WO2010133063A1 - System and method for short message monitoring - Google Patents

System and method for short message monitoring Download PDF

Info

Publication number
WO2010133063A1
WO2010133063A1 PCT/CN2009/074516 CN2009074516W WO2010133063A1 WO 2010133063 A1 WO2010133063 A1 WO 2010133063A1 CN 2009074516 W CN2009074516 W CN 2009074516W WO 2010133063 A1 WO2010133063 A1 WO 2010133063A1
Authority
WO
WIPO (PCT)
Prior art keywords
arbitration
short message
interface unit
result
fuzzy
Prior art date
Application number
PCT/CN2009/074516
Other languages
French (fr)
Chinese (zh)
Inventor
赵阳
陈苏
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2010133063A1 publication Critical patent/WO2010133063A1/en

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00Network architectures or network communication protocols for network security
    • H04L63/02Network architectures or network communication protocols for network security for separating internal from external traffic, e.g. firewalls
    • H04L63/0227Filtering policies
    • H04L63/0236Filtering by address, protocol, port number or service, e.g. IP-address or URL
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L51/00User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail
    • H04L51/58Message adaptation for wireless communication
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/12Messaging; Mailboxes; Announcements

Definitions

  • the present invention relates to short message monitoring technologies, and in particular, to a short message monitoring system and method based on machine learning arbitration results.
  • BACKGROUND Currently, various monitoring systems have been fully utilized in various fields such as short message monitoring and multimedia message monitoring, including a spam monitoring system. It should be noted here that a short message may also be referred to as a short message.
  • Most spam short message monitoring systems implement monitoring of spam short messages based on the set of limited keywords and monitoring rules of traffic statistics. For example, when the set traffic statistics count is reached, the current short message is determined to be a spam short message.
  • the main object of the present invention is to provide a short message monitoring system and method, which can improve the monitoring capability of a fuzzy short message, and can effectively determine whether the fuzzy short message is a garbage short message and realize short garbage. Monitoring of messages.
  • the technical solution of the present invention is achieved as follows: According to an aspect of the present invention, a short message monitoring system is provided.
  • the short message monitoring system includes: an interface unit and an arbitration unit; wherein, the interface unit is configured to forward the fuzzy short message to the arbitration unit for arbitration; The returned matching arbitration result is subjected to a weighting operation; when the operation result is less than the set first threshold value, the fuzzy short message is determined to be a violation short message; and the arbitration unit is configured to obtain the fuzzy short message by forwarding by the interface unit. ; Match the fuzzy short message according to different categories of arbitration rules, and return the matched arbitration results to the interface unit.
  • the foregoing arbitration rules include: a keyword rule, a user number type information rule, and a user number segment information rule.
  • the arbitration unit further includes at least two arbitration modules, and each of the arbitration rules included in each arbitration module belongs to a different category; each arbitration module is configured to match the fuzzy short message according to the respective included arbitration rules, and The matched arbitration results are returned to the sigma unit.
  • each of the foregoing arbitration modules is further configured to determine a fuzzy short message according to each arbitration result, and obtain a judgment result for the fuzzy short message by each arbitration module; and obtain a judgment result for the fuzzy short message by the interface unit,
  • the arbitration weight of the current arbitration module is increased; the probability that the current arbitration module and the interface unit have the same judgment result is smaller than the first In the expected value state, the arbitration weight of the current arbitration module is reduced; the modified arbitration weight is returned to the interface unit to update the arbitration weight.
  • the interface unit is further used to perform a weighting operation on the operation formula Result.
  • each of the foregoing arbitration modules is further configured to: when the arbitration result is greater than the set second threshold value, determine that the fuzzy short message is a violation short message; wherein, the second threshold is 1/n times the first threshold n is the number of arbitration modules.
  • the interface unit is further configured to increase the current first threshold value in a state in which the interface unit determines that the rate of the violation short message is greater than the second expected value within the set system running time; In the state where the 4 rate of the violation short message is less than 1 time of the second expected value, Reduce the current first threshold; return the modified first threshold to each arbitration module for update.
  • a short message monitoring method is provided.
  • the short message monitoring method includes: the arbitration unit respectively matches the fuzzy short message according to different categories of arbitration rules, and returns the matched arbitration results to the interface unit; the interface unit performs weighting operation according to the matched arbitration results; When the operation result is less than the set first threshold, it is determined that the fuzzy short message is a violation short message.
  • the arbitration rules include: a keyword rule, a user number type information rule, and a user number segment information rule.
  • the arbitration unit when the arbitration unit includes at least two arbitration modules, and the arbitration rules respectively included in the respective arbitration modules belong to different categories, the arbitration unit further performs: the arbitration modules respectively are short according to the respective arbitration rules and fuzzy rules. The messages match and the matching arbitration results are returned to the interface unit.
  • the method further includes: A. Each arbitration module respectively determines the fuzzy short message according to each arbitration result, and obtains a determination result for the fuzzy short message by each arbitration module;
  • the interface unit performs a force according to the returned updated arbitration weight value.
  • the weight operation and continue to perform the judgment for the fuzzy short message.
  • the operation formula used by the interface unit to perform the weighting operation is:
  • is the arbitration weight of each arbitration module, where n is the number of arbitration modules.
  • the determining, by each arbitration module, the fuzzy short message according to each arbitration result further comprises: when the arbitration result is greater than the set second threshold, determining that the fuzzy short message is a violation short message; wherein, the second valve The value is 1/n times the first threshold; n is the number of arbitration modules.
  • the method further includes: increasing, in the set system running time, the current first threshold value when the interface unit determines that the probability of the violation short message is greater than the second expected value; and determining that the violation is short in the interface unit If the probability of the message is less than 1 time of the second expected value, the current first threshold is decreased; and the modified first threshold is returned to each arbitration module for updating.
  • the arbitration unit of the present invention matches the fuzzy short message according to different types of arbitration rules, and returns the matched arbitration results to the interface unit; the interface unit performs weighting operation according to each of the matched arbitration results; when the operation result is smaller than the set When the first threshold is reached, it is determined that the fuzzy short message is a violation short message.
  • the invention provides a processing scheme for the fuzzy short message, which is a good supplement to the current monitoring system 4, and can greatly improve the ability to process the fuzzy short message without affecting the function of the current monitoring system, and actively respond to the new garbage short.
  • the message poses challenges to the current monitoring system and maximizes the real-time performance of the monitoring system.
  • the invention overcomes the defects in the current monitoring system that the fuzzy short message cannot be judged or relies on manual judgment, and proposes a monitoring system that matches the fuzzy short message and automatically judges according to different types of arbitration rules, so as to analyze the garbage more comprehensively.
  • the characteristics of short messages, and capture violations of short messages and violating users BRIEF DESCRIPTION OF THE DRAWINGS FIG.
  • FIG. 1 is a schematic structural diagram of a short message monitoring system according to an embodiment of the present invention
  • FIG. 2 is a schematic flowchart of an implementation of a short message monitoring method according to an embodiment of the present invention.
  • DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The basic idea of the present invention is: Matching fuzzy short messages according to different categories of arbitration rules, and automatically determining fuzzy short messages.
  • FIG. 1 is a schematic structural diagram of a short message monitoring system according to an embodiment of the present invention. As shown in FIG. 1, the short message monitoring system includes: an interface unit and an arbitration unit.
  • the interface unit is configured to forward the fuzzy short message to the arbitration unit for arbitration; perform weighting operation according to the matched arbitration results returned by the arbitration unit; and when the operation result is less than the first threshold value set in the interface unit, The fuzzy short message is determined to be a violation short message; when the operation result is greater than the set first threshold value, the fuzzy short message is determined to be a normal short message.
