CN104936182A - Method of managing and controlling fraud telephones intelligently and system of managing and controlling fraud telephones intelligently - Google Patents
Method of managing and controlling fraud telephones intelligently and system of managing and controlling fraud telephones intelligently Download PDFInfo
- Publication number
- CN104936182A CN104936182A CN201510190578.2A CN201510190578A CN104936182A CN 104936182 A CN104936182 A CN 104936182A CN 201510190578 A CN201510190578 A CN 201510190578A CN 104936182 A CN104936182 A CN 104936182A
- Authority
- CN
- China
- Prior art keywords
- calling
- swindle
- doubtful
- directory
- period
- 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.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
- H04W12/12—Detection or prevention of fraud
-
- 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)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Computer Networks & Wireless Communication (AREA)
- Telephonic Communication Services (AREA)
Abstract
The present invention relates to a method of managing and controlling fraud telephones intelligently and a system of managing and controlling fraud telephones intelligently. The method comprises a step 1 of triggering a calling initiated by a calling user to a calling control node SCP, using the calling control node SCP to query whether a calling number is in a suspected number table or an intercepted number table, and then adding a corresponding prefix identifier before a called number, and finally forwarding the calling to an independent and intelligent peripheral device IIP, wherein the suspected number table is used to save the suspected fraud numbers, the intercepted number table is used to save the fraud numbers that have been confirmed, and the prefix identifier is used to identify whether the calling number is in the suspected number table or the intercepted number table; a step 2 of using the intelligent peripheral device IIP to read the prefix identifier before the called number, when the prefix identifier identifies that the calling number is in the intercepted number table, simulating a called button to connect a manual service agent of a calling end, or simulating a called party to answer the telephone. The method and system of the present invention belong to the network communication technology field, and can realize the multi-path effective management and control on the fraud telephones.
Description
Technical field
The present invention relates to a kind of method and system of intelligent management and control fraudulent call, belong to network communication technology field.
Background technology
In recent years, the case implementing telephone fraud by communication network is situation occurred frequently, swindle gimmick is maked rapid progress, though all departments take certain measure carry out keypoint control and strike to telephone fraud, but effect is also not obvious, the effective management and control of multipath how realizing fraudulent call from communication network aspect progressively becomes important topic and the task of the research of Ge great operator.
Current operator is comparatively single for the management-control method of fraudulent call, and interception number is incomplete, and interception means underaction.Such as, patent application CN 201110415908.5 (application title: malice grouies busy phone counter apparatus and method, the applying date: 2011-12-14, applicant: China Telecommunication Stock Co., Ltd) disclose a kind of malice and group busy phone counter method and apparatus, this device comprises signal collecting module, analysis and processing module, counter module.To malice, this device is by grouping busy that the infecctive calls such as telephone call spacing are carried out emulation and answered, and analog subscriber button enters the manual board seat in malicious call source, take manual board seat resource and disturb, thus reach compacting and group busy class malicious call, shorten and group busy lasting duration at every turn, reduce the object of frequency.The infecctive call that this technology is only applicable to dialing spacing breaks through, and be only limited in its technical scheme and adopt the button of artificial setting to enter the manual board seat in malicious call source, its management and control means are too single, thus can not realize multipath effective management and control to fraudulent call complicated and diversified in existing network.
Therefore, how realizing multipath effective management and control to fraudulent call, is a technical problem being worth further investigation.
Summary of the invention
In view of this, the object of this invention is to provide a kind of method and system of intelligent management and control fraudulent call, multipath effective management and control can be realized to fraudulent call.
In order to achieve the above object, the invention provides a kind of method of intelligent management and control fraudulent call, include:
The calling that step one, calling subscriber initiate is triggered to call control node SCP, SCP inquires about calling number whether in doubtful directory or interception directory, then before called number, corresponding prefix identifier is increased, finally calling is forwarded to independent intelligent peripheral IIP, wherein, doubtful directory is for preserving doubtful swindle number, and interception directory is for preserving confirmed swindle number, and described prefix identifier is for identifying calling number whether in doubtful directory or interception directory;
Step 2, IIP read the prefix identifier before called number, when described prefix identifier mark calling number is when tackling in directory, then simulate called button to connect the operator attendance of calling terminal or to simulate called receiving calls.
In order to achieve the above object, present invention also offers a kind of system of intelligent management and control fraudulent call, include:
Call control node SCP, for receiving the call request that calling subscriber initiates, and whether calling number in query call request is in doubtful directory or interception directory, then before called number, corresponding prefix identifier is increased, finally calling is forwarded to IIP, wherein, doubtful directory is for preserving doubtful swindle number, interception directory is for preserving confirmed swindle number, and described prefix identifier is for identifying calling number whether in doubtful directory or interception directory;
Independent intelligent peripheral IIP, the call request of coming is forwarded for receiving SCP, and the prefix identifier read in call request before called number, when described prefix identifier mark calling number is when tackling in directory, then simulate called button to connect the operator attendance of calling terminal or to simulate called receiving calls.
Compared with prior art, the invention has the beneficial effects as follows: the present invention proposes a kind of method and system possessing the intelligent management and control fraudulent call of anti-reconnaissance capability, calling is toggled to SCP by core net, by SCP, IIP process is delivered in the calling that doubtful or confirmation swindle number are initiated again, by cooperation and the flexible dispatching of SCP and IIP entity, and adopt business function expand and introduce new technological means, thus achieve the flexible management and control of fraudulent call; The present invention first collection network side real time phone call record, and adopt BP neural network algorithm to analyze user's communication behavior, thus highly doubtful swindle number is accurately identified from the whole network calling; For doubtful swindle number, the present invention records in call conversation process, and N recording file second before intercepting call, then mated by real-time voice text-converted, in communication process, then directly issue instruction of taking out stitches as the match is successful stop this call, thus play the real-time blocking function to fraudulent call; Swindle number is adopted to analogue-key, analog answer, proceeds to the multiple management and control means such as public security operator attendance, wherein analogue-key extracts press key contents from self recording text of fraudulent call, thus accurately can start the operator attendance that switching enters fraudulent call, analog answer adopts the adaptive model based on existing network telephone traffic, thus can increase crime cost again while fascination swindle one's share of expenses for a joint undertaking; The present invention can also according to doubtful swindle number calling behavioural characteristic subsequently, and that precisely identifies that swindle molecule do not re-use or do not re-use in a period of time abandons number, thus effectively reduces system loading.
Accompanying drawing explanation
Fig. 1 is the flow chart of the method for a kind of intelligent management and control fraudulent call of the present invention.
Fig. 2 is the concrete operations flow chart of Fig. 1 step one.
Fig. 3 is the concrete operations flow chart of step 22.
Fig. 4 is that the present invention passes through to analyze user's communication behavioural information, obtains doubtful swindle number, and write the concrete operations flow chart in doubtful directory from the calling ticket in certain hour cycle T.
Fig. 5 is for the doubtful swindle number obtained new from the calling ticket in certain hour cycle T, the present invention, according to its calling behavioural characteristic subsequently, precisely identifies the concrete operations flow chart abandoning number that swindle molecule does not re-use or do not re-use in a period of time.
Fig. 6 is the composition structural representation of IIP.
Fig. 7 is the composition structural representation that non-swindle number rejects parts.
Embodiment
For making the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, the present invention is described in further detail.
As shown in Figure 1, the method for a kind of intelligent management and control fraudulent call of the present invention, includes:
The calling that step one, calling subscriber initiate is triggered to call control node SCP, SCP inquires about calling number whether in doubtful directory or interception directory, then before called number, corresponding prefix identifier is increased, finally calling is forwarded to independent intelligent peripheral IIP, wherein, doubtful directory is for preserving doubtful swindle number, and interception directory is for preserving confirmed swindle number, and described prefix identifier is for identifying calling number whether in doubtful directory or interception directory;
Step 2, IIP read the prefix identifier before called number, when described prefix identifier mark calling number is when tackling in directory, then simulate called button to connect the operator attendance of calling terminal or to simulate called receiving calls.
