The content of the invention
It is a primary object of the present invention to provide the recognition methods that a kind of life insurance with clique's risk of fraud settles a claim caseAnd device, it is intended to effectively identify the life insurance Claims Resolution case with clique's risk of fraud.
To achieve the above object, the identification side for a kind of life insurance Claims Resolution case with clique's risk of fraud that the present invention is providedMethod, the described method comprises the following steps：
If receiving the life insurance Claims Resolution case of risk to be analyzed, whether analyze in life insurance Claims Resolution case with one or moreKey risk feature in predetermined key feature combination；The key feature is combined as settling a claim from each default life insuranceThe combination of the key risk feature extracted in clique's fraud case according to preset rules；
If having in the life insurance Claims Resolution case of the risk to be analyzed in one or more predetermined key feature combinationsKey risk feature, it is determined that the risk to be analyzed life insurance Claims Resolution case be the life insurance Claims Resolution case with clique's risk of fraudPart.
Preferably, the key feature combination is determined as follows：
The sample data of the life insurance Claims Resolution case for belonging to clique's fraud of predetermined number is obtained, by same life insurance Claims Resolution groupThe sample data of life insurance Claims Resolution case under partner's fraud case is classified as same sample data sets, by different life insurance Claims Resolution groupsThe sample data of life insurance Claims Resolution case under partner's fraud case is classified as different sample data sets；
Each self-corresponding common characteristic set of each sample data sets is determined from each sample data sets；
The combinations of features being associated in all common characteristic set is determined according to predetermined correlation rule, and shouldCombinations of features is combined as key feature.
Preferably, it is described that the feature being associated in all common characteristic set is determined according to predetermined correlation ruleThe step of combining, and this feature combination is combined as key feature includes：
Feature in all common characteristic set is combined by default combination, multiple combinations of features are obtained, andCorresponding coupling index value is determined according to predetermined correlation rule to each combinations of features；
If thering is the corresponding coupling index value of combinations of features to be more than corresponding predetermined threshold value, it is determined that this feature is combined as keyCombinations of features.
Preferably, the coupling index value include support and degree of belief, it is described to each combinations of features according in advance reallyThe step of fixed correlation rule determines corresponding coupling index value includes：
For a combinations of features comprising fisrt feature and second feature, calculate have simultaneously the fisrt feature andThe life insurance Claims Resolution case sample data of the second feature accounts for the percentage of the life insurance Claims Resolution case sample data of predetermined number, withThe support combined as this feature；
Calculate while there is the life insurance Claims Resolution case sample data of the fisrt feature and the second feature, which to account for, hasThe percentage of the life insurance Claims Resolution case sample data of the fisrt feature, using the degree of belief combined as this feature.
Preferably, the coupling index value includes support and degree of belief, and the predetermined threshold value includes default support thresholdValue and default degree of belief threshold value, if described have the corresponding coupling index value of combinations of features to be more than predetermined threshold value, it is determined that this featureBeing combined as the step of key feature is combined includes：
If there is the corresponding support of combinations of features to be more than default support threshold, and corresponding degree of belief is more than default trustSpend threshold value, it is determined that this feature is combined as key feature combination.
In addition, to achieve the above object, the present invention also provides a kind of life insurance Claims Resolution case with clique's risk of fraudIdentifying device, the identifying device includes：
Analysis module, if for receive risk to be analyzed life insurance settle a claim case, analyze the life insurance Claims Resolution case in whetherWith the key risk feature in one or more predetermined key feature combinations；The key feature is combined as from defaultEach life insurance Claims Resolution clique fraud case in the combination of key risk feature that is extracted according to preset rules；
Risk determining module, if for there is one or more predefine in the life insurance Claims Resolution case of the risk to be analyzedKey feature combination in key risk feature, it is determined that the risk to be analyzed life insurance Claims Resolution case be with clique fraudThe life insurance Claims Resolution case of risk.