  • An arbitration unit configured to obtain a fuzzy short message by forwarding by the interface unit; respectively matching the fuzzy short message according to different categories of arbitration rules, and returning each matched arbitration result to the interface unit, so that the interface unit matches the matched arbitration result After the weighting operation, the first threshold value set in the interface unit is compared and the fuzzy short message is characterized.
  • the arbitration rules include: a keyword rule, a user number type information rule, and a user number segment information rule.
  • the short message monitoring system may further include: a monitoring unit and a human-machine interaction unit.
  • the monitoring unit and the human-machine interaction unit are both existing, and they are described herein.
  • the monitoring unit is configured to obtain short messages and count the number of traffic, and the number of statistics of the short messages and the monitoring unit
  • the configured traffic rules are matched to implement filtering of short message traffic; the short message with successful traffic matching is sent to the human-machine interaction unit, so that the human-computer interaction unit performs further filtering according to the keyword rule.
  • the human-machine interaction unit is configured to receive a short message with successful traffic matching from the monitoring unit, and match the keyword of the short message with successful traffic matching with the keyword rule configured in the human-computer interaction unit to implement filtering of the short message content.
  • the short message that the keyword is successfully matched is determined to be a violation short message, and the user who sends the violation short message is directly added to the blacklist, and the user who subsequently sends the violation short message can not send the short message again.
  • the short message that determines that the keyword match is unsuccessful is a fuzzy short message and is sent to the interface unit, so that further filtering is performed through the interaction between the interface unit and the arbitration unit, and the fuzzy is performed.
  • the short message is qualitatively determined to correctly and effectively determine whether the fuzzy short message is a violation short message or a normal short message.
  • the so-called fuzzy short message means: At this stage, it is impossible to correctly and effectively determine whether the short message is a suspicious message or a normal short message.
  • the non-compliant short message involved is a spam short message, which will not be described below.
  • the keyword rules used in the arbitration unit are complex, including not only a wide range of keywords, such as keywords in different categories including political, advertising, and security; but also Complex logical operation relationships between categories of keywords.
  • the monitoring unit specifically includes: a monitoring processing module and a monitoring management module.
  • the monitoring processing module is configured to obtain a short message from the short message center and collect the number of the traffic, and match the number of the statistics of the short message with the configured traffic rule; and send the short message with the successfully matched traffic to the monitoring management module.
  • the monitoring management module is configured to receive the short message with successful traffic matching and forward it to the human-machine interaction unit; and parse the user information in the short message with successful traffic matching and store it.
  • the so-called user information refers to: in the short message that the traffic matching succeeds, all the information encapsulated by the user who sends the short message includes: a primary and a called user number; a short message specific content including a keyword; Called user number type; primary and called user number segment information.
  • the human-computer interaction unit is further configured to configure an arbitration rule of the arbitration unit and synchronize to the arbitration unit.
  • the configuration may be configured by using a command line input manner, or may be configured by using a batch file command, so that the arbitration unit can obtain the time in time.
  • Arbitration rules after synchronization update according to the synchronously updated arbitration rules, can arbitrate fuzzy short messages in a timely, correct and effective manner.
  • the arbitration rule is the above-mentioned arbitration rules of a plurality of categories including a keyword rule, a user number type information rule, and a user number segment information rule.
  • the human-machine interaction unit is further configured to display monitoring information including a fuzzy short message and a judgment result obtained after the fuzzy short message is arbitrated, so as to visually display the information of the monitoring and management module and the interface unit returning the human-computer interaction unit, in time Adjusting the configuration information in the human-computer interaction unit is beneficial for efficient monitoring.
  • the arbitration unit further includes a plurality of arbitration modules, and each of the arbitration rules included in each arbitration module belongs to a different category; each arbitration module is configured to match the fuzzy short message according to the respective included arbitration rules, and each of the matched arbitration messages The arbitration result is returned to the interface unit.
  • each arbitration module is further configured to determine the fuzzy short message according to each arbitration result, and obtain a judgment result for the fuzzy short message by each arbitration module; obtain the judgment result of the fuzzy short message by the interface unit, and set During the system running time, when the current arbitration module and the interface unit determine that the same rate is greater than the first expected value, the arbitration weight of the current arbitration module is increased; the probability that the current arbitration module and the interface unit determine the same result is less than the first expected value.
  • the interface unit is further configured to perform a weighting operation according to the returned updated arbitration weight value, and continue to perform the determination for the fuzzy short message.
  • the interface unit is further used for adding
  • the operation formula used in the weight operation is: Result w ; where Result is the result of the operation, Result ⁇ is the result of each arbitration, and is the arbitration weight of each arbitration module, where n is the number of arbitration modules.
  • each arbitration module is further configured to determine that the fuzzy short message is a violation short message when the arbitration result is greater than the set second threshold value; The second threshold is 1/n times the first threshold.
  • the result of each arbitration is ReSult '-.
  • the arbitration module specifically includes: an arbitration module including a keyword rule, an arbitration module including a user number type information rule, and an arbitration module including a user number segment information rule.
  • the keyword rules include: keyword rules of at least one of the political class, the advertising class, and the security class, that is, the keyword rule can be a keyword rule of any one of the categories, such as a political class. Keyword rules; or keyword rules of combined categories in these categories, such as the keyword rules of the political category plus the security category, the keyword rules using this combination category can express complex logical operation relationships, Therefore, the arbitration of the fuzzy short message can be better realized.
  • the user number type information rule includes: a primary and a called user number type information rule.
  • the user number segment information rule includes: a primary and a called user number segment information rule.
  • the interface unit is further configured to determine, in the set system running time, that the probability of the violation short message is greater than the second expected value in the interface unit. In the state of 1 time, the current first threshold is increased; and when the interface unit determines that the probability of the violation short message is less than 1 time of the second expected value, the current first threshold is decreased. Returning the modified first threshold to each arbitration module for updating; by interacting with each arbitration module, and updating the arbitration weight, the first threshold, and implementing the judgment result of the interface unit for the fuzzy short message at the second expected value In the way of convergence in the upper and lower 20% range, the completion of the fuzzy short message is the judgment of the violation short message.
  • the core component of the system is an arbitration unit, an arbitration module in the case where the arbitration unit includes at least one arbitration module, and an interface unit.
  • the arbitration module judges whether the fuzzy short message is a violation short message: based on the configured keyword content, and the judgment of the complex logical relationship between the keywords contained in the fuzzy short message, and adds the fuzzy
  • the short message main and the called user's state judgment are comprehensively considered, and the judgment for the fuzzy short message is automatically realized by the interaction of at least one arbitration module and the interface unit.
  • the arbitration weight is learned through the interaction between the at least one arbitration module and the interface unit, and after the system runs for a period of time according to the set system running time, the optimal arbiter and the arbitration weight can be finally obtained. combination. That is, the interaction between the at least one arbitration module and the interface unit, and updating the arbitration weight, the first threshold, and the implementation of the interface unit for the fuzzy short message converge within the upper and lower 20% of the second expected value. The way to complete the fuzzy short message is to judge the violation of the short message.
  • FIG. 2 is a schematic diagram of an implementation process of a short message monitoring method according to an embodiment of the present invention. As shown in FIG.
  • the short message monitoring method includes the following steps: Step S101: The interface unit forwards the fuzzy short message to the arbitration unit for arbitration, and the arbitration unit respectively matches the fuzzy short message according to different types of arbitration rules, and The individual arbitration results are returned to the interface unit.
  • the arbitration rules include: a keyword rule, a user number type information rule, and a user number segment information rule.
  • the keyword rule includes: a keyword rule of at least one of a political class, an advertisement class, and a security class;
  • the user number type information rule includes: a primary and a called user number type information rule;
  • the user number segment information rule includes: The called user number segment information rule.