As shown in Figure 2, Fig. 1 step one may further include:
Step 11, SCP receive the call request that calling subscriber initiates, do you judge that calling number is in doubtful directory? if, before called number, then increase the prefix identifier (such as A) of doubtful identification, then calling continued to be forwarded to IIP, this flow process terminates; If not, then next step is continued;
Do step 12, SCP judge that calling number is in interception directory? if so, then next step is continued; If not, then continue to perform original operation flow, this flow process terminates;
Do step 13, SCP judge that the intercept information of tackling calling number described in directory has been mated? if, before called number, then increase the prefix identifier (such as B) of analogue-key, then calling continued to be forwarded to IIP, this flow process terminates; If not, then next step is continued;
Do step 14, SCP judge that the intercept information of tackling calling number described in directory is not mated? if, before called number, then increase the prefix identifier (such as C) of analog answer, then calling continued to be forwarded to IIP, this flow process terminates; If not, then increase before called number and forward artificial prefix identifier (such as D), then calling continued to be forwarded to IIP, this flow process terminates.
Fig. 1 step 2 can further include:
Step 21, IIP receives SCP and forwards the call request of coming, and do you judge it is the prefix identifier of doubtful identification before called number in call request? if, then bridge joint master, voice channel between called, then unidirectional recording is carried out to caller voice, the recording file reaching certain long recording time (such as 10 seconds) is changed into text, then each swindle text in described text and swindle Sample Storehouse is mated one by one, if the matching degree of a swindle text of described text and swindle Sample Storehouse exceedes certain threshold value, then illustrate that the calling number that described text is corresponding is swindle number, described calling number is moved to interception directory from doubtful directory, and identify its intercept information for mate, detect in described text simultaneously and whether have continuous print numeral, and continuous print numeral number is more than or equal to clawback bond number minimum value, if had, then digital for this string continuous print callback number as described calling number is saved in tables of data, when having multiple callback number in text, all callback numbers are all saved in tables of data, finally remove this main, voice channel between called, this flow process terminates, when the matching degree of all swindle texts in described text and swindle Sample Storehouse is all lower than threshold value, then continue next step, if not, then step 23 is turned to,
The present invention records in call conversation process, and intercept call first 10 seconds recording files, mated by real-time voice text-converted, if the match is successful, illustrate it is fraudulent call, in communication process, directly issue instruction of taking out stitches to stop this call, thus play real-time blocking function; Described clawback bond number minimum value can set according to actual conditions, such as, when clawback bond number minimum value is set to 3, if there is continuous print numeral number in text to be more than or equal to 3, then can judge that this string continuous print numeral is as a callback number;
Whether the calling number in step 22, the request of IIP detection of call is spacing, if, then calling number is moved to interception directory from doubtful directory, and identify its intercept information for not mate, described text is saved in swindle Sample Storehouse simultaneously, then remove the voice channel between these calling and called, this flow process terminates; If not, then carry out manual examination and verification to be confirmed whether being fraudulent call to recording text, if fraudulent call is confirmed as in manual examination and verification, then calling number moved to interception directory from doubtful directory, and identify its intercept information for manually to determine, this flow process terminates;
Do step 23, IIP judge it is the prefix identifier of analogue-key before called number in call request? if so, then next step is continued; If not, then step 25 is continued;
Do step 24, IIP search in tables of data the callback number of preserving calling number in call request? if had, from tables of data, then extract a callback number of calling number, touch-tone signal corresponding for described callback number is encapsulated in APM message, then by described APM Message Transmission to calling terminal, preset true man's recording is selected to play to calling terminal again, thus after making calling terminal identification, starting switching manual service, this flow process terminates; If not, then step 26 is turned to;
Swindling gang often has oneself calling system, normally adopt IVR technology, the human relay working of speech channel is realized by mobile phone key, but its media resource is limited, namely be limited to the disposal ability of method, system, in the present invention, calling terminal then can start switching manual service after identifying APM message; If source initiates calling in enormous quantities, after IIP analogue-key, the operator attendance of calling terminal will be connected, thus cause the other side's systemic breakdown.Some low frequencies are called out, also can take the artificial listening time of the other side.The present invention also proposes a kind of analogue-key method of fascination, consider that the DTMF that swindle IVR platform turns manual service does not fix, to in this swindle number pre-recording evidence obtaining process, by identifying that recording content of text obtains concrete callback number and writes tables of data, like this, the callback number that IIP inquires about this calling number corresponding in database table carrys out loopback calling system.When the corresponding multiple callback number of calling number in database table, system can adopt polling mode process, and the method both saved signalling bandwidth, again can treatment of simulated DTMF function exactly, and treatment effeciency and performance have had lifting by a relatively large margin.Voice document is recorded in advance, and is undertaken imitating different response content by different true man's sounds, and object is to obscure operator attendance, fascination offender, increase the time cost of swindle one's share of expenses for a joint undertaking, to reduce the probability of successful service comple of domestic consumer, thus reduce swindle success rate;
Do step 25, IIP judge it is the prefix identifier of analog answer before called number in call request? if so, then next step is continued; If not, then step 27 is continued;
Step 26, does IIP judge that the actual response rate of analog answer is less than the response rate of system intialization in current time period? if, then according in upper a period of time, the ring duration that the doubtful swindle of several in doubtful directory number sampling statistics goes out is come to calling terminal coloring ring back tone playing or common ring-back tone, then answer charging message ANM is sent to originating switch, and play to calling terminal to select quiet or true man's recording according to the response duration gone out doubtful swindle number sampling statistics in upper a period of time, thus simulate called connection and the scene of response, this flow process terminates, if not, then according to the call completing rate gone out doubtful swindle number sampling statistics in upper a period of time, unreachable rate with close probability, simulate the scene of called connection nonreply, unreachable or shutdown to calling terminal, this flow process terminates,
Such as, by the sampling statistics to swindle number doubtful in existing network, following numerical value can be calculated: call completing rate=92%, unreachable rate=3%, closes probability=5%, response rate=30%, response duration interval is: [0,3] are 2.7%, [3,15] be 43.9%, [15,60] are 52.9%, [60,180] are more than 0.4%, 180 seconds is 0.1%, ring duration mean value is got 15 seconds and presents with normal distribution, and CRBT ratio can with reference to each province's CRBT traffic model.Wherein, connecting nonreply scene can be: by preset ratio playing RBT and CRBT, and CRBT song is play by wheel broadcast mode; Unreachable scene can be: play unreachable prompt tone according to preset ratio; Shutdown scene can be: play shutdown prompt tone according to preset ratio;
Step 27, does IIP judge it is forward artificial prefix identifier before called number in call request? if, whether the real-time caps of call volume then calculating public security operator attendance is less than secure threshold, if the real-time caps of described call volume is less than secure threshold, then described call request is continued to be forwarded to public security operator attendance, be responsible for answering reply fraudulent call by professional, obtain this call path information (such as call path by real-time interface simultaneously, former calling number, original called party number etc.) pass to public security operator attendance, investigate and collect evidence to assist professional and the real dialing number information of instant understanding, if the real-time caps of described call volume is more than or equal to secure threshold, then turn to step 26, if not, then this flow process terminates.
As shown in Figure 3, step 22 can further include:
Step 221, IIP make a call to calling number, and receive the return messages of calling number place end office (EO);
Do you step 222, judge that described message is clear message REL and its cause value is spacing (namely 81 or 9C)? if so, then represent that described calling number is spacing, this flow process terminates; If not, then next step is continued;
Do you step 223, judge that described message is ACM message? if so, then next step is continued; If not, then represent that described calling number is not spacing, this flow process terminates;
Do you step 224, judge that the cause value in described ACM message is spacing? if so, then represent that described calling number is spacing, this flow process terminates; If not, then represent that described calling number is not spacing, and when the cause value in described ACM message is the free time, after preventing connecting caller, produce ticket, issue clear message REL to callee side, this flow process terminates.