Preferably, the identifying device also includes：
Acquisition module, the sample data of the life insurance Claims Resolution case for belonging to clique's fraud for obtaining predetermined number, will be sameThe sample data of life insurance Claims Resolution case under one life insurance Claims Resolution clique fraud case is classified as same sample data sets, will notThe sample data of life insurance Claims Resolution case under same life insurance Claims Resolution clique fraud case is classified as different sample data sets；
Common characteristic determining module, for determining that each sample data sets is each right from each sample data setsThe common characteristic set answered；
Determining module is associated, is associated for being determined according to predetermined correlation rule in all common characteristic setCombinations of features, and using this feature combination combined as key feature.
Preferably, the association determining module is additionally operable to：
Feature in all common characteristic set is combined by default combination, multiple combinations of features are obtained, andCorresponding coupling index value is determined according to predetermined correlation rule to each combinations of features；If there is combinations of features correspondingCoupling index value is more than corresponding predetermined threshold value, it is determined that this feature is combined as key feature combination.
Preferably, the coupling index value includes support and degree of belief, and the association determining module is additionally operable to：
For a combinations of features comprising fisrt feature and second feature, calculate have simultaneously the fisrt feature andThe life insurance Claims Resolution case sample data of the second feature accounts for the percentage of the life insurance Claims Resolution case sample data of predetermined number, withThe support combined as this feature；Calculate the life insurance Claims Resolution case with the fisrt feature and the second featureSample data accounts for the percentage of the life insurance Claims Resolution case sample data with the fisrt feature, using the letter combined as this featureRen Du.
Preferably, the coupling index value includes support and degree of belief, and the predetermined threshold value includes default support thresholdValue and default degree of belief threshold value, the association determining module are additionally operable to：
If there is the corresponding support of combinations of features to be more than default support threshold, and corresponding degree of belief is more than default trustSpend threshold value, it is determined that this feature is combined as key feature combination.
The recognition methods of life insurance Claims Resolution case proposed by the present invention with clique's risk of fraud and device, by analyzing the longevityWhether there is the key risk feature in one or more predetermined key feature combinations, to determine this in danger Claims Resolution caseWhether the life insurance Claims Resolution case of risk to be analyzed is the life insurance Claims Resolution case with clique's risk of fraud.Due to the key featureIt is combined as the group of key risk feature extracted from each default life insurance Claims Resolution clique fraud case according to preset rulesClose, if the life insurance Claims Resolution case of the risk to be analyzed has the key risk feature in key feature combination, illustrate that this is treatedThe life insurance Claims Resolution case of analysis risk is possible for life insurance Claims Resolution clique fraud case, you can the automatic identification risk to be analyzedLife insurance Claims Resolution case is the life insurance Claims Resolution case with clique's risk of fraud, simpler compared to the mode of manual identified, quick,Accurately.
Embodiment
In order that technical problems, technical solutions and advantages to be solved are clearer, clear, tie belowDrawings and examples are closed, the present invention will be described in further detail.It should be appreciated that specific embodiment described herein is onlyTo explain the present invention, it is not intended to limit the present invention.
The present invention provides the recognition methods that a kind of life insurance with clique's risk of fraud settles a claim case.
Reference picture 1, Fig. 1 has the embodiment of recognition methods one of the life insurance Claims Resolution case of clique's risk of fraud for the present inventionSchematic flow sheet.
In one embodiment, the recognition methods of life insurance Claims Resolution case that should be with clique's risk of fraud includes：
Step S10, whether if receiving the life insurance Claims Resolution case of risk to be analyzed, analyzing in life insurance Claims Resolution case has oneKey risk feature in individual or multiple predetermined key feature combinations；The key feature be combined as from it is default eachThe combination of the key risk feature extracted in life insurance Claims Resolution clique fraud case according to preset rules.