  • Step S102 The interface unit performs a weighting operation according to the matched arbitration results returned by the arbitration unit. When the operation result is less than the set first threshold, it is determined that the fuzzy short message is a violation short message.
  • the operation formula used by the interface unit to perform the weighting operation is: ⁇ Result. x ⁇ w.
  • the step S101 may further include: the monitoring unit acquires the short message and counts the number of the traffic, and the number of the traffic of the short message is matched with the configured traffic rule; The short message with successful traffic matching is sent to the human interaction unit. After that, the human interaction unit matches the keyword of the short message with successful traffic matching with the configured keyword rule; the short message with the keyword matching success is determined as the violation short message; The unsuccessful short message is judged as a fuzzy short message and sent to the interface unit.
  • the method may further include: initially configuring the arbitration rule and synchronizing to the arbitration unit.
  • the human interaction unit synchronously updates the arbitrated arbitration rule to the arbitration unit.
  • the human-computer interaction unit can display the monitoring information including the fuzzy short message and the judgment result obtained after the fuzzy short message is arbitrated. Therefore, after the monitoring information is displayed, the configuration of the arbitration rule can be selectively modified according to the displayed monitoring information. And update to the arbitration unit synchronously.
  • the processing of step S101 further includes: each arbitration module separately according to the arbitration rules respectively included Matches with the fuzzy short message, and returns the matched arbitration results to the interface unit.
  • Step S103 Each arbitration module determines the fuzzy short message according to each arbitration result, and obtains a determination result for the fuzzy short message by each arbitration module.
  • Step S104 Each arbitration module acquires a determination result of the fuzzy short message by the interface unit.
  • Step S105 In the set system running time, when the probability that the current arbitration module and the interface unit determine the same result is greater than the first expected value, increase the arbitration weight of the current arbitration module; and the current arbitration module and the interface unit determine the same result. When the probability is less than the first expected value, the arbitration weight of the current arbitration module is reduced; and the 4 modified arbitration weight is returned to the interface unit for updating the arbitration weight.
  • Step S106 The interface unit performs a force according to the returned updated arbitration weight value. The weight operation, and continue to perform the judgment for the fuzzy short message.
  • the method may further include: Step S107: In the set system running time, when the interface unit determines that the 3 ⁇ 4 ⁇ 4 rate of the violation short message is greater than the second expected value, the current first threshold is increased; and the probability of the violation short message is determined by the interface unit is less than In the case where the second expected value is 1 time, the current first threshold value is decreased.
  • the manner in which the fuzzy short message is completed is a violation of the short message in a manner of convergence in the upper and lower 20% of the second expected value.
  • the first embodiment of the method is as follows:
  • the monitoring unit includes a monitoring processing module and a monitoring management module.
  • the arbitration unit includes multiple arbitration modules, the arbitration function and the interface unit do not need to perform multiple interactions of the arbitration module to learn and update the arbitration weight.
  • the monitoring process for implementing the short message includes the following steps: Step S201:
  • the monitoring processing module receives the short message sent by the user from the short message center.
  • the short message encapsulates the primary and called user numbers; the short message specific content including the keyword; the primary and the called user number type; the primary and the called user number segment information and the like.
  • Step S202 The monitoring processing module performs statistical counting on the received short message, and matches the traffic rule configured by the system, and sends a short message with successful traffic matching to the monitoring management module, and the monitoring management module imports the short message with successful traffic matching.
  • the database stores and sends a short message with successful traffic matching to the human interaction unit.
  • the human-computer interaction unit may also be referred to as a console, and the short message with successful traffic matching is a violation short message, and the corresponding user is a violation user.
  • Step S203 The human-machine interaction unit displays the offending user, and filters the content of the short message with the matching of the traffic, and matches the keyword of the short message with the matching of the traffic to the keyword rule configured in the human-computer interaction unit.
  • the short message that the keyword is successfully matched is determined to be a violation short message; if the matching is unsuccessful, the short message with the unsuccessful keyword matching is determined to be fuzzy
  • the short message is sent to the interface unit.
  • the keyword rules configured in the human-computer interaction unit are the same as the existing keyword rules, and may also be referred to as specific keyword rules.
  • the violating user corresponding to the violation short message containing the specific keyword is directly added to the blacklist, and the subsequent violating user can no longer send the short message, and the fuzzy short message that does not contain the specific keyword, mainly including the political category and the security keyword, is sent. To the interface unit processing.
  • Step S204 The interface unit sends the fuzzy short message that needs to be arbitrated to the n arbitration modules, and configures the initial arbitration weight of each arbitration module, and the arbitration module times out, the first threshold After waiting for the information, wait for the response of each arbitration module.
  • Step S205 Each arbitration module arbitrates the fuzzy short message to be arbitrated.
  • Step S206 Each arbitrator returns the arbitration result to the interface unit.
  • Step S207 The interface unit performs a weighting operation according to the arbitration result returned by each arbitration module, compares and determines with the set first threshold, and characterizes the fuzzy short message.
  • the so-called qualitative short message refers to: Determine whether the nature of the fuzzy short message is a violation short message or a normal short message. Step S208, ending the current short message monitoring process.
  • the second embodiment of the method is as follows: In the case that the arbitration unit includes multiple arbitration modules, the interaction between the arbitration module and the interface unit is required to learn and update the arbitration weight.
  • the monitoring process of the short message includes the following steps. Step S301: The human-machine interaction unit of the short message monitoring completes the basic attribute configuration of the system, the configuration of the monitoring rule, and the related configuration of the arbitration module, and synchronizes the monitoring rules.
  • each arbitration module imports a related configuration for the arbitration module.
  • the related configuration of the imported arbitration module includes a monitoring rule configured for the arbitration module, and the monitoring rule includes at least one category of arbitration rules including a keyword rule, a user type information rule, and a segment information rule.
  • the keyword rule configured in the arbitration module is different from the keyword rule configured in the human-computer interaction unit in the human-computer interaction unit, and the keyword rule configured in the arbitration module has a very large amount of information.
  • Keyword basically include: all political, advertising, and security-type violation characters and complex logical operations between them.
  • the user type information includes: all blacklist and whitelist information, online charging system (OCS) arrears user information, and packet format protocol (H2) attribute information with the camp interface.
  • Type information For the segment information rule, the segment information includes: all prepaid number segments, the province number segment, the network segment segment, and the special user segment segment, that is, the whitelist segment segment.
  • the user Type information rule master and called user number type information rules; User number segment information rules include: primary and called user number segment information rules.
  • the monitoring unit receives the short message of the short message center, performs traffic counting on the short message, and matches the traffic rule configured by the system. The short message with the successfully matched traffic is determined as the violation short message, and the violation of the traffic monitoring rule is reached. The violating user corresponding to the short message is displayed by the human-computer interaction unit.
  • the human-machine interaction unit filters the content of the violation short message, that is, matches the keyword of the short message with successful traffic matching with the keyword rule configured in the human-computer interaction unit to implement the short message content. Filtering.
  • Step S305 The fuzzy short message generated after filtering is sent to each arbitration module through an interface unit for arbitration, and each arbitration module arbitrates the fuzzy short message to be arbitrated.
  • the arbitration result of initializing each arbitration module is 20, and the arbitration module can be divided into the following five categories, as follows:
  • the first type of arbitration module is: an arbitration module containing a keyword rule, used for The fuzzy short messages are classified and arbitrated according to the keyword rules. Different keywords are divided into several levels according to their importance, which are 10 points, 8 points, 5 points, 2 points and 4 levels. Among them, political categories: such as Falun Gong, Tiananmen incidents, etc. are configured as type 1 keywords; security categories: such as self-immolation, sit-down, pistol, etc.
  • the second type of arbitration module is: a calling user type arbitration module including a calling party number type information rule, configured to perform number classification and arbitration on the calling user according to the calling party number type information rule.