In step 26, in current time period, the computing formula of the actual response rate of analog answer is as follows:
wherein, MYDL (T) is the actual response rate of analog answer in current time period T, MYDN (T) is that in current time period T, IIP simulates called connection and the calls of response, and SUNC (T) is the calls in current time period T received by IIP.The response rate of system intialization can also according to the actual telephone traffic of upper a period of time, dynamic self-adapting is carried out in each time cycle, in each time cycle, the computing formula of the response rate of system intialization can be as follows: XYDL (T)=YDL (T-1)+n%, wherein, XYDL (T) is the response rate of system intialization in current time period T, YDL (T-1) is the actual response rate that upper a period of time T-1 goes out the doubtful swindle of several in doubtful directory number sampling statistics, n% is that analog answer rate improves percentage, such as n%=50%, suitable raising response rate, operator's income can be increased, crime one's share of expenses for a joint undertaking crime cost can be increased again.
It is worth mentioning that, the present invention can also by analyzing user's communication behavioural information, and from the calling ticket in certain hour cycle T, obtain doubtful swindle number, and write in doubtful directory, as shown in Figure 4, the present invention can also include:
Steps A 1, from the calling ticket in certain hour cycle T, extract several calling numbers, build number sample group to be confirmed and confirm swindle sample group, and judge whether the calling number of each extraction is tackling in directory, when calling number is when tackling in directory, then described calling number being divided into and confirming swindle sample group, when calling number is not when tackling in directory, then described calling number being divided into number sample group to be confirmed;
Steps A 2, from the calling ticket in period of time T, extract number sample group to be confirmed and confirmed to swindle the N number of calling parameter value in sample group corresponding to all calling numbers, N is a natural number, its value can be determined according to actual conditions, and described calling parameter value can include but not limited to: numbers of calling and called parties, calling time started, end of calling time, call duration, ring duration or calling and called ownership area code etc.;
Steps A 3, a structure BP neural net, with N number of calling parameter value of all calling numbers in number sample group to be confirmed for input, and according to confirming that the N number of calling parameter value swindling all calling numbers in sample group carrys out calculation expectation output valve, train described BP neural net;
Steps A 4, from the calling ticket in period of time T, extract N number of calling parameter value of each calling number one by one, and be input to described BP neural net to calculate its output valve, when error between described output valve and desired output is within the scope of certain numerical value, then illustrate that described calling number is doubtful swindle number, calling number is write in doubtful directory.
For a better understanding of the present invention, be necessary simply to introduce the general principle of BP neural net.BP neural net contains one deck input layer, one deck output layer and at least one deck hidden layer, its general principle is the error by calculating between output layer and desired output, then from output layer, oppositely adjust each neuronic weights and bias, finally make the error between the output at networking and desired output meet the requirement pre-set.Specifically, the training of BP neural net is divided into following two stages:
1, in the propagated forward stage, input vector is introduced by input layer, conducts to output layer with feed-forward mode via hidden layer, and calculates network output valve, and now, the weights of network are all fixing;
2, back-propagation phase, the weights of network are then revised according to error correction rule, to make the output valve of network trend towards desired output, be namely specifically deduct network output valve with desired output to obtain error signal, then this error signal backpropagation returned in network.
The steps A 3 of Fig. 4 can further include:
Steps A 31, a structure BP neural net, to the weight matrix W of its input layer to hidden layer
1with deviation matrix B
1, and hidden layer to the weight matrix W of output layer
2with deviation matrix B
2value carry out random initializtion;
Steps A 32, from number sample group to be confirmed extract a calling number;
Steps A 33, using the N number of calling parameter value corresponding to the calling number extracted as the input value of BP neural net, according to W
1, B
1, W
2and B
2calculate the neuronic output valve of output layer, and come W according to the error between the neuronic output valve of output layer and desired output
1, B
1, W
2, B
2value oppositely adjust, until the error between the neuronic output valve of output layer and desired output is less than an error threshold pre-set;
In steps A 33, the computing formula of desired output σ is as follows:
m has confirmed to swindle the calling number number in sample group,
the neuronic output valve sum of output layer having confirmed that N number of calling parameter value of all calling numbers in swindle sample group obtains as the input value of BP neural net, θ
iconfirmed to swindle the calling number Z in sample group
in number of calling parameter value as the input value of BP neural net and the neuronic output valve of output layer obtained;
Steps A 34, judge whether to have extracted all calling numbers from number sample group to be confirmed? if so, then this flow process terminates; If not, then continue to extract next calling number from number sample group to be confirmed, then turn to steps A 33.
Simultaneously, for the doubtful swindle number of above-mentioned new acquisition from the calling ticket in certain hour cycle T, the present invention can also according to its calling behavioural characteristic subsequently, precisely identify swindle molecule do not re-use or do not re-use in a period of time abandon number, thus it is deleted from doubtful directory, effectively reduce system loading.As shown in Figure 5, the present invention can also include:
Step B1, all doubtful swindle number in the doubtful directory of new write in a period of time T is formed an outlier analysis bank, and the number of calls extracted in outlier analysis bank in K the continuous print period of time T of all doubtful swindle numbers in period of time T and after period of time T, K is a natural number, its value can set according to actual conditions, such as K=5;
Step B2, from outlier analysis bank extract a doubtful swindle number;
Step B3, the number of calls sum judged in K the continuous print period of time T of described doubtful swindle number after period of time T
0? if so, then described doubtful swindle number is deleted from doubtful directory, and turn to step B6; If not, then next step is continued; Wherein, CallingTA (t) is the number of calls in t the period of time T of doubtful swindle number after period of time T;
Step B4, calculate the average suppression ratio Calling_minus_ratio of total number of calls of described doubtful swindle number:
Wherein, CallingT is the number of calls of doubtful swindle number in period of time T, and do you judge that Calling_minus_ratio is less than the minimum threshold of the average suppression ratio of total number of calls? if so, then described doubtful swindle number is deleted from doubtful directory, and turn to step B6; If not, then next step is continued;
Step B5, calculate the least square slope ratio Slope_ratio of described doubtful swindle number:
Wherein, Calling_ratio (t) is the number of calls suppression ratio in t the period of time T of doubtful swindle number after period of time T, and further, the computing formula of Calling_ratio (t) is as follows:
And judge whether that Slope_ratio is at (A
1, A
2) numerical intervals within the scope of and Calling_minus_ratio is less than the max-thresholds of the average suppression ratio of total number of calls? if so, then described doubtful swindle number is deleted from doubtful directory, then continue next step; If not, then next step is continued; Wherein, A
1, A
2minimum, the large threshold value of least square slope ratio, the most large and small threshold value of the average suppression ratio of total number of calls and least square slope ratio can according to actual conditions value;
Step B6, judge whether all doubtful swindle number that extracted in outlier analysis bank? if so, then this flow process terminates; If not, then step B2 is turned to.
According to practical application effect, as K=5, max-thresholds=-0.5013 of minimum threshold=-0.7443 of total average suppression ratio of the number of calls, the always average suppression ratio of the number of calls, A
1=-0.2906, A
2when=0.1257, the present invention can very effective precisely identify swindle molecule do not re-use or do not re-use in a period of time abandon number.
