CN103716471A - User call behavior model generating method applicable to spam voice filtering - Google Patents

Abstract

The invention discloses a user call behavior model generating method applicable to spam voice filtering. The user call behavior model generating method comprises the following steps that: call interactive behavior characteristics (CI) are established and are used for describing related behavior characteristics of specified users which are adopted as a calling user and a called user, wherein the call interactive behavior characteristics (CI) further include incoming call/outgoing call ratios, call interactive recording characteristic value and interaction strength, and distribution of the incoming call/outgoing call ratios and the call interactive recording characteristic value and interaction strength; call frequency and distribution thereof FCD are established and are used for describing the characteristics of call time in call records, wherein the characteristics of the call time in call records include call frequency values and call frequency value time distribution in a specified statistical duration; and call lasting time and distribution thereof DCT are established and are used for describing the characteristics of lasting time in the call records, wherein the characteristics of the lasting time in the call records include call answering refusal proportions, call average lasting time and histogram distribution of the call lasting time. With the user call behavior model generating method applicable to the spam voice filtering of the invention adopted, a technical problem that a camouflaged call behavior of a spam voice transmitter is hard to find can be solved.

Description

A kind of user who is applicable to rubbish voice filtering calls out the generation method of behavior model

Technical field

The invention belongs to rubbish voice filtering and Data Mining, more specifically, relate to a kind of generation method that user who is applicable to rubbish voice filtering calls out behavior model.

Background technology

Along with the combination of fixed network, mobile communications network and the Internet, voice service is widely used.But due to the impact of rubbish voice, voice service has run into the obstruction of business development, when guaranteeing user's proper communication, need in time malicious user to be limited.Existing rubbish voice filtering technology is improved on the basis of spam and filtering junk short messages technology, can play the effect that detects and filter to rubbish voice to a certain extent.But the content character of rubbish voice is different from spam, and the filtering technique of spam has certain limitation.Spam mostly is text filtering, and rubbish voice content is multimedia messages.Spam filtering allows time delay, and rubbish voice is very high to the requirement of real-time.

Effective and reasonable strobe utility requires few as much as possible with calling and called user's interchange, can adopt the filter method based on call model.Call model sends the behavioural characteristic of calling according to user, reflect objectively whether user is rubbish voice.Existing call model has proposed the rubbish voice feature of observing from calling behavior in a large number, and adopts decision tree or the method for Bayes classifier, realizes rubbish voice filtering.User, call out in the immovable situation of behavior, existing call model is mature on the whole effectively.But the existing testing mechanism based on call model is difficult to find the calling behavior of rubbish voice sender camouflage, has certain defect.

Summary of the invention

Above defect or Improvement requirement for prior art, the invention provides a kind of generation method that user who is applicable to rubbish voice filtering calls out behavior model, its object is, solves the technical problem of the calling behavior of the very difficult discovery rubbish voice sender camouflage existing in prior art.

For achieving the above object, according to one aspect of the present invention, a kind of generation method that provides user who is applicable to rubbish voice filtering to call out behavior model comprises the following steps:

(1) set up and call out interbehavior feature CI, for describing designated user as calling subscriber and called subscriber's corelation behaviour feature, wherein call out interbehavior feature CI and further comprise again incoming call exhalation ratio, call out intersection record characteristic value and mutual intensity and three parts that distribute thereof, specifically can be expressed as:

CI＝{R _in/out，C _out，C _in，C _in/out，F _in/out}

Wherein, R _in/outfor the ratio of user as calling subscriber and called subscriber, C _outthat user only has the quantity of exhalation behavior, C in the calling of all and different user as calling subscriber _inthat user only has the quantity of the behavior of answering, C in the calling of all and different user as called subscriber _in/outuser in the calling of all and different user simultaneously as calling subscriber and called subscriber's quantity, F _in/outit is the mutual intensity frequency distribution with other users;

(2) set up calling frequency and distribution F thereof _cD, for describing the call record feature of call time, comprising the interior calling frequency value of timing statistics section and calling frequency value time distribution two parts of appointment, specifically can be expressed as:

$F_{\mathrm{CD}}=\left\{F_{\mathrm{in}/\mathrm{out}}^T{,\stackrel{\‾}{D}}_{\mathrm{out}}^T\right\}$

Wherein,

for the absolute frequency value that in timing statistics section, user makes a call,

be user's calling frequency in timing statistics section be the distribution in the timeslice of 2 hours one day 12 length.