In the present embodiment, the risk analysis for receiving the case of being settled a claim to life insurance that user sends is asked, and is existed for example, receiving userRelevant information is inputted in the terminals such as mobile phone, tablet personal computer, self-help terminal equipment (for example, applicant's identity of life insurance Claims Resolution casePersonal relations information, the amount for which loss settled etc. of information, life insurance business person's information, participant of settling a claim) send afterwards to life insurance Claims Resolution caseThe risk analysis request of part, can such as receive user's preassembled longevity in mobile phone, tablet personal computer, self-help terminal equipment terminalThe risk analysis request sent in danger Claims Resolution case risk analysis application program after input relevant information, or user is received in handThe risk analysis sent after relevant information is inputted on browser in the terminals such as machine, tablet personal computer, self-help terminal equipmentRequest.
Whether after the risk analysis request for case of being settled a claim to life insurance is received, analyzing in life insurance Claims Resolution case has oneKey risk feature in individual or multiple predetermined key feature combinations, such as analyzes the correlated characteristic of life insurance Claims Resolution caseWhether such as applicant's identity information, life insurance business person's information, settle a claim the personal relations information of participant, amount for which loss settled are advanceKey risk feature in the key feature combination of determination.Wherein, the key feature is combined as managing from each default life insuranceThe combination of the key risk feature extracted in clique's fraud case according to preset rules is paid for, for example, being taken advantage of in life insurance Claims Resolution cliqueIn fraud case part, between doctor and client, between doctor and life insurance business person, between client and life insurance business person, client and clientBetween be likely to exist and constitute the risk that case is cheated by life insurance Claims Resolution clique, multiple life insurances Claims Resolutions are obtained from historical dataCase sample is cheated by clique, and the feature in clique's fraud case sample of being settled a claim to multiple life insurances is analyzed, and can such as calculate manyThe occurrence number for the feature that is mutually related in individual life insurance Claims Resolution clique fraud case sample, can also calculate multiple life insurance Claims Resolution cliquesBe mutually related the frequency of occurrences of feature in preset time period, etc. in fraud case sample, does not limit herein.Can will be manyOccurrence number exceedes the feature that is mutually related of certain number of times as key feature in individual life insurance Claims Resolution clique fraud case sampleCombination, or the frequency of occurrences in multiple life insurances Claims Resolution clique fraud case sample in preset time period is exceeded into certain frequencyThe feature that is mutually related is combined as key feature.If for example, multiple life insurances Claims Resolution clique fraud case sample in Y1 hospitals withThe number of times that Y2 life insurance business persons occur together exceedes preset times (such as 3 times), then can be by Y1 hospitals and the group of Y2 life insurance business personsCooperate to combine for key feature.After the risk analysis request for case of being settled a claim to life insurance is received, you can analyze life insurance Claims ResolutionKey risk feature such as Y1 hospitals, Y2 life insurance business persons in whether being combined in case comprising key feature.
Step S20, if having in the life insurance Claims Resolution case of the risk to be analyzed one or more predetermined crucial specialLevy the key risk feature in combination, it is determined that the life insurance Claims Resolution case of the risk to be analyzed is the longevity with clique's risk of fraudDanger Claims Resolution case.
If there are one or more predetermined key features in the life insurance Claims Resolution case for analyzing the risk to be analyzedKey risk feature in combination, such as comprising Y1 hospitals and/or Y2 life insurance business persons, then illustrates the life insurance reason of the risk to be analyzedThe feature for including high clique's risk of fraud in case is paid for, then the life insurance Claims Resolution case for automatically determining and recognizing risk to be analyzed is heightThe life insurance Claims Resolution case of clique's risk of fraud.
The present embodiment is by analyzing in life insurance Claims Resolution case whether have one or more predetermined key feature groupsKey risk feature in conjunction, to determine whether the life insurance Claims Resolution case of the risk to be analyzed is the longevity with clique's risk of fraudDanger Claims Resolution case.Because the key feature is combined as from each default life insurance Claims Resolution clique fraud case according to default ruleThe combination of the key risk feature then extracted, if the life insurance Claims Resolution case of the risk to be analyzed has in key feature combinationKey risk feature, then illustrate the risk to be analyzed life insurance Claims Resolution case be possible for life insurance Claims Resolution clique fraud case,Can the life insurance Claims Resolution case of the automatic identification risk to be analyzed be the life insurance Claims Resolution case with clique's risk of fraud, compared to peopleIt is simpler otherwise, quick, accurate that work is known.