The system of a kind of intelligent management and control fraudulent call of the present invention, includes:
Call control node SCP, for receiving the call request that calling subscriber initiates, and whether calling number in query call request is in doubtful directory or interception directory, then before called number, corresponding prefix identifier is increased, finally calling is forwarded to IIP, wherein, doubtful directory is for preserving doubtful swindle number, interception directory is for preserving confirmed swindle number, and described prefix identifier is for identifying calling number whether in doubtful directory or interception directory;
Independent intelligent peripheral IIP, the call request of coming is forwarded for receiving SCP, and the prefix identifier read in call request before called number, when described prefix identifier mark calling number is when tackling in directory, then simulate called button to connect the operator attendance of calling terminal or to simulate called receiving calls.
Described SCP includes further:
Prefix identifier device, for judging calling number in call request whether in doubtful directory or interception directory, when calling number is in doubtful directory, then increases the prefix identifier of doubtful identification before called number; When calling number to tackle in directory and its intercept information be mate time, then before called number, increase the prefix identifier of analogue-key; When calling number to tackle in directory and its intercept information be do not mate time, then before called number, increase the prefix identifier of analog answer; When calling number to tackle in directory and its intercept information be not mated and do not mated time, then increase before called number and forward artificial prefix identifier, finally calling is continued to be forwarded to IIP.
As shown in Figure 6, described IIP includes further:
Doubtful recognition device, for when before called number in call request being the prefix identifier of doubtful identification, bridge joint master, voice channel between called, and unidirectional recording is carried out to caller voice, the recording file reaching certain long recording time is changed into text, then each swindle text in described text and swindle Sample Storehouse is mated one by one, if the matching degree of a swindle text of described text and swindle Sample Storehouse exceedes certain threshold value, then illustrate that the calling number that described text is corresponding is swindle number, described calling number is moved to interception directory from doubtful directory, and identify its intercept information for mate, detect in described text simultaneously and whether have continuous print numeral, and continuous print numeral number is more than or equal to clawback bond number minimum value, if had, then digital for this string continuous print callback number as described calling number is saved in tables of data, when having multiple callback number in text, all callback numbers are all saved in tables of data, finally remove this main, voice channel between called, when the matching degree of all swindle texts in described text and swindle Sample Storehouse is all lower than threshold value, then notify vacant number detection device,
Vacant number detection device, whether be spacing for the calling number in detection of call request, if, then calling number is moved to interception directory from doubtful directory, and identify its intercept information for not mate, described text is saved in swindle Sample Storehouse simultaneously, then removes the voice channel between these calling and called; If not, then manual examination and verification are carried out to be confirmed whether being fraudulent call to recording text, if fraudulent call is confirmed as in manual examination and verification, then calling number is moved to interception directory from doubtful directory, and identify its intercept information for manually to determine;
Analogue-key device, for when before called number in call request being the prefix identifier of analogue-key, search the callback number of whether preserving calling number in call request in tables of data, if had, from tables of data, then extract a callback number of calling number, touch-tone signal corresponding for described callback number is encapsulated in APM message, then by described APM Message Transmission to calling terminal, select preset true man's recording to play to calling terminal again, thus after making calling terminal identification, start switching manual service;
Analog answer device, for when before called number in call request being the prefix identifier of analog answer, judge in current time period, whether the actual response rate of analog answer is less than the response rate of system intialization, if, then according in upper a period of time, the ring duration that the doubtful swindle of several in doubtful directory number sampling statistics goes out is come to calling terminal coloring ring back tone playing or common ring-back tone, then answer charging message ANM is sent to originating switch, and play to calling terminal to select quiet or true man's recording according to the response duration gone out doubtful swindle number sampling statistics in upper a period of time, thus simulate called connection and the scene of response, if not, then according to the call completing rate gone out doubtful swindle number sampling statistics in upper a period of time, unreachable rate with close probability, simulate the scene of called connection nonreply, unreachable or shutdown to calling terminal,
Forward Prosthesis, for when being when forwarding artificial prefix identifier before called number in call request, whether the real-time caps of call volume calculating public security operator attendance is less than secure threshold, if the real-time caps of described call volume is less than secure threshold, then described call request is continued to be forwarded to public security operator attendance, be responsible for answering reply fraudulent call by professional, obtain this call path information (such as call path by real-time interface simultaneously, former calling number, original called party number etc.) pass to public security operator attendance, investigate and collect evidence to assist professional and the real dialing number information of instant understanding, if the real-time caps of described call volume is more than or equal to secure threshold, then notify analog answer device.
In analog answer device, in current time period, the computing formula of the actual response rate of analog answer can be as follows:
wherein, MYDL (T) is the actual response rate of analog answer in current time period T, MYDN (T) is that in current time period T, IIP simulates called connection and the calls of response, and SUNC (T) is the calls in current time period T received by IIP.The response rate of system intialization can also according to the actual telephone traffic of upper a period of time, dynamic self-adapting is carried out in each time cycle, in each time cycle, the computing formula of the response rate of system intialization can be as follows: XYDL (T)=YDL (T-1)+n%, wherein, XYDL (T) is the response rate of system intialization in current time period T, YDL (T-1) is the actual response rate that upper a period of time T-1 goes out the doubtful swindle of several in doubtful directory number sampling statistics, n% is that analog answer rate improves percentage, such as n%=50%.
It is worth mentioning that, the present invention can also by analyzing user's communication behavioural information, new doubtful swindle number is obtained from the calling ticket in certain hour cycle T, doubtful swindle number calling behavioural characteristic subsequently that simultaneously can also be new according to these, precisely identify swindle molecule do not re-use or do not re-use in a period of time abandon number, thus while effectively improving system works efficiency, system loading can also be reduced further.The present invention can also include swindle number analysis device, and described swindle number analysis device includes further:
Sample group builds parts, for from the calling ticket in certain hour cycle T, extract several calling numbers, build number sample group to be confirmed and confirm swindle sample group, and judge whether the calling number of each extraction is tackling in directory, when calling number is when tackling in directory, then described calling number is divided into and confirms swindle sample group, when calling number is not when tackling in directory, then described calling number is divided into number sample group to be confirmed;
Swindle Number Reorganization parts, for extracting number sample group to be confirmed and confirmed to swindle the N number of calling parameter value in sample group corresponding to all calling numbers from the calling ticket in period of time T, described calling parameter value can include but not limited to: numbers of calling and called parties, calling time started, end of calling time, call duration, ring duration or calling and called ownership area code etc.; Then a BP neural net is built, with N number of calling parameter value of all calling numbers in number sample group to be confirmed for input, and according to confirming that the N number of calling parameter value swindling all calling numbers in sample group carrys out calculation expectation output valve, train described BP neural net; The last N number of calling parameter value extracting each calling number from the calling ticket in period of time T one by one, and be input to described BP neural net to calculate its output valve, when error between described output valve and desired output is within the scope of certain numerical value, then illustrate that described calling number is doubtful swindle number, calling number is write in doubtful directory;
Non-swindle number rejects parts, for all doubtful swindle number in the doubtful directory of new write in a period of time T is formed an outlier analysis bank, and extract the number of calls in outlier analysis bank in K the continuous print period of time T of all doubtful swindle numbers in period of time T and after period of time T; According to the number of calls downward trend in K the continuous time cycle T of doubtful swindle number after period of time T, judge whether each doubtful swindle number in outlier analysis bank is abandoned by swindle one's share of expenses for a joint undertaking one by one, if so, then the doubtful swindle number abandoned by swindle one's share of expenses for a joint undertaking is deleted from doubtful directory.
In swindle Number Reorganization parts, the computing formula of desired output σ is as follows:
m has confirmed to swindle the calling number number in sample group,
the neuronic output valve sum of output layer having confirmed that N number of calling parameter value of all calling numbers in swindle sample group obtains as the input value of BP neural net, θ
iconfirmed to swindle the calling number Z in sample group
in number of calling parameter value as the input value of BP neural net and the neuronic output valve of output layer obtained.