  • the number is divided into the following levels: whitelist users plus 10 points, blacklist users minus 20 points, OCS arrears users deduct 10 points, H2 attributes are diamond card users plus 10 points, gold card users plus 5, silver card users force. 3 points, ordinary card users do not add points, the main and called users are matched to get the arbitration result, the score is deducted to 0 and then treated as 0 points. Among them, the diamond card user, the gold card user, the 4 Leica user, and the ordinary card user are different ratings of the operator in the camping system.
  • the third type of arbitration module is: a called user type arbitration module that includes a called user number type information rule, configured to perform number classification and arbitration on the called user according to the called user number type information rule.
  • the number is divided into the following levels: whitelist users plus 10 points, blacklist users minus 20 points, OCS arrears users deduct 10 points, H2 attributes are diamond card users plus 10 points, gold card users plus 5, silver card users force. 3 points, the ordinary card does not add points, the matching user is matched to get the arbitration result, and the score is deducted to 0 and then treated as 0.
  • the fourth type of arbitration module is: a calling number segment information arbitration module including a calling party number segment information rule, configured to classify and arbitrate the calling party according to the calling party number segment information rule.
  • the number is divided into the following levels: VIP group number plus 10 points, global number section plus 5 points, Monternet gateway plus 5 points, industry gateway does not add points, prepaid number paragraph does not add points, foreign province number minus 2
  • the sub-group and the external network number segment are reduced by 5 points, etc., and the matching result is obtained after matching the calling and called users, and the score is deducted to 0 and then treated as 0 points. All of the above segment information is configured by the operator. Among them, the VIP group number segment is the number provided by the operator to the large group users.
  • the number of calls made by the users in the number segment is low, and the short number can be dialed directly;
  • the Monternet Gateway is a service provider directly operated by the operator;
  • the gateway is a service provider operated by a non-operator;
  • the prepaid number segment is a number segment set by the operator, and all numbers in the number segment are all prepaid users.
  • the fifth type of arbitration module is: a called number segment information arbitration module that includes a called user number segment information rule, and is used for classifying and arbitrating the called user according to the called user number segment information rule.
  • the number is divided into the following levels: VIP group number plus 10 points, global number section plus 5 points, Monternet gateway plus 5 points, industry gateway does not add points, prepaid number paragraph does not add points, foreign province number minus 2
  • the sub-network and the external network number segment are deducted by 5 points, etc., and the matching result is obtained after matching the called user, and the score is deducted to 0 and then treated as 0.
  • Step S306 Each arbitration module sends the arbitration result to the interface unit, and the interface unit performs a weighting operation on the arbitration result.
  • Step S307 the interface unit returns the operation result to each arbitration module, and the arbitration module first
  • the operation result is compared with the set second threshold, where the second threshold of each arbitration module is the same, which is 1/5 times the first threshold set by the interface unit.
  • the operation result is less than the set second threshold, it is determined that the fuzzy short message is a normal short message; when the operation result is greater than the set second threshold, it is determined that the fuzzy short message is a violation short message.
  • Step S308 comparing whether the judgment result of the fuzzy short message by each arbitration module is the same as the judgment result of the fuzzy short message by the interface unit, if the same, determining that the judgment result by the arbitration module for the fuzzy short message is correct, otherwise determining that the error is incorrect .
  • Step S309 After the arbitration module runs for a period of time, the weight learning is performed.
  • the period is the default system running time, which can be 1 hour, and the current arbitration module determines the correct rate and the first time during the calculated period.
  • the first expected value here may be set to 50%; if the probability of the current arbitration module determining that the correctness is greater than 50%, the arbitration weight of the current arbitration module is increased, for example, the arbitration weight of the next stage is increased by 0.1; If the current arbitration module judges that the correct phase ratio is less than 50%, the arbitration weight of the current arbitration module is reduced, for example, the arbitration weight of the next stage is reduced by 0.1; the modified arbitration weight is sent to the interface unit, and the interface machine performs The renewal of the arbitration weight.
  • Step S310 After the interface machine runs for a period of time, the period of time refers to the default system running time, which may be 1 hour, and compares the calculated probability of the violation short message with the second expected value during the period, where the second The expected value can be initially set to 40%, but not more than 50%; if the probability of the violation short message is less than 1 times of the second expected value, then the current first threshold is reduced, such as reducing the first threshold by 5; If the 3 ⁇ 4 ⁇ 4 rate of the short message is greater than the second expected value by one time, the current first threshold is increased, for example, the first threshold is increased by 5; and the fourth modified threshold is sent to each arbitration module, and each arbitration is performed.
  • Step S311 Write a result of the percentage of the violation short message to the total message within 24 hours every day, and the file is a file including the judgment result of the interface unit for the fuzzy short message, and observe whether the judgment result converges within a range, if the convergence The range is small, and around the second expected value, such as convergence in the upper and lower 20% of the second expected value, the system's arbitration weight and the first and second threshold training results are considered normal; if the result is divergent, then After the second expected value is modified, the training is performed, and the judgment result is observed one day later and the corresponding processing is performed.
  • the present invention it takes a certain time to perform training in the initial stage of the system operation, but after training the optimal expected value and the arbitration weight, the automatic arbitration result of the fuzzy short message is compared with the existing manual. Judgment has been greatly improved, including the accuracy and automation of judgment. It is a good complement to the existing short message monitoring system. At the same time, it can analyze some behavior habits of users sending spam messages, and provides some risks for improving existing systems. Since the existing short message monitoring system can only monitor based on the traffic rules or the keyword rules of the single ticket, that is, the keyword rules configured in the human-computer interaction unit mentioned above, it is easy to disable the monitoring and monitor There are fewer and fewer short messages.
  • the threshold of traffic reduction is reduced, the number of monitored short messages will increase by an order of magnitude, and the workload of manual operation and maintenance will be large.
  • the content of new violation short messages is endless. Only adding limited and individual keywords will not meet the monitoring requirements. In particular, some new content violation short message delivery modes are difficult to configure only keywords for monitoring.
  • the invention greatly improves the performance of the system by increasing the complex operation relationship between the keywords; and the monitoring system of the present invention automatically learns the arbitration weight through the machine based on the interaction between the arbitration module and the interface unit, thereby A reliable arbitration result can be obtained, which is a great improvement for the existing monitoring system.
  • the present invention classifies and arbitrates the short message of the violation, and arbitrates the short message that does not contain the specific keyword in the message content, and the fuzzy short message is automatically arbitrated by the arbitration module.
  • the message content of the message, the type of the main and called numbers, the type of the main and the called number, and the like, are comprehensively analyzed.
  • whether the fuzzy short message is a junk short message is judged, and the present invention is adopted than the existing monitoring system.
  • the single maintenance personnel carry out manual and subjective judgment programs, and the accuracy and efficiency are greatly improved.
  • the above modules or steps of the present invention can be implemented by a general-purpose computing device, which can be concentrated on a single computing device or distributed over a network composed of multiple computing devices.
  • the invention is not limited to any specific combination of hardware and software.
  • the above is only the preferred embodiment of the present invention and is not intended to limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications and changes can be made in the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Hardware Design (AREA)
  • Computer Security & Cryptography (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer And Data Communications (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

A system and a method for short message monitoring are provided, in the system, the interface unit (1) is used for weighting the arbitrating results after being matched returned by each arbitrating unit (2) and judging whether the blur short message is the short message of breaking the rule; the arbitrating unit (2) is used for obtaining the blur short message transmitted by the interface unit (1), matching different kinds of arbitrating rules with blur short message respectively and returning each arbitrating result after being matched to the interface unit (1). The method for short message monitoring includes that: the arbitrating unit (2) matches different kinds of arbitrating rules with blur short message respectively, and returns each arbitrating result after being matched to the interface unit (1); the interface unit (1) weights each arbitrating result after being matched; if the arbitrating result is less than the preset first threshold, the blur short message is judged as the short message of breaking the rule. The system and the method of the invention improve the monitoring ability for the blur short message.