(3) set up call duration and distribution D thereof _cT, for describing the feature of call record duration, comprising histogram distribution three parts that in timing statistics section, call denial is answered ratio, called out average duration and call duration, specifically can be expressed as:

$D_{\mathrm{CT}}=\{f_E^T,{\mathrm{CT}}_{\mathrm{avg}}^T,{\mathrm{CTD}}^T\}$

Wherein,

for user is rejected the probable value of answering in making a call,

for user's mean call duration, CTD ^tdistribution for call duration.

Preferably, step (1) specifically comprises following sub-step:

(1-1) history of counting user is called out interbehavior characteristic parameter C _out, C _inand C _in/out, for the ease of across comparison between user, further it is normalized:

${\stackrel{\‾}{C}}_{\mathrm{out}}=\frac{C_{\mathrm{out}}}{C_{\mathrm{out}}+C_{\mathrm{in}}+C_{\mathrm{in}/\mathrm{out}}}$

${\stackrel{\‾}{C}}_{\mathrm{in}}=\frac{C_{\mathrm{in}}}{C_{\mathrm{out}}+C_{\mathrm{in}}+C_{\mathrm{in}/\mathrm{out}}}$

${\stackrel{\‾}{C}}_{\mathrm{in}/\mathrm{out}}=\frac{C_{\mathrm{in}/\mathrm{out}}}{C_{\mathrm{out}}+C_{\mathrm{in}}+C_{\mathrm{in}/\mathrm{out}}}$

Wherein, with

three's sum is 1.

(1-2) statistics incoming call exhalation ratio R _in/out, computing formula is as follows:

$R_{\mathrm{in}/\mathrm{out}}=\frac{C_{\mathrm{out}}}{C_{\mathrm{in}}}$

(1-3) statistics designated user and other users' the mutual intensity CD of calling _in/out.The mutual intensity of calling between this user and user j is expressed as

its computing formula is as follows:

${\mathrm{CD}}_{\mathrm{in}/\mathrm{out}}^j=\mathrm{INT}\left[\log (C_{\mathrm{out}}^j+C_{\mathrm{in}}^j)\right]$

Wherein, INT[] expression bracket function, the number of times that represents this user's active call user j,

represent that this user answers the number of calls from user j.

(1-4) statistics is called out mutual intensity distributions CDD _in/out

CDD _in/out＝{CDN ₀，CDN ₁，CDN ₂，CDN ₃}

Wherein, CDN _ithe computing formula of (i=0,1,2,3) is as follows:

Be CDN _ifor value equals i or is more than or equal to the CD of i _in/outquantity, COUNT[wherein] be counting function,

represent the mutual intensity of calling between this user and user j, n represents this user All Contacts's quantity.

(1-5) to CDD _in/outbe normalized, the mutual intensity distributions of calling after normalization is designated as

specifically be expressed as:

${\stackrel{\‾}{\mathrm{CDN}}}_{\mathrm{in}/\mathrm{out}}=\{{\stackrel{\‾}{\mathrm{CDN}}}_0,{\stackrel{\‾}{\mathrm{CDN}}}_1,{\stackrel{\‾}{\mathrm{CDN}}}_2,{\stackrel{\‾}{\mathrm{CDN}}}_3\}$

Wherein,

for the CDN after normalization _i(i=0,1,2,3), computing formula is:

${\stackrel{\‾}{\mathrm{CDN}}}_i=\frac{{\mathrm{CDN}}_i}{{\mathrm{\Σ}}_{k+0}^a{\mathrm{CDN}}_k}.$

Preferably, step (2) specifically comprises following sub-step:

(2-1) absolute frequency value that counting user makes a call

$F_{\mathrm{in}/\mathrm{out}}^T=\frac{C_{\mathrm{out}}^T}{T}$

Wherein, the timing statistics segment length that T is appointment, unit is hour that this time period should start from 0 o'clock of first day, end at 24 o'clock of last day, so T should be the integral multiple of 24 hours;

for the number of calls of user in the time period as calling subscriber.