As shown in Figure 7, non-swindle number rejecting parts include further:
Number of calls accumulated unit, for calculating the number of calls sum in K the continuous time cycle T of doubtful swindle number after period of time T
and judge whether number of calls sum is 0, if, then represent that doubtful swindle number is swindled one's share of expenses for a joint undertaking and abandoned, wherein, CallingTA (t) is the number of calls in t the period of time T of doubtful swindle number after period of time T, K is a natural number, and its value can set according to actual conditions, such as K=5;
The average suppression ratio unit of total number of calls, for calculating the average suppression ratio Calling_minus_ratio of total number of calls of doubtful swindle number:
Wherein, CallingT is the number of calls of doubtful swindle number in period of time T, CallingTA (t) is the number of calls in t the period of time T of doubtful swindle number after period of time T, and judge whether Calling_minus_ratio is less than the minimum threshold of the average suppression ratio of total number of calls, if so, then represent that doubtful swindle number is swindled one's share of expenses for a joint undertaking and abandoned;
Often organize number of calls suppression ratio unit, for calculating the least square slope ratio Slope_ratio of doubtful swindle number:
Wherein, Calling_ratio (t) is the number of calls suppression ratio in t the period of time T of doubtful swindle number after period of time T, and further, the computing formula of Calling_ratio (t) is as follows:
And judge whether that Slope_ratio is at (A
1, A
2) numerical intervals within the scope of and Calling_minus_ratio is less than the max-thresholds of the average suppression ratio of total number of calls, if so, then represent that doubtful swindle number is swindled one's share of expenses for a joint undertaking and abandoned, wherein, A
1, A
2it is minimum, the large threshold value of least square slope ratio.
The most large and small threshold value of the average suppression ratio of total number of calls and least square slope ratio can according to actual conditions value.
The present invention has carried out test of many times at existing network, according to test effect, as K=5, max-thresholds=-0.5013 of minimum threshold=-0.7443 of total average suppression ratio of the number of calls, the always average suppression ratio of the number of calls, A
1=-0.2906, A
2when=0.1257, the present invention can obtain good effect, precisely identify swindle one's share of expenses for a joint undertaking do not re-use or do not re-use in a period of time abandon number.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within the scope of protection of the invention.
Claims (16)
1. a method for intelligent management and control fraudulent call, is characterized in that, include:
The calling that step one, calling subscriber initiate is triggered to call control node SCP, SCP inquires about calling number whether in doubtful directory or interception directory, then before called number, corresponding prefix identifier is increased, finally calling is forwarded to independent intelligent peripheral IIP, wherein, doubtful directory is for preserving doubtful swindle number, and interception directory is for preserving confirmed swindle number, and described prefix identifier is for identifying calling number whether in doubtful directory or interception directory;
Step 2, IIP read the prefix identifier before called number, when described prefix identifier mark calling number is when tackling in directory, then simulate called button to connect the operator attendance of calling terminal or to simulate called receiving calls.
2. method according to claim 1, is characterized in that, step one includes further:
Step 11, SCP receive the call request that calling subscriber initiates, and judge calling number whether in doubtful directory, if so, then before called number, increase the prefix identifier of doubtful identification, then calling continued to be forwarded to IIP, this flow process terminates; If not, then next step is continued;
Step 12, SCP judge that whether calling number is tackling in directory, if so, then continues next step; If not, then continue to perform original operation flow, this flow process terminates;
Step 13, SCP judge whether the intercept information of tackling calling number described in directory is mated, and if so, then before called number, increase the prefix identifier of analogue-key, then calling continued to be forwarded to IIP, this flow process terminates; If not, then next step is continued;
Step 14, SCP judge whether the intercept information of tackling calling number described in directory is do not mate, and if so, then before called number, increase the prefix identifier of analog answer, then calling continued to be forwarded to IIP, this flow process terminates; If not, then increase before called number and forward artificial prefix identifier, then calling continued to be forwarded to IIP, this flow process terminates.
3. method according to claim 1, is characterized in that, step 2 includes further:
Step 21, IIP to judge in call request before called number being the whether prefix identifier of analogue-key, if so, then continue step 22; If not, then step 23 is continued;
Step 22, IIP search the callback number of whether preserving calling number in call request in tables of data, if had, from tables of data, then extract a callback number of calling number, touch-tone signal corresponding for described callback number is encapsulated in APM message, then by described APM Message Transmission to calling terminal, select preset true man's recording to play to calling terminal again, thus start switching manual service after making calling terminal identification, this flow process terminates; If not, then step 24 is turned to;
Step 23, IIP to judge in call request before called number being the whether prefix identifier of analog answer, if so, then continue step 24;
Step 24, IIP judge in current time period, whether the actual response rate of analog answer is less than the response rate of system intialization, if, then according in upper a period of time, the ring duration that the doubtful swindle of several in doubtful directory number sampling statistics goes out is come to calling terminal coloring ring back tone playing or common ring-back tone, then answer charging message ANM is sent to originating switch, and play to calling terminal to select quiet or true man's recording according to the response duration gone out doubtful swindle number sampling statistics in upper a period of time, thus simulate called connection and the scene of response; If not, then according to the call completing rate gone out doubtful swindle number sampling statistics in upper a period of time, unreachable rate with close probability, simulate the scene of called connection nonreply, unreachable or shutdown to calling terminal.
4. method according to claim 3, is characterized in that, in step 24, in current time period, the computing formula of the actual response rate of analog answer is as follows:
wherein, MYDL (T) is the actual response rate of analog answer in current time period T, MYDN (T) is that in current time period T, IIP simulates called connection and the calls of response, and SUNC (T) is the calls in current time period T received by IIP
In each time cycle, the computing formula of the response rate of system intialization is: XYDL (T)=YDL (T-1)+n%, wherein, XYDL (T) is the response rate of system intialization in current time period T, YDL (T-1) is the actual response rate that upper a period of time T-1 goes out the doubtful swindle of several in doubtful directory number sampling statistics, and n% is that analog answer rate improves percentage.
5. method according to claim 1, is characterized in that, step 2 also includes:
IIP to judge in call request before called number being the whether prefix identifier of doubtful identification, if, then bridge joint master, voice channel between called, then unidirectional recording is carried out to caller voice, the recording file reaching certain long recording time is changed into text, then each swindle text in described text and swindle Sample Storehouse is mated one by one, if the matching degree of a swindle text of described text and swindle Sample Storehouse exceedes certain threshold value, then illustrate that the calling number that described text is corresponding is swindle number, described calling number is moved to interception directory from doubtful directory, and identify its intercept information for mate, detect in described text simultaneously and whether have continuous print numeral, and continuous print numeral number is more than or equal to clawback bond number minimum value, if had, then digital for this string continuous print callback number as described calling number is saved in tables of data, when having multiple callback number in text, all callback numbers are all saved in tables of data, finally remove this main, voice channel between called, this flow process terminates, when the matching degree of all swindle texts in described text and swindle Sample Storehouse is all lower than threshold value, then continue next step,
Whether the calling number in the request of IIP detection of call is spacing, if, then calling number is moved to interception directory from doubtful directory, and identify its intercept information for not mate, described text is saved in swindle Sample Storehouse simultaneously, then removes the voice channel between these calling and called; If not, then manual examination and verification are carried out to be confirmed whether being fraudulent call to recording text, if fraudulent call is confirmed as in manual examination and verification, then calling number is moved to interception directory from doubtful directory, and identify its intercept information for manually to determine.