Further, in other embodiments, as shown in Fig. 2 key feature combination is determined in accordance with the following steps：
Step S30, obtains the sample data of the life insurance Claims Resolution case for belonging to clique's fraud of predetermined number, by the same longevityThe sample data of life insurance Claims Resolution case under danger Claims Resolution clique fraud case is classified as same sample data sets, by the different longevityThe sample data of life insurance Claims Resolution case under danger Claims Resolution clique fraud case is classified as different sample data sets, certain to obtainThe sample data sets of case are cheated by the life insurance Claims Resolution clique of quantity.If for example, associated by life insurance Claims Resolution clique fraud case XLife insurance Claims Resolution case includes tetra- life insurance Claims Resolution cases of X1, X2, X3 and X4, then by the corresponding life insurance Claims Resolution number of X1, X2, X3 and X4The sample data sets corresponding according to being classified as same and settling a claim with the life insurance clique fraud case X.
Step S40, determines each self-corresponding common characteristic of each sample data sets from each sample data setsSet, if for example, having the sample data sets of the different life insurance Claims Resolution clique fraud cases of A, B two, from sample data sets AIn determine the corresponding common characteristic set a of A, the corresponding common characteristic set b of B are determined from sample data sets B.ItsIn, the common characteristic is that all life insurances Claims Resolution case belonged under same life insurance Claims Resolution clique fraud case is jointly comprisedFeature, for example, the common characteristic can be Y1 hospitals, Y2 life insurance business persons, Y3 work units, Y4 personal relations, Y5 agesInterval and/or Y6 native places etc..
Step S50, the feature group being associated in all common characteristic set is determined according to predetermined correlation ruleClose, and this feature combination is combined as key feature.For example, will can jointly occur in the feature in all common characteristic setNumber of times does many combinations of features and combined as key feature.
In a kind of optional embodiment, above-mentioned steps S50 can include：
Feature in all common characteristic set is combined by default combination, the default combination include butCombination of two, multi-to-multi combination etc. are not limited to, multiple combinations of features are obtained, and each combinations of features is closed according to predeterminedConnection rule determines corresponding coupling index value；If there is the corresponding coupling index value of combinations of features to be more than corresponding predetermined threshold value,Then determine that this feature is combined as key feature combination.
Wherein, correlation rule is shape such as X → Y implications, wherein, X and Y are referred to as the guide of correlation ruleIt is (antecedent or left-hand-side, LHS) and follow-up (consequent or right-hand-side, RHS), at thisIn embodiment, X and Y can be each features in combinations of features, and the coupling index value includes support (support) and letterAppoint degree (confidence).For example, for a combinations of features for including fisrt feature X and second feature Y, calculating has simultaneouslyThe fisrt feature X and the second feature Y life insurance Claims Resolution case sample data account for the life insurance Claims Resolution case of predetermined numberThe percentage of sample data, using the support combined as this feature, i.e. life insurance Claims Resolution case of this feature combination in predetermined numberSupport in part sample data be predetermined number life insurance Claims Resolution case sample data in while comprising fisrt feature X, secondThe percentage of characteristic Y, i.e. probability.Calculate the life insurance Claims Resolution case with the fisrt feature X and the second feature YSample data accounts for the percentage of the life insurance Claims Resolution case sample data with the fisrt feature X, using what is combined as this featureThe degree of belief that degree of belief, i.e. this feature combine in the life insurance Claims Resolution case sample data of predetermined number is the life insurance of predetermined numberIn case sample data of settling a claim in the case of included fisrt feature X, second feature Y percentage, i.e. condition are included generalRate.Can be according to default support threshold and default degree of belief threshold value be preset the need for practical application, if there is combinations of features pairThe support answered is more than default support threshold, and corresponding degree of belief is more than default degree of belief threshold value, it is determined that this feature groupIt is combined into key feature combination；If there is the corresponding support of combinations of features to be less than or equal to default support threshold, and/or, it is rightThe degree of belief answered is less than or equal to default degree of belief threshold value, it is determined that this feature combination is not key feature combination, so as to controlMake whether case of being settled a claim to the life insurance of the risk to be analyzed is that the life insurance with clique's risk of fraud is settled a claim the analysis result of case.
