature value, contrast characteristic's value of key feature, key feature comparing result etc., can also together be shown in screen
Show.
In conclusion scheme provided by the present application is without establishing presetting database, it is special by the key for obtaining current call
The identification feature value of sign, with by search engine retrieving to contrast characteristic's value of the key feature be compared, final root
According to the comparing result of the identification feature value and contrast characteristic's value of the key feature, it may be determined that whether the current call is different
Often incoming call.Since contrast characteristic's value for being compared is directly to retrieve acquisition by search engine, it is not rely on default
Database, therefore in Non-precondition database or not perfect presetting database, higher accuracy can also be reached.
In addition, the part of the application can be applied to computer program product, such as computer program instructions, when its quilt
When computer executes, by the operation of the computer, it can call or provide according to the present processes and/or technical solution.
And the program instruction of the present processes is called, it is possibly stored in fixed or moveable recording medium, and/or pass through
Broadcast or the data flow in other signal loaded mediums and be transmitted, and/or be stored according to described program instruction operation
In the working storage of computer equipment.Here, including that an acquisition as shown in Figure 5 passes according to one embodiment of the application
The equipment of defeated file, the equipment include memory 510 for storing computer program instructions and for executing program instructions
Processor 520, wherein when the computer program instructions are executed by the processor, trigger equipment operation and be based on aforementioned basis
The method and/or technology scheme of multiple embodiments of the application.
It should be noted that the application can be carried out in the assembly of software and/or software and hardware, for example, can adopt
With application-specific integrated circuit (ASIC), general purpose computer or any other realized similar to hardware device.In one embodiment
In, the software program of the application can be executed by processor to realize steps described above or function.Similarly, the application
Software program (including relevant data structure) can be stored in computer readable recording medium storing program for performing, for example, RAM memory,
Magnetic or optical driver or floppy disc and similar devices.In addition, hardware can be used to realize in some steps or function of the application, example
Such as, coordinate to execute the circuit of each step or function as with processor.
It is obvious to a person skilled in the art that the application is not limited to the details of above-mentioned exemplary embodiment, Er Qie
In the case of without departing substantially from spirit herein or essential characteristic, the application can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and scope of the present application is by appended power
Profit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent requirements of the claims
Variation is included in the application.Any reference signs in the claims should not be construed as limiting the involved claims.This
Outside, it is clear that one word of " comprising " is not excluded for other units or step, and odd number is not excluded for plural number.That is stated in device claim is multiple
Unit or device can also be realized by a unit or device by software or hardware.The first, the second equal words are used for table
Show title, and does not represent any particular order.
Claims (15)
1. a kind of identifying the method sent a telegram here extremely, wherein the method includes:
Obtain the identification feature value of the key feature of current call;
Pass through contrast characteristic's value of key feature described in search engine retrieving;
According to the comparing result of the identification feature value and contrast characteristic's value of the key feature, determine the current call whether be
Abnormal incoming call.
2. according to the method described in claim 1, wherein, by contrast characteristic's value of key feature described in search engine retrieving,
Including:
In the communication process of the current call, pass through contrast characteristic's value of key feature described in search engine retrieving.
3. according to the method described in claim 1, wherein, according to the identification feature value of the key feature and contrast characteristic's value
Comparing result determines whether the current call is abnormal incoming call, including:
In the communication process of the current call, according to the comparison of the identification feature value and contrast characteristic's value of the key feature
As a result, determining whether the current call is abnormal incoming call.
4. according to the method described in claim 1, wherein, according to the identification feature value of the key feature and contrast characteristic's value
Comparing result determines whether the current call is abnormal incoming call, including:
Determine the similarity of the identification feature value and contrast characteristic's value of the key feature;
Determine whether the current call is abnormal incoming call according to the similarity of the key feature.
5. according to the method described in claim 1, wherein, after determining that the current call is abnormal incoming call, further including:
According to the identification feature value of the key feature and the similarity of contrast characteristic's value and/or the source of contrast characteristic's value
Obtain the confidence level of recognition result, wherein the recognition result includes that the current call is abnormal sends a telegram here.