6. method according to claim 1, is characterized in that, also includes:
Steps A 1, from the calling ticket in certain hour cycle T, extract several calling numbers, build number sample group to be confirmed and confirm swindle sample group, and judge whether the calling number of each extraction is tackling in directory, when calling number is when tackling in directory, then described calling number being divided into and confirming swindle sample group, when calling number is not when tackling in directory, then described calling number being divided into number sample group to be confirmed;
Steps A 2, from the calling ticket in period of time T, extract number sample group to be confirmed and confirmed to swindle the N number of calling parameter value in sample group corresponding to all calling numbers;
Steps A 3, a structure BP neural net, with N number of calling parameter value of all calling numbers in number sample group to be confirmed for input, and according to confirming that the N number of calling parameter value swindling all calling numbers in sample group carrys out calculation expectation output valve, train described BP neural net;
Steps A 4, from the calling ticket in period of time T, extract N number of calling parameter value of each calling number one by one, and be input to described BP neural net to calculate its output valve, when error between described output valve and desired output is within the scope of certain numerical value, then illustrate that described calling number is doubtful swindle number, calling number is write in doubtful directory.
7. method according to claim 6, is characterized in that, in steps A 3, the computing formula of desired output σ is as follows:
m has confirmed to swindle the calling number number in sample group,
the neuronic output valve sum of output layer having confirmed that N number of calling parameter value of all calling numbers in swindle sample group obtains as the input value of BP neural net, θ
iconfirmed to swindle the calling number Z in sample group
in number of calling parameter value as the input value of BP neural net and the neuronic output valve of output layer obtained.
8. method according to claim 1, is characterized in that, also includes:
Step B1, all doubtful swindle number in the doubtful directory of new write in a period of time T is formed an outlier analysis bank, and extract the number of calls in outlier analysis bank in K the continuous print period of time T of all doubtful swindle numbers in period of time T and after period of time T;
Step B2, from outlier analysis bank extract a doubtful swindle number;
Step B3, the number of calls sum judged in K the continuous print period of time T of described doubtful swindle number after period of time T
whether be 0, if so, then described doubtful swindle number deleted from doubtful directory, and turn to step B6; If not, then next step is continued; Wherein, CallingTA (t) is the number of calls in t the period of time T of doubtful swindle number after period of time T;
Step B4, calculate the average suppression ratio Calling_minus_ratio of total number of calls of described doubtful swindle number:
Wherein, CallingT is the number of calls of doubtful swindle number in period of time T, and judges whether Calling_minus_ratio is less than the minimum threshold of the average suppression ratio of total number of calls, if, then described doubtful swindle number is deleted from doubtful directory, and turn to step B6; If not, then next step is continued;
Step B5, calculate the least square slope ratio Slope_ratio of described doubtful swindle number:
Wherein, Calling_ratio (t) is the number of calls suppression ratio in t the period of time T of doubtful swindle number after period of time T, and further, the computing formula of Calling_ratio (t) is as follows:
and judge whether that Slope_ratio is at (A
1, A
2) numerical intervals within the scope of and Calling_minus_ratio is less than the max-thresholds of the average suppression ratio of total number of calls, if so, then described doubtful swindle number is deleted from doubtful directory, then continues next step; If not, then next step is continued; Wherein, A
1, A
2it is minimum, the large threshold value of least square slope ratio;
Step B6, judge whether all doubtful swindle number that extracted in outlier analysis bank, if so, then this flow process terminates; If not, then step B2 is turned to.
9. a system for intelligent management and control fraudulent call, is characterized in that, include:
Call control node SCP, for receiving the call request that calling subscriber initiates, and whether calling number in query call request is in doubtful directory or interception directory, then before called number, corresponding prefix identifier is increased, finally calling is forwarded to IIP, wherein, doubtful directory is for preserving doubtful swindle number, interception directory is for preserving confirmed swindle number, and described prefix identifier is for identifying calling number whether in doubtful directory or interception directory;
Independent intelligent peripheral IIP, the call request of coming is forwarded for receiving SCP, and the prefix identifier read in call request before called number, when described prefix identifier mark calling number is when tackling in directory, then simulate called button to connect the operator attendance of calling terminal or to simulate called receiving calls.
10. system according to claim 9, is characterized in that, described SCP includes further:
Prefix identifier device, for judging calling number in call request whether in doubtful directory or interception directory, when calling number is in doubtful directory, then increases the prefix identifier of doubtful identification before called number; When calling number to tackle in directory and its intercept information be mate time, then before called number, increase the prefix identifier of analogue-key; When calling number to tackle in directory and its intercept information be do not mate time, then before called number, increase the prefix identifier of analog answer; When calling number to tackle in directory and its intercept information be not mated and do not mated time, then increase before called number and forward artificial prefix identifier, finally calling is continued to be forwarded to IIP.
11. systems according to claim 9, is characterized in that, described IIP includes further:
Analogue-key device, for when before called number in call request being the prefix identifier of analogue-key, search the callback number of whether preserving calling number in call request in tables of data, if had, from tables of data, then extract a callback number of calling number, touch-tone signal corresponding for described callback number is encapsulated in APM message, then by described APM Message Transmission to calling terminal, select preset true man's recording to play to calling terminal again, thus after making calling terminal identification, start switching manual service; Or,
Analog answer device, for when before called number in call request being the prefix identifier of analog answer, judge in current time period, whether the actual response rate of analog answer is less than the response rate of system intialization, if, then according in upper a period of time, the ring duration that the doubtful swindle of several in doubtful directory number sampling statistics goes out is come to calling terminal coloring ring back tone playing or common ring-back tone, then answer charging message ANM is sent to originating switch, and play to calling terminal to select quiet or true man's recording according to the response duration gone out doubtful swindle number sampling statistics in upper a period of time, thus simulate called connection and the scene of response, if not, then according to the call completing rate gone out doubtful swindle number sampling statistics in upper a period of time, unreachable rate with close probability, simulate the scene of called connection nonreply, unreachable or shutdown to calling terminal.
12. systems according to claim 11, is characterized in that, in analog answer device, in current time period, the computing formula of the actual response rate of analog answer is as follows:
wherein, MYDL (T) is the actual response rate of analog answer in current time period T, MYDN (T) is that in current time period T, IIP simulates called connection and the calls of response, and SUNC (T) is the calls in current time period T received by IIP
In each time cycle, the computing formula of the response rate of system intialization is: XYDL (T)=YDL (T-1)+n%, wherein, XYDL (T) is the response rate of system intialization in current time period T, YDL (T-1) is the actual response rate that upper a period of time T-1 goes out the doubtful swindle of several in doubtful directory number sampling statistics, and n% is that analog answer rate improves percentage.
13. systems according to claim 9, it is characterized in that, IIP also includes:
Doubtful recognition device, for when before called number in call request being the prefix identifier of doubtful identification, bridge joint master, voice channel between called, and unidirectional recording is carried out to caller voice, the recording file reaching certain long recording time is changed into text, then each swindle text in described text and swindle Sample Storehouse is mated one by one, if the matching degree of a swindle text of described text and swindle Sample Storehouse exceedes certain threshold value, then illustrate that the calling number that described text is corresponding is swindle number, described calling number is moved to interception directory from doubtful directory, and identify its intercept information for mate, detect in described text simultaneously and whether have continuous print numeral, and continuous print numeral number is more than or equal to clawback bond number minimum value, if had, then digital for this string continuous print callback number as described calling number is saved in tables of data, when having multiple callback number in text, all callback numbers are all saved in tables of data, finally remove this main, voice channel between called, when the matching degree of all swindle texts in described text and swindle Sample Storehouse is all lower than threshold value, then notify vacant number detection device,
Vacant number detection device, whether be spacing for the calling number in detection of call request, if, then calling number is moved to interception directory from doubtful directory, and identify its intercept information for not mate, described text is saved in swindle Sample Storehouse simultaneously, then removes the voice channel between these calling and called; If not, then manual examination and verification are carried out to be confirmed whether being fraudulent call to recording text, if fraudulent call is confirmed as in manual examination and verification, then calling number is moved to interception directory from doubtful directory, and identify its intercept information for manually to determine.