If for example, all common characteristic set I={ close by Y1 hospitals, Y2 life insurance business persons, Y3 work units, Y4 individualsSystem }, there are 5 life insurance Claims Resolution case sample datas to be cured comprising Y1 in predetermined number is the life insurance Claims Resolution case sample data of 6, there are 3 life insurance Claims Resolution case sample datas in institute while comprising Y1 hospitals and Y2 life insurance business persons.Then for including Y1 hospitals and Y2For the combinations of features of life insurance business person, the support of this feature combination is while has the longevity of Y1 hospitals and Y2 life insurance business personsDanger Claims Resolution case sample data accounts for the percentage of the life insurance Claims Resolution case sample data of predetermined number, i.e. 3/6=0.5；This featureThe confidence level of combination is while has Y1 hospitals and the life insurance Claims Resolution case sample data of Y2 life insurance business persons to account for Y1 hospitalsLife insurance settle a claim case sample data percentage, i.e. 3/5=0.6.If presetting support threshold α=0.4, degree of belief thresholdValue β=0.5, then judge because Y1 hospitals support 0.5 corresponding with the combinations of features of Y2 life insurance business persons is more than default supportThreshold value 0.4 is spent, and corresponding degree of belief 0.6 is more than default degree of belief threshold value 0.5, accordingly, it can be determined that Y1 hospitals and Y2 life insurancesThe combinations of features of business person combines for key feature.
In the present embodiment, due to generally being entered by for example multiple participants of multiple features in life insurance Claims Resolution clique fraud caseRow clique is cheated, therefore, and being settled a claim by correlation rule from multiple life insurances, it is associated to be found during the common characteristic of case is cheated by cliqueKey feature combination, it is ensured that the key feature in the key feature combination found is used to carry out clique's fraud to be usualFeature.Risk analysis is carried out to life insurance Claims Resolution case using the key feature in the key feature combination found, can judgedWhen there is the key risk feature in the combination of one or more key features in the life insurance Claims Resolution case of the risk to be analyzed, automaticallyIt is the life insurance Claims Resolution case with clique's risk of fraud to recognize the life insurance Claims Resolution case of the risk to be analyzed.Compared to manual identifiedMode, the present embodiment can recognize that the internal association of different life insurances Claims Resolution cases so that the clique for case of being settled a claim to life insuranceRisk of fraud analysis more accurate and effective.
A kind of identifying device for case of being settled a claim the present invention further provides life insurance with clique's risk of fraud.
Reference picture 3, Fig. 3 has the identifying device first embodiment of the life insurance Claims Resolution case of clique's risk of fraud for the present inventionHigh-level schematic functional block diagram.
In the first embodiment, the identifying device of life insurance Claims Resolution case that should be with clique's risk of fraud includes：
Analysis module 01, if the life insurance Claims Resolution case for receiving risk to be analyzed, analyzing in life insurance Claims Resolution case isThe no key risk feature with one or more predetermined key feature combinations；The key feature is combined as from pre-If each life insurance Claims Resolution clique fraud case in the combination of key risk feature that is extracted according to preset rules；
In the present embodiment, the risk analysis for receiving the case of being settled a claim to life insurance that user sends is asked, and is existed for example, receiving userRelevant information is inputted in the terminals such as mobile phone, tablet personal computer, self-help terminal equipment (for example, applicant's identity of life insurance Claims Resolution casePersonal relations information, the amount for which loss settled etc. of information, life insurance business person's information, participant of settling a claim) send afterwards to life insurance Claims Resolution caseThe risk analysis request of part, can such as receive user's preassembled longevity in mobile phone, tablet personal computer, self-help terminal equipment terminalThe risk analysis request sent in danger Claims Resolution case risk analysis application program after input relevant information, or user is received in handThe risk analysis sent after relevant information is inputted on browser in the terminals such as machine, tablet personal computer, self-help terminal equipmentRequest.