(2-2) the calling frequency distributed constant of the designated user in statistics fixed time section T

specifically be expressed as:

$D_{\mathrm{out}}^T=\{D_{\mathrm{out}}^1,D_{\mathrm{out}}^2,...,D_{\mathrm{out}}^{12},\}$

Wherein,

[2* (t-1), 2*t) the call number sum in period that represent every day in timing statistics section T.Time every day can be divided into 12 timeslices, and each timeslice includes 2 hours, and 12 timeslices are specially: [0,2), [2,4), [4,6), [6,8), [8,10), [10,12), [12,14), [14,16), [16,18), [18,20), [20,22), [22,24).

(2-3) to user's calling frequency distributed constant

be normalized, processing procedure is as follows:

${\stackrel{\‾}{D}}_{\mathrm{out}}^t=\frac{D_{\mathrm{out}}^t}{C_{\mathrm{out}}^T},(t=1,...,12)$

The calling frequency after normalization is distributed as:

${\stackrel{\‾}{D}}_{\mathrm{out}}^T=\{{\stackrel{\‾}{D}}_{\mathrm{out}}^1,{\stackrel{\‾}{D}}_{\mathrm{out}}^2,...,{\stackrel{\‾}{D}}_{\mathrm{out}}^{12},\}.$

(2-4) obtain finally having calling frequency and the distributed constant F thereof of service efficiency _cD

$F_{\mathrm{CD}}=\{F_{\mathrm{in}/\mathrm{out}}^T,{\stackrel{\‾}{D}}_{\mathrm{out}}^T\}.$

Preferably, step (3) specifically comprises following sub-step:

(3-1) counting user is rejected the ratio of answering in making a call

computing formula is as follows:

$f_E^T=\frac{{\mathrm{CR}}_{\mathrm{out}}^T}{C_{\mathrm{out}}^T}$

Wherein, represent the all-calls number of times that in timing statistics section, user sends,

represent unaccepted active call number of times in the statistical piece time;

(3-2) mean call duration of counting user

computing formula is as follows:

${\mathrm{CT}}_{\mathrm{avg}}^T=\frac{{\mathrm{\Σ}}_{r=1}^{C_{\mathrm{in}/\mathrm{out}}^T}{t}_r}{C_{\mathrm{in}/\mathrm{out}}^T}$

Wherein,

represent the number of calls of statistical piece in the time, t _rbe call duration corresponding to r bar call record, unit is second;

(3-3) the histogram distribution CTD of statistics call duration ^t, be specifically expressed as follows:

${\mathrm{CTD}}^T=\{{\mathrm{CT}}_D^0,{\mathrm{CT}}_D^1,...,{\mathrm{CT}}_D^6,{\mathrm{CT}}_D^7\}$

Wherein,

represent to call out the call record ratio of lasting duration within certain time period, specific formula for calculation is as follows:

In general, the above technical scheme of conceiving by the present invention compared with prior art, can obtain following beneficial effect:

1, the present invention has completed and has been applicable to the foundation that the user of rubbish voice filtering calls out behavior model, provides user to call out the computational methods of behavioural characteristic.

2, the present invention is directed to the inadequate defect of original call behavioral data, proposed the definition mode of three kinds of reinforced characteristic parameter combinations.

3, the present invention is based on the calling and called user's of calling procedure calling interaction feature, except basic incoming call exhalation ratio, also consider mutual intensity and distribution situation thereof, add and understand the feature that calling is mutual, be more suitable for the filtration of rubbish voice.

Accompanying drawing explanation

Fig. 1 is that the user that the present invention is applicable to rubbish voice filtering calls out the schematic diagram of the generation method of behavior model.

Fig. 2 is the refinement flow chart of step in the inventive method (1).

Fig. 3 is the refinement flow chart of step in the inventive method (2).

Fig. 4 is the refinement flow chart of step in the inventive method (3).

Embodiment

In order to make object of the present invention, technical scheme and advantage clearer, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not intended to limit the present invention.In addition,, in each execution mode of described the present invention, involved technical characterictic just can not combine mutually as long as do not form each other conflict.

The model that description user provided by the invention calls out behavior comprises a plurality of calling behavior basic parameters, mainly four parameters such as calling subscriber, called subscriber, call time, duration, consists of.On the basis of these four parameters, further built the calling behavioural characteristic of three different enhancings, comprised respectively and call out interaction feature CI, calling frequency and distribution F thereof _cD, call duration and distribution D thereof _cT, each call features is comprised of each autocorrelative calling behavior.