14. systems according to claim 9, is characterized in that, also include swindle number analysis device, described swindle number analysis device includes further:
Sample group builds parts, for from the calling ticket in certain hour cycle T, extract several calling numbers, build number sample group to be confirmed and confirm swindle sample group, and judge whether the calling number of each extraction is tackling in directory, when calling number is when tackling in directory, then described calling number is divided into and confirms swindle sample group, when calling number is not when tackling in directory, then described calling number is divided into number sample group to be confirmed;
Swindle Number Reorganization parts, for extracting number sample group to be confirmed and having confirmed to swindle the N number of calling parameter value in sample group corresponding to all calling numbers from the calling ticket in period of time T; Then a BP neural net is built, with N number of calling parameter value of all calling numbers in number sample group to be confirmed for input, and according to confirming that the N number of calling parameter value swindling all calling numbers in sample group carrys out calculation expectation output valve, train described BP neural net; The last N number of calling parameter value extracting each calling number from the calling ticket in period of time T one by one, and be input to described BP neural net to calculate its output valve, when error between described output valve and desired output is within the scope of certain numerical value, then illustrate that described calling number is doubtful swindle number, calling number is write in doubtful directory
In described swindle Number Reorganization parts, the computing formula of desired output σ is as follows:
m has confirmed to swindle the calling number number in sample group,
the neuronic output valve sum of output layer having confirmed that N number of calling parameter value of all calling numbers in swindle sample group obtains as the input value of BP neural net, θ
iconfirmed to swindle the calling number Z in sample group
in number of calling parameter value as the input value of BP neural net and the neuronic output valve of output layer obtained.
15. systems according to claim 9, is characterized in that, also include swindle number analysis device, described swindle number analysis device includes further:
Non-swindle number rejects parts, for all doubtful swindle number in the doubtful directory of new write in a period of time T is formed an outlier analysis bank, and extract the number of calls in outlier analysis bank in K the continuous print period of time T of all doubtful swindle numbers in period of time T and after period of time T; According to the number of calls downward trend in K the continuous time cycle T of doubtful swindle number after period of time T, judge whether each doubtful swindle number in outlier analysis bank is abandoned by swindle one's share of expenses for a joint undertaking one by one, if so, then the doubtful swindle number abandoned by swindle one's share of expenses for a joint undertaking is deleted from doubtful directory.
16. systems according to claim 15, is characterized in that, non-swindle number is rejected parts and included further:
Number of calls accumulated unit, for calculating the number of calls sum in K the continuous time cycle T of doubtful swindle number after period of time T
and judge whether described number of calls sum is 0, if so, then represent that doubtful swindle number is swindled one's share of expenses for a joint undertaking and abandoned, wherein, CallingTA (t) is the number of calls in t the period of time T of doubtful swindle number after period of time T;
The average suppression ratio unit of total number of calls, for calculating the average suppression ratio Calling_minus_ratio of total number of calls of doubtful swindle number:
Wherein, CallingT is the number of calls of doubtful swindle number in period of time T, CallingTA (t) is the number of calls in t the period of time T of doubtful swindle number after period of time T, and judge whether Calling_minus_ratio is less than the minimum threshold of the average suppression ratio of total number of calls, if so, then represent that doubtful swindle number is swindled one's share of expenses for a joint undertaking and abandoned;
Often organize number of calls suppression ratio unit, for calculating the least square slope ratio Slope_ratio of doubtful swindle number:
Wherein, Calling_ratio (t) is the number of calls suppression ratio in t the period of time T of doubtful swindle number after period of time T, and further, the computing formula of Calling_ratio (t) is as follows:
and judge whether that Slope_ratio is at (A
1, A
2) numerical intervals within the scope of and Calling_minus_ratio is less than the max-thresholds of the average suppression ratio of total number of calls, if so, then represent that doubtful swindle number is swindled one's share of expenses for a joint undertaking and abandoned, wherein, A
1, A
2it is minimum, the large threshold value of least square slope ratio.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510190578.2A CN104936182B (en) | 2015-04-21 | 2015-04-21 | A kind of method and system of intelligence management and control fraudulent call |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510190578.2A CN104936182B (en) | 2015-04-21 | 2015-04-21 | A kind of method and system of intelligence management and control fraudulent call |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104936182A true CN104936182A (en) | 2015-09-23 |
| CN104936182B CN104936182B (en) | 2018-05-25 |
Family
ID=54123085
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510190578.2A Active CN104936182B (en) | 2015-04-21 | 2015-04-21 | A kind of method and system of intelligence management and control fraudulent call |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104936182B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106327221A (en) * | 2016-09-30 | 2017-01-11 | 北京中科寒武纪科技有限公司 | Device and method for preventing telemarketing scam |
| CN106850567A (en) * | 2016-12-29 | 2017-06-13 | 恒安嘉新(北京)科技有限公司 | A kind of method and device that doubtful harmful call treatment is realized based on preceding slotting code |
| DE102016106018A1 (en) * | 2016-04-01 | 2017-10-05 | Deutsche Telekom Ag | Telephony communication system for detecting abuse in a public telephone network |
| CN108234395A (en) * | 2016-12-14 | 2018-06-29 | 中兴通讯股份有限公司 | A kind of call monitoring method and device |
| CN108737620A (en) * | 2017-04-25 | 2018-11-02 | 中国移动通信集团重庆有限公司 | Voice call hold-up interception method and device |
| CN109600520A (en) * | 2017-09-30 | 2019-04-09 | 上海触乐信息科技有限公司 | Harassing call number identification method, device and equipment |
| CN110072016A (en) * | 2019-01-29 | 2019-07-30 | 浙江鹏信信息科技股份有限公司 | A method of bad Classification of Speech is realized using call behavior time-domain filtering |
| CN110177179A (en) * | 2019-05-16 | 2019-08-27 | 国家计算机网络与信息安全管理中心 | A kind of swindle number identification method based on figure insertion |
| CN110581924A (en) * | 2018-06-07 | 2019-12-17 | 中国电信股份有限公司 | Method and system for prompting risk of fraud |
| CN110730150A (en) * | 2018-07-16 | 2020-01-24 | 中国移动通信有限公司研究院 | Call control method, session control center and management and control platform |
| CN110798330A (en) * | 2018-08-01 | 2020-02-14 | 中国移动通信集团浙江有限公司 | Telecommunication fraud library updating processing method and device |
| CN111813811A (en) * | 2020-06-30 | 2020-10-23 | 杭州科度科技有限公司 | Telecommunication fraud early warning method, device, electronic equipment and medium |
| CN113692022A (en) * | 2021-07-09 | 2021-11-23 | 深圳市赛云数据有限公司 | Operator communication system based on empty channel pre-judgment utilization |
| CN113766065A (en) * | 2021-09-09 | 2021-12-07 | 上海欣方智能系统有限公司 | IMS network fraud call interception based realization method and system |
| CN113992801A (en) * | 2020-07-27 | 2022-01-28 | 中国移动通信集团有限公司 | Violation number identification method and device, storage medium and computer equipment |