Description

一种短消息监控系统及方法  Short message monitoring system and method
技术领域 本发明涉及短消息监控技术,尤其涉及一种基于机器学习仲裁结果的短 消息监控系统及方法。 背景技术 当前 , 各种监控系统已经在短信监控、 彩信监控等各种领域中得到充分 使用, 其中也包括垃圾短信监控系统, 此处需要指出的是短信也可以称为短 消息。 大多数垃圾短消息监控系统都是基于设置的有限关键字、 和流量统计 计数的监控规则实现对垃圾短消息的监控, 例如, 当达到设置的流量统计计 数时确定当前短消息为垃圾短消息。 随着时代的发展, 用户发送消息的行为和内容发生了很大的变化, 呈现 多样化和复杂化的特点 , 仅仅凭现有流量和有限的关键字这种仅体现筒单逻 辑的监控规则, 是无法正确判断出短消息是否为垃圾短消息的。 这里, 通过 流量规则和筒单的关键字规则这种现有的监控规则, 无法正确判断出是否为 垃圾短消息的可疑消息称为模糊短消息。 如果无上限的不断增加关键字的数 量, 以?丈善对模糊短消息的监控能力 , 则会大大影响监控系统的实时性能。 目前, 针对如何有效地判断出模糊短消息是否为垃圾短消息, 并实现对垃圾 短消息的监控, 尚不存在有效的解决方案。 发明内容 有鉴于此, 本发明的主要目的在于提供一种短消息监控系统及方法, 改 善对模糊短消息的监控能力, 可以有效地判断出模糊短消息是否为垃圾短消 息, 并实现对垃圾短消息的监控。 为达到上述目的, 本发明的技术方案是这样实现的: 才艮据本发明的一个方面, 提供了一种短消息监控系统。 根据本发明的短消息监控系统包括: 接口单元和仲裁单元; 其中, 接口单元, 用于将模糊短消息转发给仲裁单元进行仲裁; 根据仲裁单元 返回的匹配后的各个仲裁结果进行加权运算; 在运算结果小于设置的第一阀 值状态下, 判断出模糊短消息为违规短消息; 仲裁单元, 用于经由接口单元的转发获取到模糊短消息; 根据不同类别 的仲裁规则分别与模糊短消息匹配 , 并将匹配后的各个仲裁结果返回接口单 元。 优选地, 上述仲裁规则包括: 关键字规则、 用户号码类型信息规则、 用 户号段信息规则。 优选地, 上述仲裁单元, 进一步包括至少两个仲裁模块, 且各个仲裁模 块中各自包含的仲裁规则分别属于不同类别; 各个仲裁模块, 用于分别根据 各自包含的仲裁规则与模糊短消息匹配, 并将匹配后的各个仲裁结果返回接 σ单元。 优选地, 上述各个仲裁模块, 进一步用于分别根据各个仲裁结果对模糊 短消息进行判断, 并获得通过各个仲裁模块针对模糊短消息的判断结果; 获取通过接口单元针对模糊短消息的判断结果,在设置的系统运行时间 内,在当前仲裁模块与接口单元判断结果相同的概率大于第一期望值状态下, 增加当前仲裁模块的仲裁权值; 在当前仲裁模块与接口单元判断结果相同的 概率小于第一期望值状态下, 减少当前仲裁模块的仲裁权值; 将修改后的仲 裁权值返回接口单元进行仲裁权值的更新。 优选地, 上述接口单元, 进一步用于进行加权运算时所采用的运算公式 Result. X ω{ TECHNICAL FIELD The present invention relates to short message monitoring technologies, and in particular, to a short message monitoring system and method based on machine learning arbitration results. BACKGROUND Currently, various monitoring systems have been fully utilized in various fields such as short message monitoring and multimedia message monitoring, including a spam monitoring system. It should be noted here that a short message may also be referred to as a short message. Most spam short message monitoring systems implement monitoring of spam short messages based on the set of limited keywords and monitoring rules of traffic statistics. For example, when the set traffic statistics count is reached, the current short message is determined to be a spam short message. With the development of the times, the behavior and content of users sending messages have undergone great changes, presenting diverse and complex features, and only relying on existing traffic and limited keywords, this only reflects the logic of the single logic. It is impossible to correctly determine whether the short message is a spam short message. Here, the existing monitoring rule, such as the traffic rule and the keyword rule of the ticket, cannot correctly determine whether the suspicious message of the garbage short message is called a fuzzy short message. If there is no limit, keep increasing the number of keywords to? The ability of Shanshan to monitor fuzzy short messages will greatly affect the real-time performance of the monitoring system. At present, there is no effective solution to how to effectively determine whether a fuzzy short message is a garbage short message and implement monitoring of the garbage short message. SUMMARY OF THE INVENTION In view of this, the main object of the present invention is to provide a short message monitoring system and method, which can improve the monitoring capability of a fuzzy short message, and can effectively determine whether the fuzzy short message is a garbage short message and realize short garbage. Monitoring of messages. To achieve the above object, the technical solution of the present invention is achieved as follows: According to an aspect of the present invention, a short message monitoring system is provided. The short message monitoring system according to the present invention includes: an interface unit and an arbitration unit; wherein, the interface unit is configured to forward the fuzzy short message to the arbitration unit for arbitration; The returned matching arbitration result is subjected to a weighting operation; when the operation result is less than the set first threshold value, the fuzzy short message is determined to be a violation short message; and the arbitration unit is configured to obtain the fuzzy short message by forwarding by the interface unit. ; Match the fuzzy short message according to different categories of arbitration rules, and return the matched arbitration results to the interface unit. Preferably, the foregoing arbitration rules include: a keyword rule, a user number type information rule, and a user number segment information rule. Preferably, the arbitration unit further includes at least two arbitration modules, and each of the arbitration rules included in each arbitration module belongs to a different category; each arbitration module is configured to match the fuzzy short message according to the respective included arbitration rules, and The matched arbitration results are returned to the sigma unit. Preferably, each of the foregoing arbitration modules is further configured to determine a fuzzy short message according to each arbitration result, and obtain a judgment result for the fuzzy short message by each arbitration module; and obtain a judgment result for the fuzzy short message by the interface unit, During the set system running time, when the probability that the current arbitration module and the interface unit determine the same result is greater than the first expected value, the arbitration weight of the current arbitration module is increased; the probability that the current arbitration module and the interface unit have the same judgment result is smaller than the first In the expected value state, the arbitration weight of the current arbitration module is reduced; the modified arbitration weight is returned to the interface unit to update the arbitration weight. Preferably, the interface unit is further used to perform a weighting operation on the operation formula Result. X ω {