As shown in Figure 1, the generation method that the user that the present invention is applicable to rubbish voice filtering calls out behavior model comprises the following steps:

(1) set up and call out interbehavior feature CI, for describing designated user as calling subscriber and called subscriber's corelation behaviour feature, wherein call out interbehavior feature CI and further comprise again incoming call exhalation ratio, call out intersection record characteristic value and mutual intensity and three parts that distribute thereof, specifically can be expressed as:

CI＝{R _in/out，C _out，C _in，C _in/out，F _in/out}

Wherein, R _in/outfor the ratio of user as calling subscriber and called subscriber, C _outthat user only has the quantity of exhalation behavior, C in the calling of all and different user as calling subscriber _inthat user only has the quantity of the behavior of answering, C in the calling of all and different user as called subscriber _in/outuser in the calling of all and different user simultaneously as calling subscriber and called subscriber's quantity, F _in/outit is the mutual intensity frequency distribution with other users;

(2) set up calling frequency and distribution F thereof _cD, for describing the call record feature of call time, comprising the interior calling frequency value of timing statistics section and calling frequency value time distribution two parts of appointment, specifically can be expressed as:

$F_{\mathrm{CD}}=\{F_{\mathrm{in}/\mathrm{out}}^T,{\stackrel{\‾}{D}}_{\mathrm{out}}^T\}$

Wherein,

for the absolute frequency value that in timing statistics section, user makes a call,

be user's calling frequency in timing statistics section be the distribution in the timeslice of 2 hours one day 12 length.

(3) set up call duration and distribution D thereof _cT, for describing the feature of call record duration, comprising histogram distribution three parts that in timing statistics section, call denial is answered ratio, called out average duration and call duration, specifically can be expressed as:

$D_{\mathrm{CT}}=\{f_E^T,{\mathrm{CT}}_{\mathrm{avg}}^T,{\mathrm{CTD}}^T\}$

Wherein, for user is rejected the probable value of answering in making a call,

for user's mean call duration, CTD ^tdistribution for call duration.

As shown in Figure 2, the step of the inventive method (1) comprises the following steps:

201, all historical call record of inquiry designated user, adds up total exhalation number of times C _out, total incoming call number of times C _inwith total incoming call exhalation number of times C _in/out;

202, the C that step 201 is counted _out, C _inand C _in/outbe normalized.Historical interaction feature value in call record after normalization is designated as:

with

computing formula is as follows:

${\stackrel{\‾}{C}}_{\mathrm{out}}=\frac{C_{\mathrm{out}}}{C_{\mathrm{out}}+C_{\mathrm{in}}+C_{\mathrm{in}/\mathrm{out}}}$

${\stackrel{\‾}{C}}_{\mathrm{in}}=\frac{C_{\mathrm{in}}}{C_{\mathrm{out}}+C_{\mathrm{in}}+C_{\mathrm{in}/\mathrm{out}}}$

${\stackrel{\‾}{C}}_{\mathrm{in}/\mathrm{out}}=\frac{C_{\mathrm{in}/\mathrm{out}}}{C_{\mathrm{out}}+C_{\mathrm{in}}+C_{\mathrm{in}/\mathrm{out}}}$

Wherein, with

three's sum is 1;

203, the incoming call exhalation ratio R of statistics designated user _in/out, computing formula is as follows:

$R_{\mathrm{in}/\mathrm{out}}=\frac{C_{\mathrm{out}}}{C_{\mathrm{in}}}$

204, exhalation, incoming call number of times between statistics designated user and other users, be designated as C _outand C _in. the number of times that represents this user's active call user j,

represent that this user answers the number of calls from user j.Only preserve this user and other statisticses that exists exhalation, incoming call to record.

205, according to the result of step 204, further add up designated user and other users' the mutual intensity CD of calling _in/out.

represent the mutual intensity of calling between this user and user j, its computing formula is as follows:

${\mathrm{CD}}_{\mathrm{in}/\mathrm{out}}^j=\mathrm{INT}\left[\log (C_{\mathrm{out}}^j+C_{\mathrm{in}}^j)\right]$

Wherein, INT[] expression bracket function, the number of times that represents this user's active call user j,

represent that this user answers the number of calls from user j.

206, the mutual intensity distributions CDD of calling of statistics designated user _in/out,

CDD _in/out＝{CDN ₀，CDN ₁，CDN ₂，CDN ₃}

Wherein, CDN _ithe computing formula of (i=0,1,2,3) is as follows:

Be CDN _ifor value equals i or is more than or equal to the CD of i _in/outquantity, COUNT[wherein] be counting function,

represent the mutual intensity of calling between this user and user j, n represents this user All Contacts's quantity.