Whether after the risk analysis request for case of being settled a claim to life insurance is received, analyzing in life insurance Claims Resolution case has oneKey risk feature in individual or multiple predetermined key feature combinations, such as analyzes the correlated characteristic of life insurance Claims Resolution caseWhether such as applicant's identity information, life insurance business person's information, settle a claim the personal relations information of participant, amount for which loss settled are advanceKey risk feature in the key feature combination of determination.Wherein, the key feature is combined as managing from each default life insuranceThe combination of the key risk feature extracted in clique's fraud case according to preset rules is paid for, for example, being taken advantage of in life insurance Claims Resolution cliqueIn fraud case part, between doctor and client, between doctor and life insurance business person, between client and life insurance business person, client and clientBetween be likely to exist and constitute the risk that case is cheated by life insurance Claims Resolution clique, multiple life insurances Claims Resolutions are obtained from historical dataCase sample is cheated by clique, and the feature in clique's fraud case sample of being settled a claim to multiple life insurances is analyzed, and can such as calculate manyThe occurrence number for the feature that is mutually related in individual life insurance Claims Resolution clique fraud case sample, can also calculate multiple life insurance Claims Resolution cliquesBe mutually related the frequency of occurrences of feature in preset time period, etc. in fraud case sample, does not limit herein.Can will be manyOccurrence number exceedes the feature that is mutually related of certain number of times as key feature in individual life insurance Claims Resolution clique fraud case sampleCombination, or the frequency of occurrences in multiple life insurances Claims Resolution clique fraud case sample in preset time period is exceeded into certain frequencyThe feature that is mutually related is combined as key feature.If for example, multiple life insurances Claims Resolution clique fraud case sample in Y1 hospitals withThe number of times that Y2 life insurance business persons occur together exceedes preset times (such as 3 times), then can be by Y1 hospitals and the group of Y2 life insurance business personsCooperate to combine for key feature.After the risk analysis request for case of being settled a claim to life insurance is received, you can analyze life insurance Claims ResolutionKey risk feature such as Y1 hospitals, Y2 life insurance business persons in whether being combined in case comprising key feature.
Risk determining module 02, if one or more true in advance for having in the life insurance Claims Resolution case of the risk to be analyzedKey risk feature in fixed key feature combination, it is determined that the life insurance Claims Resolution case of the risk to be analyzed is to be taken advantage of with cliqueCheat the life insurance Claims Resolution case of risk.
If there are one or more predetermined key features in the life insurance Claims Resolution case for analyzing the risk to be analyzedKey risk feature in combination, such as comprising Y1 hospitals and/or Y2 life insurance business persons, then illustrates the life insurance reason of the risk to be analyzedThe feature for including high clique's risk of fraud in case is paid for, then the life insurance Claims Resolution case for automatically determining and recognizing risk to be analyzed is heightThe life insurance Claims Resolution case of clique's risk of fraud.
The present embodiment is by analyzing in life insurance Claims Resolution case whether have one or more predetermined key feature groupsKey risk feature in conjunction, to determine whether the life insurance Claims Resolution case of the risk to be analyzed is the longevity with clique's risk of fraudDanger Claims Resolution case.Because the key feature is combined as from each default life insurance Claims Resolution clique fraud case according to default ruleThe combination of the key risk feature then extracted, if the life insurance Claims Resolution case of the risk to be analyzed has in key feature combinationKey risk feature, then illustrate the risk to be analyzed life insurance Claims Resolution case be possible for life insurance Claims Resolution clique fraud case,Can the life insurance Claims Resolution case of the automatic identification risk to be analyzed be the life insurance Claims Resolution case with clique's risk of fraud, compared to peopleIt is simpler otherwise, accurate that work is known.