为: Result '4 ; 其中, Result 为运算结果, Result'为各个仲裁结 果, ^为各个仲裁模块的仲裁权值, n为仲裁模块的个数。 优选地, 上述各个仲裁模块, 进一步用于在仲裁结果大于设置的第二阀 值状态下, 判断出模糊短消息为违规短消息; 其中, 第二阀值为第一阀值的 1/n倍; n为仲裁模块的个数。 优选地, 上述接口单元, 进一步用于在设置的系统运行时间内, 在接口 单元判断出违规短消息的 率大于第二期望值 1倍的状态下, 增加当前第一 阀值; 在接口单元判断出违规短消息的 4既率小于第二期望值 1倍的状态下, 减少当前第一阀值; 将修改后的第一阀值返回各个仲裁模块进行更新。 才艮据本发明的另一方面, 提供了一种短消息监控方法。 根据本发明的短消息监控方法包括: 仲裁单元根据不同类别的仲裁规则分别与模糊短消息匹配 ,并将匹配后 的各个仲裁结果返回接口单元; 接口单元根据匹配后的各个仲裁结果进行加权运算; 当运算结果小于设 置的第一阀值时, 判断出模糊短消息为违规短消息。 优选地, 仲裁规则包括: 关键字规则、 用户号码类型信息规则、 用户号 段信息规则。 优选地, 当仲裁单元包括至少两个仲裁模块, 且各个仲裁模块中各自包 含的仲裁规则分别属于不同类别情况下, 仲裁单元进行匹配进一步包括: 各 个仲裁模块分别根据各自包含的仲裁规则与模糊短消息匹配, 并将匹配后的 各个仲裁结果返回接口单元。 其中, 上述方法进一步包括: A、 各个仲裁模块分别根据各个仲裁结果对模糊短消息进行判断, 并获 得通过各个仲裁模块针对模糊短消息的判断结果; It is: Result '4; where Result is the result of the operation, Result ' is the result of each arbitration, ^ is the arbitration weight of each arbitration module, and n is the number of arbitration modules. Preferably, each of the foregoing arbitration modules is further configured to: when the arbitration result is greater than the set second threshold value, determine that the fuzzy short message is a violation short message; wherein, the second threshold is 1/n times the first threshold n is the number of arbitration modules. Preferably, the interface unit is further configured to increase the current first threshold value in a state in which the interface unit determines that the rate of the violation short message is greater than the second expected value within the set system running time; In the state where the 4 rate of the violation short message is less than 1 time of the second expected value, Reduce the current first threshold; return the modified first threshold to each arbitration module for update. According to another aspect of the present invention, a short message monitoring method is provided. The short message monitoring method according to the present invention includes: the arbitration unit respectively matches the fuzzy short message according to different categories of arbitration rules, and returns the matched arbitration results to the interface unit; the interface unit performs weighting operation according to the matched arbitration results; When the operation result is less than the set first threshold, it is determined that the fuzzy short message is a violation short message. Preferably, the arbitration rules include: a keyword rule, a user number type information rule, and a user number segment information rule. Preferably, when the arbitration unit includes at least two arbitration modules, and the arbitration rules respectively included in the respective arbitration modules belong to different categories, the arbitration unit further performs: the arbitration modules respectively are short according to the respective arbitration rules and fuzzy rules. The messages match and the matching arbitration results are returned to the interface unit. The method further includes: A. Each arbitration module respectively determines the fuzzy short message according to each arbitration result, and obtains a determination result for the fuzzy short message by each arbitration module;
B、 获取通过接口单元针对模糊短消息的判断结果; 在设置的系统运行 时间内 , 在当前仲裁模块与接口单元判断结果相同的概率大于第一期望值情 况下, 增加当前仲裁模块的仲裁权值; 在当前仲裁模块与接口单元判断结果 相同的概率小于第一期望值情况下, 减少当前仲裁模块的仲裁权值; 将修改 后的仲裁权值返回接口单元进行仲裁权值的更新; B. Obtain a judgment result of the fuzzy short message by the interface unit; and increase the arbitration weight of the current arbitration module when the probability that the current arbitration module and the interface unit have the same judgment result is greater than the first expected value in the set system running time; If the probability that the current arbitration module and the interface unit determine the same result is less than the first expected value, reduce the arbitration weight of the current arbitration module; and return the modified arbitration weight to the interface unit to update the arbitration weight;
C、 接口单元根据返回的更新后的仲裁权值进行力。权运算, 并继续执行 针对模糊短消息的判断。 优选地, 上述接口单元进行加权运算时所采用的运算公式为: C. The interface unit performs a force according to the returned updated arbitration weight value. The weight operation, and continue to perform the judgment for the fuzzy short message. Preferably, the operation formula used by the interface unit to perform the weighting operation is:
^Result. x<w. ^Result. x<w.
Result= '=i ; 其中, Result 为运算结果, Result'为各个仲裁结果, Result= '=i ; where Result is the result of the operation and Result ' is the result of each arbitration.
^为各个仲裁模块的仲裁权值 , n为仲裁模块的个数。 优选地, A中, 各个仲裁模块分别根据各个仲裁结果对模糊短消息进行 判断进一步包括: 当仲裁结果大于设置的第二阀值时, 判断出模糊短消息为违规短消息; 其中, 第二阀值为第一阀值的 1/n倍; n为仲裁模块的个数。 优选地, C后还包括: 在设置的系统运行时间内 ,在接口单元判断出违规短消息的概率大于第 二期望值 1倍的情况下, 增加当前第一阀值; 在接口单元判断出违规短消息 的概率小于第二期望值 1倍的情况下, 减少当前第一阀值; 将修改后的第一 阀值返回各个仲裁模块进行更新。 本发明仲裁单元才艮据不同类别的仲裁规则分别与模糊短消息匹配 ,并将 匹配后的各个仲裁结果返回接口单元; 接口单元根据匹配后的各个仲裁结果 进行加权运算; 当运算结果小于设置的第一阀值时, 判断出模糊短消息为违 规短消息。 本发明提供了对模糊短消息的处理方案 , 是对当前监控系统 4艮好的补 充, 可以在不影响当前监控系统功能的前提下大大提高其处理模糊短消息的 能力 , 积极应对新的垃圾短消息对当前监控系统提出的挑战 , 并且最大限度 的保证监控系统的实时性。 采用本发明, 克服了当前监控系统中对于模糊短 消息无法判断或者依赖于人工判断的缺陷, 提出根据不同类别的仲裁规则分 别与模糊短消息匹配并进行自动判断的监控系统 , 以便更全面分析垃圾短消 息的特征, 并捕获违规短消息和违规用户。 附图说明 图 1为才艮据本发明实施例的短消息监控系统的组成结构示意图; 图 2为根据本发明实施例的短消息监控方法的实现流程示意图。 具体实施方式 本发明的基本思想是: 根据不同类别的仲裁规则分别与模糊短消息匹 配, 并对模糊短消息进行自动判断。 下面结合附图对技术方案的实施作进一步的详细描述。 图 1为才艮据本发明实施例的短消息监控系统的组成结构示意图。 如图 1 所示, 该短消息监控系统包括: 接口单元和仲裁单元。 其中, 接口单元, 用 于将模糊短消息转发给仲裁单元进行仲裁; 根据仲裁单元返回的匹配后的各 个仲裁结果进行加权运算; 在运算结果小于接口单元中所设置的第一阀值状 态下, 判断出模糊短消息为违规短消息; 在运算结果大于设置的第一阀值状 态下, 判断出模糊短消息为正常短消息。 仲裁单元, 用于经由接口单元的转 发获取到模糊短消息; 根据不同类别的仲裁规则分别与模糊短消息匹配, 并 将匹配后的各个仲裁结果返回接口单元 , 以便接口单元对匹配后的仲裁结果 进行加权运算后, 与接口单元中设置的第一阀值进行比较并对模糊短消息定 性。 其中, 仲裁规则包括: 关键字规则、 用户号码类型信息规则、 用户号段 信息规则。 上述短消息监控系统还可以包括: 监控单元和人机交互单元。 其中, 监 控单元和人机交互单元都是现有的, 在此对它们筒单阐述, 监控单元用于获 取短消息并统计流量个数, 将统计的短消息的流量个数与监控单元中所配置 的流量规则匹配, 以实现对短消息流量的过滤; 将流量匹配成功的短消息发 送给人机交互单元 , 以便人机交互单元按照关键字规则执行进一步的过滤。 人机交互单元, 用于从监控单元接收流量匹配成功的短消息, 将流量匹 配成功的短消息的关键字与人机交互单元中所配置的关键字规则匹配, 以实 现对短消息内容的过滤; 在匹配成功的 ^犬态下, 判断出关键字匹配成功的短 消息为违规短消息, 并将发送违规短消息的用户直接添加入黑名单, 后续该 发送违规短消息的用户不能再发送短消息; 在匹配不成功的状态下, 判断出 关键字匹配不成功的短消息为模糊短消息并发送给接口单元, 以便后续通过 接口单元与仲裁单元之间的交互执行进一步的过滤, 并对模糊短消息定性, 从而正确地、有效地判断出模糊短消息是违规短消息还是正常短消息。其中, 所谓模糊短消息指: 现阶段无法正确地、 有效地判断出到底是违规短消息还 是正常短消息的可疑消息。 所涉及的违规短消息即为垃圾短消息 , 以下不作 赘述。 这里需要指出的是, 以上通过监控单元的流量规则进行过滤,以及通过 人机交互单元的关键字规则进行过滤都是基于现有监控规则进行的过滤。 尤 其, 这里的关键字规则是很筒单的, 区别于后续仲裁单元中采用的关键字规 贝1 J。 仲裁单元中采用的关键字规则是复杂的, 不仅包括范围广泛的关键字, 比如包括政治类、 广告类、 安全类中不同类别的关键字; 而且还包括各个不 同类别关键字之间复杂的逻辑运算关系。 针对以上本发明的系统组成结构而言, 监控单元具体包括: 监控处理模 块和监控管理模块。 其中, 监控处理模块, 用于从短信中心获取短消息并统 计流量个数, 将统计的短消息的流量个数与配置的流量规则匹配; 将流量匹 配成功的短消息发送给监控管理模块。 监控管理模块, 用于接收流量匹配成 功的短消息并转发给人机交互单元; 解析出流量匹配成功的短消息中的用户 信息并存储。 这里, 所谓用户信息指: 流量匹配成功的短消息中, 所封装的 所有与发送该短消息的用户有关的信息, 包括: 主、 被叫用户号码; 包含关 键字的短消息具体内容; 主、 被叫用户号码类型; 主、 被叫用户号段信息等。 人机交互单元, 进一步用于配置仲裁单元的仲裁规则并同步到仲裁单 元, 配置时可以采用命令行输入的方式进行配置, 也可以采用批处理文件命 令的方式进行配置, 以便仲裁单元能及时获取到同步更新后的仲裁规则, 根 据同步更新后的仲裁规则, 能及时地、 正确地、 有效地对模糊短消息进行仲 裁。 这里, 仲裁规则即为上述涉及的包括关键字规则、 用户号码类型信息规 则、 用户号段信息规则在内的多个类别的仲裁规则。 人机交互单元, 进一步 用于显示包括模糊短消息、 对模糊短消息仲裁后得到的判断结果在内的监控 信息, 以便对监控管理模块和接口单元返回人机交互单元的信息进行直观显 示, 及时调整人机交互单元中的配置信息, 有利于高效的完成监控。 仲裁单元, 进一步包括多个仲裁模块, 且各个仲裁模块中各自包含的仲 裁规则分别属于不同类别; 各个仲裁模块, 用于分别根据各自包含的仲裁规 则与模糊短消息匹配, 并将匹配后的各个仲裁结果返回接口单元。 