6. the method according to any one of claims 1 to 5, wherein, this method further includes:
In the communication process of the current call, at least one of display is following information:
The identification feature value of the key feature;
Contrast characteristic's value of the key feature;
The comparing result of the key feature;
The recognition result of the current call;
The confidence level of the recognition result.
7. the method according to any one of claims 1 to 5, wherein, determining that the current call is that exception sends a telegram here it
Afterwards, further include:
The abnormal incoming call is reported.
8. a kind of identifying the equipment sent a telegram here extremely, wherein the equipment includes:
Identification module, the identification feature value of the key feature for obtaining current call;
Module is retrieved, for contrast characteristic's value by key feature described in search engine retrieving;
Matching module, for the comparing result according to the identification feature value and contrast characteristic's value of the key feature, described in determination
Whether current call is abnormal incoming call.
9. equipment according to claim 8, wherein the retrieval module, for the communication process in the current call
In, pass through contrast characteristic's value of key feature described in search engine retrieving.
10. equipment according to claim 8, wherein the matching module, for the communication process in the current call
In, according to the comparing result of the identification feature value and contrast characteristic's value of the key feature, determine that the current call is determined as
Abnormal incoming call.
11. equipment according to claim 8, wherein the matching module, the identification for determining the key feature are special
The similarity of value indicative and contrast characteristic's value;And determine whether the current call is different according to the similarity of the key feature
Often incoming call.
12. equipment according to claim 8, wherein the matching module is additionally operable to determining that the current call is different
Often after incoming call, according to the similarity and/or the contrast characteristic of the identification feature value of the key feature and contrast characteristic's value
The source of value obtains the confidence level of recognition result, wherein the recognition result includes that the current call is abnormal sends a telegram here.
13. the equipment according to any one of claim 8 to 12, wherein the equipment further includes:
Display module, in the communication process of the current call, showing at least one of following information:
The identification feature value of the key feature;
Contrast characteristic's value of the key feature;
The comparing result of the key feature;
The recognition result of the current call;
The confidence level of the recognition result.
14. the equipment according to any one of in claim 8 to 12, wherein the equipment further includes:
Processing module is reported, for after determining that the current call is abnormal incoming call, reporting the abnormal incoming call.
15. a kind of identifying the equipment sent a telegram here extremely, wherein the equipment includes:
Processor;And
It is arranged to the memory of storage computer executable instructions, the executable instruction makes the processing when executed
Device:The identification feature value for obtaining the key feature of current call, passes through the contrast characteristic of key feature described in search engine retrieving
Value, and identification feature value and contrast characteristic's value according to the key feature comparing result, determine that the current call is
It is no to send a telegram here to be abnormal.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710010573.6A CN108280089A (en) | 2017-01-06 | 2017-01-06 | Identify the method and apparatus sent a telegram here extremely |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710010573.6A CN108280089A (en) | 2017-01-06 | 2017-01-06 | Identify the method and apparatus sent a telegram here extremely |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108280089A true CN108280089A (en) | 2018-07-13 |
Family
ID=62800928
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710010573.6A Pending CN108280089A (en) | 2017-01-06 | 2017-01-06 | Identify the method and apparatus sent a telegram here extremely |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108280089A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109451182A (en) * | 2018-10-19 | 2019-03-08 | 北京邮电大学 | A kind of detection method and device of fraudulent call |
CN111311411A (en) * | 2020-02-14 | 2020-06-19 | 北京三快在线科技有限公司 | Illegal behavior identification method and device |