| CN117132392A (en) * | 2023-10-23 | 2023-11-28 | 蓝色火焰科技成都有限公司 | Vehicle loan fraud risk early warning method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1999033285A1 (en) * | 1997-12-19 | 1999-07-01 | Ericsson Inc. | Architecture independent application invocation over a telephony network |
| WO2007140007A3 (en) * | 2006-05-26 | 2008-05-08 | Telcordia Tech Inc | Flexible rating rules and calender rules implemented in a real-time charging system for a telecommunications network |
| CN102143461A (en) * | 2010-02-02 | 2011-08-03 | 王滨 | Intelligent call forwarding method and device |
| CN102821370A (en) * | 2011-06-08 | 2012-12-12 | 中兴通讯股份有限公司 | Method and system for reporting suspicious numbers based on intelligent network |
| CN104244216A (en) * | 2014-09-29 | 2014-12-24 | 中国移动通信集团浙江有限公司 | Method and system for intercepting fraud phones in real time during calling |
| CN104469025A (en) * | 2014-11-26 | 2015-03-25 | 杭州东信北邮信息技术有限公司 | Clustering-algorithm-based method and system for intercepting fraud phone in real time |
-
2015
- 2015-04-21 CN CN201510190578.2A patent/CN104936182B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1999033285A1 (en) * | 1997-12-19 | 1999-07-01 | Ericsson Inc. | Architecture independent application invocation over a telephony network |
| WO2007140007A3 (en) * | 2006-05-26 | 2008-05-08 | Telcordia Tech Inc | Flexible rating rules and calender rules implemented in a real-time charging system for a telecommunications network |
| CN102143461A (en) * | 2010-02-02 | 2011-08-03 | 王滨 | Intelligent call forwarding method and device |
| CN102821370A (en) * | 2011-06-08 | 2012-12-12 | 中兴通讯股份有限公司 | Method and system for reporting suspicious numbers based on intelligent network |
| CN104244216A (en) * | 2014-09-29 | 2014-12-24 | 中国移动通信集团浙江有限公司 | Method and system for intercepting fraud phones in real time during calling |
| CN104469025A (en) * | 2014-11-26 | 2015-03-25 | 杭州东信北邮信息技术有限公司 | Clustering-algorithm-based method and system for intercepting fraud phone in real time |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10291772B2 (en) | 2016-04-01 | 2019-05-14 | Deutsche Telekon Ag | Telephony communications system for detecting abuse in a public telephone network |
| DE102016106018A1 (en) * | 2016-04-01 | 2017-10-05 | Deutsche Telekom Ag | Telephony communication system for detecting abuse in a public telephone network |
| DE102016106018B4 (en) * | 2016-04-01 | 2017-10-26 | Deutsche Telekom Ag | Telephony communication system for detecting abuse in a public telephone network |
| CN106327221A (en) * | 2016-09-30 | 2017-01-11 | 北京中科寒武纪科技有限公司 | Device and method for preventing telemarketing scam |
| CN108234395A (en) * | 2016-12-14 | 2018-06-29 | 中兴通讯股份有限公司 | A kind of call monitoring method and device |
| CN106850567A (en) * | 2016-12-29 | 2017-06-13 | 恒安嘉新(北京)科技有限公司 | A kind of method and device that doubtful harmful call treatment is realized based on preceding slotting code |
| CN108737620A (en) * | 2017-04-25 | 2018-11-02 | 中国移动通信集团重庆有限公司 | Voice call hold-up interception method and device |
| CN109600520B (en) * | 2017-09-30 | 2021-03-16 | 上海触乐信息科技有限公司 | Crank call number identification method, device and equipment |
| CN109600520A (en) * | 2017-09-30 | 2019-04-09 | 上海触乐信息科技有限公司 | Harassing call number identification method, device and equipment |
| CN110581924B (en) * | 2018-06-07 | 2021-07-30 | 中国电信股份有限公司 | Method and system for prompting risk of fraud |
| CN110581924A (en) * | 2018-06-07 | 2019-12-17 | 中国电信股份有限公司 | Method and system for prompting risk of fraud |
| CN110730150A (en) * | 2018-07-16 | 2020-01-24 | 中国移动通信有限公司研究院 | Call control method, session control center and management and control platform |
| CN110798330A (en) * | 2018-08-01 | 2020-02-14 | 中国移动通信集团浙江有限公司 | Telecommunication fraud library updating processing method and device |
| CN110072016A (en) * | 2019-01-29 | 2019-07-30 | 浙江鹏信信息科技股份有限公司 | A method of bad Classification of Speech is realized using call behavior time-domain filtering |
| CN110177179B (en) * | 2019-05-16 | 2020-12-29 | 国家计算机网络与信息安全管理中心 | Fraud number identification method based on graph embedding |
| CN110177179A (en) * | 2019-05-16 | 2019-08-27 | 国家计算机网络与信息安全管理中心 | A kind of swindle number identification method based on figure insertion |
| CN111813811A (en) * | 2020-06-30 | 2020-10-23 | 杭州科度科技有限公司 | Telecommunication fraud early warning method, device, electronic equipment and medium |
| CN113992801A (en) * | 2020-07-27 | 2022-01-28 | 中国移动通信集团有限公司 | Violation number identification method and device, storage medium and computer equipment |
| CN113692022A (en) * | 2021-07-09 | 2021-11-23 | 深圳市赛云数据有限公司 | Operator communication system based on empty channel pre-judgment utilization |
| CN113766065A (en) * | 2021-09-09 | 2021-12-07 | 上海欣方智能系统有限公司 | IMS network fraud call interception based realization method and system |
| CN117132392A (en) * | 2023-10-23 | 2023-11-28 | 蓝色火焰科技成都有限公司 | Vehicle loan fraud risk early warning method and system |
| CN117132392B (en) * | 2023-10-23 | 2024-01-30 | 蓝色火焰科技成都有限公司 | Vehicle loan fraud risk early warning method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104936182B (en) | 2018-05-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104936182A (en) | Method of managing and controlling fraud telephones intelligently and system of managing and controlling fraud telephones intelligently | |
| US10194022B2 (en) | System and method for automatically detecting undesired calls | |
| CN104244216B (en) | The method and system of real-time blocking fraudulent call in a kind of communication process | |
| CN103095889B (en) | Junk call intercepting system based on talk mode identification and operating method thereof | |
| CN107770777B (en) | Method for identifying recorded fraud calls | |
| WO2016197675A1 (en) | Method and apparatus for identifying crank call | |
| CN100456881C (en) | Subscriber identy identifying method and calling control method and system | |
| US8391445B2 (en) | Caller identification using voice recognition | |
| CN104317934B (en) | A kind of method and system of rapid batch outgoing call | |
| WO2000014730A1 (en) | Method for checking access authorisation for voice telephony at a fixed network or mobile telephone connection and corresponding communications network | |
| CN109151220A (en) | A kind of communication session call failure scene analysis system | |
| CN108366173A (en) | A kind of phone recognition methods, relevant device and system | |
| CN110072019A (en) | A kind of method and device shielding harassing call | |
| CN101674353A (en) | Incoming call filter method | |
| CN106936997B (en) | A kind of rubbish voice recognition methods and system based on social networks map | |
| CN110337059A (en) | A kind of parser, server and the network system of subscriber household relationship | |
| CN1937663A (en) | Method, system and device for realizing variable voice telephone business | |
| CN109859032A (en) | Funds on account collection method, apparatus, equipment and storage medium based on intelligent sound | |
| CN104410973A (en) | Recognition method and system for tape played phone fraud | |
| CN107231494A (en) | A kind of acquisition methods of user communication characteristic, storage medium and electronic equipment | |
| CN108989587A (en) | Application method of the IVR voice in the work of bank's post-loan management | |
| CN104580649B (en) | A kind of method and system for auditing automatic speech service content | |
| CN101127949A (en) | A method for realizing instant recording service based on mobile communication network | |
| CA2456667A1 (en) | Methods and systems for providing on-line bills for use in communications services | |
| CN106453927A (en) | A number state automatic identification method and apparatus based on a PSTN network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder |
Address after: No. 19, Jiefang East Road, Hangzhou, Zhejiang Province, 310016 Patentee after: CHINA MOBILE GROUP ZHEJIANG Co.,Ltd. Patentee after: Xinxun Digital Technology (Hangzhou) Co.,Ltd. Address before: No. 19, Jiefang East Road, Hangzhou, Zhejiang Province, 310016 Patentee before: CHINA MOBILE GROUP ZHEJIANG Co.,Ltd. Patentee before: EB Information Technology Ltd. |
|
| CP01 | Change in the name or title of a patent holder |