此处, 各个仲裁模块, 进一步用于分别根据各个仲裁结果对模糊短消息 进行判断, 并获得通过各个仲裁模块针对模糊短消息的判断结果; 获取通过 接口单元针对模糊短消息的判断结果, 在设置的系统运行时间内, 在当前仲 裁模块与接口单元判断结果相同的 率大于第一期望值状态下, 增加当前仲 裁模块的仲裁权值; 在当前仲裁模块与接口单元判断结果相同的概率小于第 一期望值状态下, 减少当前仲裁模块的仲裁权值; 将修改后的仲裁权值返回 接口单元进行仲裁权值的更新。 相应地,接口单元, 进一步用于才艮据返回的更新后的仲裁权值进行加权 运算, 并继续执行针对模糊短消息的判断。 此处, 针对接口单元执行的加权运算而言, 接口单元进一步用于进行加 ^ is the arbitration weight of each arbitration module, where n is the number of arbitration modules. Preferably, in A, the determining, by each arbitration module, the fuzzy short message according to each arbitration result further comprises: when the arbitration result is greater than the set second threshold, determining that the fuzzy short message is a violation short message; wherein, the second valve The value is 1/n times the first threshold; n is the number of arbitration modules. Preferably, after C, the method further includes: increasing, in the set system running time, the current first threshold value when the interface unit determines that the probability of the violation short message is greater than the second expected value; and determining that the violation is short in the interface unit If the probability of the message is less than 1 time of the second expected value, the current first threshold is decreased; and the modified first threshold is returned to each arbitration module for updating. The arbitration unit of the present invention matches the fuzzy short message according to different types of arbitration rules, and returns the matched arbitration results to the interface unit; the interface unit performs weighting operation according to each of the matched arbitration results; when the operation result is smaller than the set When the first threshold is reached, it is determined that the fuzzy short message is a violation short message. The invention provides a processing scheme for the fuzzy short message, which is a good supplement to the current monitoring system 4, and can greatly improve the ability to process the fuzzy short message without affecting the function of the current monitoring system, and actively respond to the new garbage short. The message poses challenges to the current monitoring system and maximizes the real-time performance of the monitoring system. The invention overcomes the defects in the current monitoring system that the fuzzy short message cannot be judged or relies on manual judgment, and proposes a monitoring system that matches the fuzzy short message and automatically judges according to different types of arbitration rules, so as to analyze the garbage more comprehensively. The characteristics of short messages, and capture violations of short messages and violating users. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic structural diagram of a short message monitoring system according to an embodiment of the present invention; FIG. 2 is a schematic flowchart of an implementation of a short message monitoring method according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The basic idea of the present invention is: Matching fuzzy short messages according to different categories of arbitration rules, and automatically determining fuzzy short messages. The implementation of the technical solution will be further described in detail below with reference to the accompanying drawings. FIG. 1 is a schematic structural diagram of a short message monitoring system according to an embodiment of the present invention. As shown in FIG. 1, the short message monitoring system includes: an interface unit and an arbitration unit. The interface unit is configured to forward the fuzzy short message to the arbitration unit for arbitration; perform weighting operation according to the matched arbitration results returned by the arbitration unit; and when the operation result is less than the first threshold value set in the interface unit, The fuzzy short message is determined to be a violation short message; when the operation result is greater than the set first threshold value, the fuzzy short message is determined to be a normal short message. An arbitration unit, configured to obtain a fuzzy short message by forwarding by the interface unit; respectively matching the fuzzy short message according to different categories of arbitration rules, and returning each matched arbitration result to the interface unit, so that the interface unit matches the matched arbitration result After the weighting operation, the first threshold value set in the interface unit is compared and the fuzzy short message is characterized. The arbitration rules include: a keyword rule, a user number type information rule, and a user number segment information rule. The short message monitoring system may further include: a monitoring unit and a human-machine interaction unit. The monitoring unit and the human-machine interaction unit are both existing, and they are described herein. The monitoring unit is configured to obtain short messages and count the number of traffic, and the number of statistics of the short messages and the monitoring unit The configured traffic rules are matched to implement filtering of short message traffic; the short message with successful traffic matching is sent to the human-machine interaction unit, so that the human-computer interaction unit performs further filtering according to the keyword rule. The human-machine interaction unit is configured to receive a short message with successful traffic matching from the monitoring unit, and match the keyword of the short message with successful traffic matching with the keyword rule configured in the human-computer interaction unit to implement filtering of the short message content. In the matching dog state, the short message that the keyword is successfully matched is determined to be a violation short message, and the user who sends the violation short message is directly added to the blacklist, and the user who subsequently sends the violation short message can not send the short message again. In the state that the matching is unsuccessful, the short message that determines that the keyword match is unsuccessful is a fuzzy short message and is sent to the interface unit, so that further filtering is performed through the interaction between the interface unit and the arbitration unit, and the fuzzy is performed. The short message is qualitatively determined to correctly and effectively determine whether the fuzzy short message is a violation