As shown in figure 4, second embodiment of the invention proposes a kind of knowledge of the life insurance Claims Resolution case with clique's risk of fraudOther device, on the basis of above-described embodiment, the identifying device also includes：
Acquisition module 03, the sample data of the life insurance Claims Resolution case for belonging to clique's fraud for obtaining predetermined number, willThe sample data of life insurance Claims Resolution case under same life insurance Claims Resolution clique fraud case is classified as same sample data sets, willThe sample data of life insurance Claims Resolution case under different life insurance Claims Resolution clique fraud cases is classified as different sample data sets, withObtain the sample data sets that case is cheated by a number of life insurance Claims Resolution clique.If for example, life insurance Claims Resolution clique fraud case XAssociated life insurance Claims Resolution case includes tetra- life insurance Claims Resolution cases of X1, X2, X3 and X4, then by X1, X2, X3 and X4 corresponding longevityDanger Claims Resolution data are classified as the same sample data sets corresponding with life insurance Claims Resolution clique fraud case X.
Common characteristic determining module 04, for determining each sample data sets each from each sample data setsCorresponding common characteristic set, if for example, there is the different life insurances Claims Resolution cliques of A, B two to cheat the sample data sets of case,The corresponding common characteristic set a of A are determined from sample data sets A, determine that B is corresponding common from sample data sets BThere is characteristic set b.Wherein, the common characteristic is all life insurances Claims Resolution belonged under same life insurance Claims Resolution clique fraud caseThe feature that case is jointly comprised, for example, the common characteristic can be Y1 hospitals, Y2 life insurance business persons, Y3 work units, Y4 privatesRelationship, Y5 age ranges and/or Y6 native places etc..
Determining module 05 is associated, it is related in all common characteristic set for being determined according to predetermined correlation ruleThe combinations of features of connection, and this feature combination is combined as key feature.For example, can be by the feature in all common characteristic setIn common occurrence number do many combinations of features and combined as key feature.
In a kind of optional embodiment, above-mentioned association determining module 05 can be used for：
Feature in all common characteristic set is combined by default combination, the default combination include butCombination of two, multi-to-multi combination etc. are not limited to, multiple combinations of features are obtained, and each combinations of features is closed according to predeterminedConnection rule determines corresponding coupling index value；If there is the corresponding coupling index value of combinations of features to be more than corresponding predetermined threshold value,Then determine that this feature is combined as key feature combination.
Wherein, correlation rule is shape such as X → Y implications, wherein, X and Y are referred to as the guide of correlation ruleIt is (antecedent or left-hand-side, LHS) and follow-up (consequent or right-hand-side, RHS), at thisIn embodiment, X and Y can be each features in combinations of features, and the coupling index value includes support (support) and letterAppoint degree (confidence).For example, for a combinations of features for including fisrt feature X and second feature Y, calculating has simultaneouslyThe fisrt feature X and the second feature Y life insurance Claims Resolution case sample data account for the life insurance Claims Resolution case of predetermined numberThe percentage of sample data, using the support combined as this feature, i.e. life insurance Claims Resolution case of this feature combination in predetermined numberSupport in part sample data be predetermined number life insurance Claims Resolution case sample data in while comprising fisrt feature X, secondThe percentage of characteristic Y, i.e. probability.Calculate the life insurance Claims Resolution case with the fisrt feature X and the second feature YSample data accounts for the percentage of the life insurance Claims Resolution case sample data with the fisrt feature X, using what is combined as this featureThe degree of belief that degree of belief, i.e. this feature combine in the life insurance Claims Resolution case sample data of predetermined number is the life insurance of predetermined numberIn case sample data of settling a claim in the case of included fisrt feature X, second feature Y percentage, i.e. condition are included generalRate.Can be according to default support threshold and default degree of belief threshold value be preset the need for practical application, if there is combinations of features pairThe support answered is more than default support threshold, and corresponding degree of belief is more than default degree of belief threshold value, it is determined that this feature groupIt is combined into key feature combination；If there is the corresponding support of combinations of features to be less than or equal to default support threshold, and/or, it is rightThe degree of belief answered is less than or equal to default degree of belief threshold value, it is determined that this feature combination is not key feature combination, so as to controlMake whether case of being settled a claim to the life insurance of the risk to be analyzed is that the life insurance with clique's risk of fraud is settled a claim the analysis result of case.