WO2021051504A1 (en) * | 2019-09-18 | 2021-03-25 | 平安科技(深圳)有限公司 | Method for identifying abnormal call party, device, computer apparatus, and storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20090079003A (en) * | 2008-01-16 | 2009-07-21 | 에스케이 텔레콤주식회사 | Method and apparatus for collection of abnormal call and communication service system by using the collection apparatus |
CN101719967A (en) * | 2009-12-04 | 2010-06-02 | 明泰科技股份有限公司 | Method, system and device for providing caller information |
CN102801844A (en) * | 2012-08-22 | 2012-11-28 | 广东欧珀移动通信有限公司 | Incoming call information display method and mobile terminal |
CN103179122A (en) * | 2013-03-22 | 2013-06-26 | 马博 | Telcom phone phishing-resistant method and system based on discrimination and identification content analysis |
CN103458412A (en) * | 2012-06-04 | 2013-12-18 | 百度在线网络技术(北京)有限公司 | System and method for preventing phone fraud, mobile terminal and cloud terminal analysis server |
CN104427079A (en) * | 2013-09-09 | 2015-03-18 | 中兴通讯股份有限公司 | Early warning method and device of user voice calls |
-
2017
- 2017-01-06 CN CN201710010573.6A patent/CN108280089A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20090079003A (en) * | 2008-01-16 | 2009-07-21 | 에스케이 텔레콤주식회사 | Method and apparatus for collection of abnormal call and communication service system by using the collection apparatus |
CN101719967A (en) * | 2009-12-04 | 2010-06-02 | 明泰科技股份有限公司 | Method, system and device for providing caller information |
CN103458412A (en) * | 2012-06-04 | 2013-12-18 | 百度在线网络技术(北京)有限公司 | System and method for preventing phone fraud, mobile terminal and cloud terminal analysis server |
CN102801844A (en) * | 2012-08-22 | 2012-11-28 | 广东欧珀移动通信有限公司 | Incoming call information display method and mobile terminal |
CN103179122A (en) * | 2013-03-22 | 2013-06-26 | 马博 | Telcom phone phishing-resistant method and system based on discrimination and identification content analysis |
CN104427079A (en) * | 2013-09-09 | 2015-03-18 | 中兴通讯股份有限公司 | Early warning method and device of user voice calls |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109451182A (en) * | 2018-10-19 | 2019-03-08 | 北京邮电大学 | A kind of detection method and device of fraudulent call |
CN109451182B (en) * | 2018-10-19 | 2021-08-13 | 北京邮电大学 | Detection method and device for fraud telephone |
WO2021051504A1 (en) * | 2019-09-18 | 2021-03-25 | 平安科技(深圳)有限公司 | Method for identifying abnormal call party, device, computer apparatus, and storage medium |
CN111311411A (en) * | 2020-02-14 | 2020-06-19 | 北京三快在线科技有限公司 | Illegal behavior identification method and device |
CN111311411B (en) * | 2020-02-14 | 2022-03-08 | 北京三快在线科技有限公司 | Illegal behavior identification method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109815314B (en) | Intent recognition method, recognition device and computer readable storage medium | |
CN106384273B (en) | Malicious bill-swiping detection system and method | |
WO2019184217A1 (en) | Hotspot event classification method and apparatus, and storage medium | |
CN109147797B (en) | Customer service method, device, computer equipment and storage medium based on voiceprint recognition | |
US11727053B2 (en) | Entity recognition from an image | |
WO2020207167A1 (en) | Text classification method, apparatus and device, and computer-readable storage medium | |
CN110275965B (en) | False news detection method, electronic device and computer readable storage medium | |
US9229974B1 (en) | Classifying queries | |
WO2019085064A1 (en) | Medical claim denial determination method, device, terminal apparatus, and storage medium | |
WO2021068635A1 (en) | Information processing method and apparatus, and electronic device | |
US9251395B1 (en) | Providing resources to users in a social network system | |
CN110472027A (en) | Intension recognizing method, equipment and computer readable storage medium | |
WO2019174073A1 (en) | Method and device for modifying client information in conversation, computer device and storage medium | |
CN108280089A (en) | Identify the method and apparatus sent a telegram here extremely | |
US11367311B2 (en) | Face recognition method and apparatus, server, and storage medium | |
CN111950267B (en) | Text triplet extraction method and device, electronic equipment and storage medium | |
CN116246287B (en) | Target object recognition method, training device and storage medium | |
EP4092564A1 (en) | Method and apparatus for constructing event library, electronic device and computer readable medium | |
CN110737677A (en) | data searching system and method | |
CN114817518A (en) | License handling method, system and medium based on big data archive identification | |
CN113962302A (en) | Sensitive data intelligent identification method based on label distribution learning | |
CN111475657A (en) | Display device, display system and entity alignment method | |
CN116401410B (en) | Method, device, storage medium and equipment for accessing map data to multi-scene graph database | |
CN115690932B (en) | Outdoor operation digital attendance checking method and system based on NFC technology | |
KR102524828B1 (en) | Detective agency brokerage system based on big data |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180713 |