207, to CDD _in/outbe normalized, the mutual intensity distributions of calling after normalization is designated as specifically be expressed as:

${\stackrel{\‾}{\mathrm{CDN}}}_{\mathrm{in}/\mathrm{out}}=\{{\stackrel{\‾}{\mathrm{CDN}}}_0,{\stackrel{\‾}{\mathrm{CDN}}}_1,{\stackrel{\‾}{\mathrm{CDN}}}_2,{\stackrel{\‾}{\mathrm{CDN}}}_3\}$

Wherein,

for the CDN after normalization _i(i=0,1,2,3) computing formula is:

${\stackrel{\‾}{\mathrm{CDN}}}_i=\frac{{\mathrm{CDN}}_i}{{\mathrm{\Σ}}_{k+0}^a{\mathrm{CDN}}_k}.$

208, the calling interaction feature parameters C I of output designated user.

As shown in Figure 3, the step of the inventive method (2) comprises the following steps:

301, all historical call record in the fixed time section of inquiry designated user, comprises incoming call and exhalation;

302, the absolute frequency value that statistics designated user makes a call

$F_{\mathrm{in}/\mathrm{out}}^T=\frac{C_{\mathrm{out}}^T}{T}$

Wherein, the timing statistics segment length that T is appointment, unit is hour, and T is the integral multiple of 24 hours;

for the number of calls of user in the time period as calling subscriber;

303, the calling frequency distributed constant of the designated user in statistics fixed time section T specifically be expressed as:

$D_{\mathrm{out}}^T=\{D_{\mathrm{out}}^1,D_{\mathrm{out}}^2,...,D_{\mathrm{out}}^{12},\}$

Wherein,

[2* (t-1), 2*t) the call number sum in period that represent every day in timing statistics section T.Time every day can be divided into 12 timeslices, and each timeslice includes 2 hours, and 12 timeslices are: [0,2), [and 2,4) ..., [22,24);

304, to user's calling frequency distributed constant be normalized, processing procedure is as follows:

${\stackrel{\‾}{D}}_{\mathrm{out}}^t=\frac{D_{\mathrm{out}}^t}{C_{\mathrm{out}}^T},(t=1,...,12)$

The calling frequency after normalization is distributed as

${\stackrel{\‾}{D}}_{\mathrm{out}}^T=\{{\stackrel{\‾}{D}}_{\mathrm{out}}^1,{\stackrel{\‾}{D}}_{\mathrm{out}}^2,...,{\stackrel{\‾}{D}}_{\mathrm{out}}^{12},\};$

305, calling frequency and the distributed constant F thereof of designated user in the output designated statistics time _cD.

As shown in Figure 4, the step of the inventive method (3) comprises the following steps:

401, all historical call record in the fixed time section T of inquiry designated user, comprises incoming call and exhalation;

402, during making a call, counting user is rejected the ratio of answering

computing formula is as follows:

$f_E^T=\frac{{\mathrm{CR}}_{\mathrm{out}}^T}{C_{\mathrm{out}}^T}$

Wherein,

represent the all-calls number of times that in the statistical piece time, user sends,

represent unaccepted active call number of times in the statistical piece time;

403, the mean call duration of counting user

computing formula is as follows:

${\mathrm{CT}}_{\mathrm{avg}}^T=\frac{{\mathrm{\Σ}}_{r=1}^{C_{\mathrm{in}/\mathrm{out}}^T}{t}_r}{C_{\mathrm{in}/\mathrm{out}}^T}$

Wherein,

represent the number of calls in fixed time section T, t _rbe call duration corresponding to r bar call record, unit is second;

404, the histogram distribution CTD of statistics call duration ^t, be specifically expressed as follows:

${\mathrm{CTD}}^T=\{{\mathrm{CT}}_D^0,{\mathrm{CT}}_D^1,...,{\mathrm{CT}}_D^6,{\mathrm{CT}}_D^7\}$

Wherein,

represent the call record ratio of call duration within certain time period in fixed time section T, specific formula for calculation is as follows:

405, call duration and the distributed constant D thereof of designated user in the output designated statistics time _cT.