If for example, all common characteristic set I={ close by Y1 hospitals, Y2 life insurance business persons, Y3 work units, Y4 individualsSystem }, there are 5 life insurance Claims Resolution case sample datas to be cured comprising Y1 in predetermined number is the life insurance Claims Resolution case sample data of 6, there are 3 life insurance Claims Resolution case sample datas in institute while comprising Y1 hospitals and Y2 life insurance business persons.Then for including Y1 hospitals and Y2For the combinations of features of life insurance business person, the support of this feature combination is while has the longevity of Y1 hospitals and Y2 life insurance business personsDanger Claims Resolution case sample data accounts for the percentage of the life insurance Claims Resolution case sample data of predetermined number, i.e. 3/6=0.5；This featureThe confidence level of combination is while has Y1 hospitals and the life insurance Claims Resolution case sample data of Y2 life insurance business persons to account for Y1 hospitalsLife insurance settle a claim case sample data percentage, i.e. 3/5=0.6.If presetting support threshold α=0.4, degree of belief thresholdValue β=0.5, then judge because Y1 hospitals support 0.5 corresponding with the combinations of features of Y2 life insurance business persons is more than default supportThreshold value 0.4 is spent, and corresponding degree of belief 0.6 is more than default degree of belief threshold value 0.5, accordingly, it can be determined that Y1 hospitals and Y2 life insurancesThe combinations of features of business person combines for key feature.
In the present embodiment, due to generally being entered by for example multiple participants of multiple features in life insurance Claims Resolution clique fraud caseRow clique is cheated, therefore, and being settled a claim by correlation rule from multiple life insurances, it is associated to be found during the common characteristic of case is cheated by cliqueKey feature combination, it is ensured that the key feature in the key feature combination found is used to carry out clique's fraud to be usualFeature.Risk analysis is carried out to life insurance Claims Resolution case using the key feature in the key feature combination found, can judgedWhen there is the key risk feature in the combination of one or more key features in the life insurance Claims Resolution case of the risk to be analyzed, automaticallyIt is the life insurance Claims Resolution case with clique's risk of fraud to recognize the life insurance Claims Resolution case of the risk to be analyzed.Compared to manual identifiedMode, the present embodiment can recognize that the internal association of different life insurances Claims Resolution cases so that the clique for case of being settled a claim to life insuranceRisk of fraud analysis more accurate and effective.
It should be noted that herein, term " comprising ", "comprising" or its any other variant are intended to non-rowHis property is included, so that process, method, article or device including a series of key elements not only include those key elements, andAnd also including other key elements being not expressly set out, or also include for this process, method, article or device institute inherentlyKey element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that including thisAlso there is other identical element in process, method, article or the device of key element.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment sideMethod can add the mode of required general hardware platform to realize by software, naturally it is also possible to be realized by hardware, but a lotIn the case of the former be more preferably embodiment.Understood based on such, technical scheme is substantially in other words to existingThe part that technology contributes can be embodied in the form of software product, and the computer software product is stored in a storageIn medium (such as ROM/RAM, magnetic disc, CD), including some instructions are to cause a station terminal equipment (can be mobile phone, calculateMachine, server, air conditioner, or network equipment etc.) perform method described in each of the invention embodiment.
Above by reference to the preferred embodiments of the present invention have been illustrated, not thereby limit to the interest field of the present invention.OnState that sequence number of the embodiment of the present invention is for illustration only, the quality of embodiment is not represented.Patrolled in addition, though showing in flow chartsOrder is collected, but in some cases, can be with the step shown or described by being performed different from order herein.
Those skilled in the art do not depart from the scope of the present invention and essence, can have a variety of flexible programs to realize the present invention,Feature for example as one embodiment can be used for another embodiment and obtain another embodiment.All technologies with the present inventionAny modifications, equivalent substitutions and improvements made within design, all should be within the interest field of the present invention.