Those skilled in the art will readily understand; the foregoing is only preferred embodiment of the present invention; not in order to limit the present invention, all any modifications of doing within the spirit and principles in the present invention, be equal to and replace and improvement etc., within all should being included in protection scope of the present invention.

Claims (4)

1. the user who is applicable to rubbish voice filtering calls out a generation method for behavior model, it is characterized in that, comprises the following steps:

(1) set up and call out interbehavior feature CI, for describing designated user as calling subscriber and called subscriber's corelation behaviour feature, wherein call out interbehavior feature CI and further comprise again incoming call exhalation ratio, call out intersection record characteristic value and mutual intensity and three parts that distribute thereof, specifically can be expressed as:

CI＝{R _in/out，C _out，C _in，C _in/out，F _in/out}

Wherein, R _in/outfor the ratio of user as calling subscriber and called subscriber, C _outthat user only has the quantity of exhalation behavior, C in the calling of all and different user as calling subscriber _inthat user only has the quantity of the behavior of answering, C in the calling of all and different user as called subscriber _in/outuser in the calling of all and different user simultaneously as calling subscriber and called subscriber's quantity, F _in/outit is the mutual intensity frequency distribution with other users;

(2) set up calling frequency and distribution F thereof _cD, for describing the call record feature of call time, comprising the interior calling frequency value of timing statistics section and calling frequency value time distribution two parts of appointment, specifically can be expressed as:

$F_{\mathrm{CD}}=\{F_{\mathrm{in}/\mathrm{out}}^T,{\stackrel{\‾}{D}}_{\mathrm{out}}^T\}$

Wherein,

for the absolute frequency value that in timing statistics section, user makes a call,

be user's calling frequency in timing statistics section be the distribution in the timeslice of 2 hours one day 12 length.

(3) set up call duration and distribution D thereof _cT, for describing the feature of call record duration, comprising histogram distribution three parts that in timing statistics section, call denial is answered ratio, called out average duration and call duration, specifically can be expressed as:

$D_{\mathrm{CT}}=\{f_E^T,{\mathrm{CT}}_{\mathrm{avg}}^T,{\mathrm{CTD}}^T\}$

Wherein,

for user is rejected the probable value of answering in making a call,

for user's mean call duration, CTD ^tdistribution for call duration.

2. generation method according to claim 1, is characterized in that, step (1) specifically comprises following sub-step:

(1-1) history of counting user is called out interbehavior characteristic parameter C _out, C _inand C _in/out, for the ease of across comparison between user, further it is normalized:

${\stackrel{\‾}{C}}_{\mathrm{out}}=\frac{C_{\mathrm{out}}}{C_{\mathrm{out}}+C_{\mathrm{in}}+C_{\mathrm{in}/\mathrm{out}}}$

${\stackrel{\‾}{C}}_{\mathrm{in}}=\frac{C_{\mathrm{in}}}{C_{\mathrm{out}}+C_{\mathrm{in}}+C_{\mathrm{in}/\mathrm{out}}}$

${\stackrel{\‾}{C}}_{\mathrm{in}/\mathrm{out}}=\frac{C_{\mathrm{in}/\mathrm{out}}}{C_{\mathrm{out}}+C_{\mathrm{in}}+C_{\mathrm{in}/\mathrm{out}}}$

Wherein, with

three's sum is 1;

(1-2) statistics incoming call exhalation ratio R _in/out, computing formula is as follows:

$R_{\mathrm{in}/\mathrm{out}}=\frac{C_{\mathrm{out}}}{C_{\mathrm{in}}}$

(1-3) statistics designated user and other users' the mutual intensity CD of calling _in/out.The mutual intensity of calling between this user and user j is expressed as

its computing formula is as follows:

${\mathrm{CD}}_{\mathrm{in}/\mathrm{out}}^j=\mathrm{INT}\left[\log (C_{\mathrm{out}}^j+C_{\mathrm{in}}^j)\right]$

Wherein, INT[] expression bracket function,

the number of times that represents this user's active call user j,

represent that this user answers the number of calls from user j;

(1-4) statistics is called out mutual intensity distributions CDD _in/out

CDD _in/out＝{CDN ₀，CDN ₁，CDN ₂，CDN ₃}

Wherein, CDN _ithe computing formula of (i=0,1,2,3) is as follows:

Be CDN _ifor value equals i or is more than or equal to the DC of i _in/outquantity, COUNT[wherein] be counting function,

represent the mutual intensity of calling between this user and user j, n represents this user All Contacts's quantity;

(1-5) to CDD _in/ _outbe normalized, the mutual intensity distributions of calling after normalization is designated as specifically be expressed as:

${\stackrel{\‾}{\mathrm{CDN}}}_{\mathrm{in}/\mathrm{out}}=\{{\stackrel{\‾}{\mathrm{CDN}}}_0,{\stackrel{\‾}{\mathrm{CDN}}}_1,{\stackrel{\‾}{\mathrm{CDN}}}_2,{\stackrel{\‾}{\mathrm{CDN}}}_3\}$

Wherein,

for the CDN after normalization _i(i=0,1,2,3), computing formula is:

${\stackrel{\‾}{\mathrm{CDN}}}_i=\frac{{\mathrm{CDN}}_i}{{\mathrm{\Σ}}_{k+0}^a{\mathrm{CDN}}_k}.$

3. generation method according to claim 1, is characterized in that, step (2) specifically comprises following sub-step:

(2-1) absolute frequency value that counting user makes a call

$F_{\mathrm{in}/\mathrm{out}}^T=\frac{C_{\mathrm{out}}^T}{T}$

Wherein, the timing statistics segment length that T is appointment, unit is hour that this time period should start from 0 o'clock of first day, end at 24 o'clock of last day, so T should be the integral multiple of 24 hours;

for the number of calls of user in the time period as calling subscriber;

(2-2) the calling frequency distributed constant of the designated user in statistics fixed time section T

specifically be expressed as:

$D_{\mathrm{out}}^T=\{D_{\mathrm{out}}^1,D_{\mathrm{out}}^2,...,D_{\mathrm{out}}^{12},\}$

Wherein, [2* (t-1), 2*t) the call number sum in period that represent every day in timing statistics section T.Time every day can be divided into 12 timeslices, and each timeslice includes 2 hours, and 12 timeslices are specially: [0,2), [2,4), [4,6), [6,8), [8,10), [10,12), [12,14), [14,16), [16,18), [18,20), [20,22), [22,24).

(2-3) to user's calling frequency distributed constant

be normalized, processing procedure is as follows:

${\stackrel{\‾}{D}}_{\mathrm{out}}^t=\frac{D_{\mathrm{out}}^t}{C_{\mathrm{out}}^T},(t=1,...,12);$

The calling frequency after normalization is distributed as:

${\stackrel{\‾}{D}}_{\mathrm{out}}^T=\{{\stackrel{\‾}{D}}_{\mathrm{out}}^1,{\stackrel{\‾}{D}}_{\mathrm{out}}^2,...,{\stackrel{\‾}{D}}_{\mathrm{out}}^{12},\};$

(2-4) obtain finally having calling frequency and the distributed constant F thereof of service efficiency _cD

$F_{\mathrm{CD}}=\{F_{\mathrm{in}/\mathrm{out}}^T,{\stackrel{\‾}{D}}_{\mathrm{out}}^T\}.$

4. generation method according to claim 1, is characterized in that, step (3) specifically comprises following sub-step:

(3-1) counting user is rejected the ratio of answering in making a call

computing formula is as follows:

$f_E^T=\frac{{\mathrm{CR}}_{\mathrm{out}}^T}{C_{\mathrm{out}}^T}$

Wherein, represent the all-calls number of times that in timing statistics section, user sends,

represent unaccepted active call number of times in the statistical piece time;

(3-2) mean call duration of counting user

computing formula is as follows:

${\mathrm{CT}}_{\mathrm{avg}}^T=\frac{{\mathrm{\Σ}}_{r=1}^{C_{\mathrm{in}/\mathrm{out}}^T}{t}_r}{C_{\mathrm{in}/\mathrm{out}}^T}$

Wherein,

represent the number of calls of statistical piece in the time, t _rbe call duration corresponding to r bar call record;

(3-3) the histogram distribution CTD of statistics call duration ^t, be specifically expressed as follows:

${\mathrm{CTD}}^T=\{{\mathrm{CT}}_D^0,{\mathrm{CT}}_D^1,...,{\mathrm{CT}}_D^6,{\mathrm{CT}}_D^7\}$

Wherein,

represent to call out the call record ratio of lasting duration within certain time period, specific formula for calculation is as follows:

Images