short message or a normal short message. Among them, the so-called fuzzy short message means: At this stage, it is impossible to correctly and effectively determine whether the short message is a suspicious message or a normal short message. The non-compliant short message involved is a spam short message, which will not be described below. It should be noted here that the above filtering by the traffic rules of the monitoring unit and the filtering by the keyword rules of the human-machine interaction unit are all filtering based on the existing monitoring rules. In particular, where the cartridge is a single rule keyword, keywords, rules different from shellfish subsequent arbitration unit employed in the 1 J. The keyword rules used in the arbitration unit are complex, including not only a wide range of keywords, such as keywords in different categories including political, advertising, and security; but also Complex logical operation relationships between categories of keywords. For the system component structure of the present invention, the monitoring unit specifically includes: a monitoring processing module and a monitoring management module. The monitoring processing module is configured to obtain a short message from the short message center and collect the number of the traffic, and match the number of the statistics of the short message with the configured traffic rule; and send the short message with the successfully matched traffic to the monitoring management module. The monitoring management module is configured to receive the short message with successful traffic matching and forward it to the human-machine interaction unit; and parse the user information in the short message with successful traffic matching and store it. Here, the so-called user information refers to: in the short message that the traffic matching succeeds, all the information encapsulated by the user who sends the short message includes: a primary and a called user number; a short message specific content including a keyword; Called user number type; primary and called user number segment information. The human-computer interaction unit is further configured to configure an arbitration rule of the arbitration unit and synchronize to the arbitration unit. The configuration may be configured by using a command line input manner, or may be configured by using a batch file command, so that the arbitration unit can obtain the time in time. Arbitration rules after synchronization update, according to the synchronously updated arbitration rules, can arbitrate fuzzy short messages in a timely, correct and effective manner. Here, the arbitration rule is the above-mentioned arbitration rules of a plurality of categories including a keyword rule, a user number type information rule, and a user number segment information rule. The human-machine interaction unit is further configured to display monitoring information including a fuzzy short message and a judgment result obtained after the fuzzy short message is arbitrated, so as to visually display the information of the monitoring and management module and the interface unit returning the human-computer interaction unit, in time Adjusting the configuration information in the human-computer interaction unit is beneficial for efficient monitoring. The arbitration unit further includes a plurality of arbitration modules, and each of the arbitration rules included in each arbitration module belongs to a different category; each arbitration module is configured to match the fuzzy short message according to the respective included arbitration rules, and each of the matched arbitration messages The arbitration result is returned to the interface unit. Here, each arbitration module is further configured to determine the fuzzy short message according to each arbitration result, and obtain a judgment result for the fuzzy short message by each arbitration module; obtain the judgment result of the fuzzy short message by the interface unit, and set During the system running time, when the current arbitration module and the interface unit determine that the same rate is greater than the first expected value, the arbitration weight of the current arbitration module is increased; the probability that the current arbitration module and the interface unit determine the same result is less than the first expected value. In the state, the arbitration weight of the current arbitration module is reduced; the modified arbitration weight is returned to the interface unit to update the arbitration weight. Correspondingly, the interface unit is further configured to perform a weighting operation according to the returned updated arbitration weight value, and continue to perform the determination for the fuzzy short message. Here, for the weighting operation performed by the interface unit, the interface unit is further used for adding
^Result. x<w. ^Result. x<w.
权运算时所采用的运算公式为: Result w ; 其中, Result 为运 算结果, Result<为各个仲裁结果, 为各个仲裁模块的仲裁权值, n 为仲裁 模块的个数。 这里 , 针对各个仲裁模块根据各个仲裁结果对模糊短消息进行判断而 言, 各个仲裁模块进一步用于在仲裁结果大于设置的第二阀值状态下, 判断 出模糊短消息为违规短消息; 其中, 第二阀值为第一阀值的 1/n倍。 其中, 各个仲裁结果即为 ReSult'—。 这里, 针对仲裁模块的类型而言, 仲裁模块具体包括: 包含关键字规则 的仲裁模块、 包含用户号码类型信息规则的仲裁模块、 包含用户号段信息规 则的仲裁模块。 其中, 关键字规则包括: 政治类、 广告类、 安全类中至少一 种类别的关键字规则 , 也就是说, 关键字规则既可以是这些类别中任一个单 一类别的关键字规则, 比如政治类的关键字规则; 又可以是这些类别中组合 类别的关键字规则 , 比如政治类加安全类这一组合类别的关键字规则 , 使用 这种组合类别的关键字规则能表达复杂的逻辑运算关系, 从而能更好地实现 对模糊短消息的仲裁。 用户号码类型信息规则包括: 主、 被叫用户号码类型 信息规则。 用户号段信息规则包括: 主、 被叫用户号段信息规则。 这里,针对接口单元根据更新后的仲裁权值继续执行针对模糊短消息的 判断而言, 接口单元进一步用于在设置的系统运行时间内 , 在接口单元判断 出违规短消息的概率大于第二期望值 1倍的状态下, 增加当前第一阀值; 在 接口单元判断出违规短消息的概率小于第二期望值 1倍的状态下, 减少当前 第一阀值。 将修改后的第一阀值返回各个仲裁模块进行更新; 通过与各个仲 裁模块的交互, 并更新仲裁权值、 第一阀值、 和实现接口单元针对模糊短消 息的判断结果在第二期望值的上、 下 20%范围内收敛的方式, 完成模糊短消 息为违规短消息的判断。 综上 , 才艮据本发明的系统的核心部件是仲裁单元 , 在仲裁单元包括至少 一个仲裁模块情况下的仲裁模块, 以及接口单元。 针对仲裁模块而言, 仲裁 模块对模糊短消息是否为违规短消息的判断是: 基于配置的关键字内容, 以 及模糊短消息中含有的关键字之间复杂逻辑关系的判断, 并且加入了对模糊 短消息主、 被叫用户的状态判断等来综合的考量, 并通过至少一个仲裁模块 与接口单元的交互自动实现针对模糊短消息的判断。 进一步地, 采用本发明, 通过至少一个仲裁模块与接口单元的交互进行 仲裁权值的学习 , 根据设置的系统运行时间系统运行一段时间后 , 最终可以 得到结果最优的仲裁器和仲裁权值的组合。 也就是说, 通过至少一个仲裁模 块与接口单元的交互, 并更新仲裁权值、 第一阀值、 和实现接口单元针对模 糊短消息的判断结果在第二期望值的上、 下 20%范围内收敛的方式, 完成模 糊短消息为违规短消息的判断。 图 2为根据本发明实施例的短消息监控方法的实现流程示意图。 如图 2 所示, 该短消息监控方法包括以下步骤: 步骤 S101、 接口单元将模糊短消息转发给仲裁单元进行仲裁, 仲裁单 元根据不同类别的仲裁规则分别与模糊短消息匹配, 并将匹配后的各个仲裁 结果返回接口单元。 这里, 仲裁规则包括: 关键字规则、 用户号码类型信息规则、 用户号段 信息规则。 其中, 关键字规则包括: 政治类、 广告类、 安全类中至少一种类 别的关键字规则; 用户号码类型信息规则包括: 主、 被叫用户号码类型信息 规则; 用户号段信息规则包括: 主、 被叫用户号段信息规则。 步骤 S102、 接口单元才艮据仲裁单元返回的匹配后的各个仲裁结果进行 加权运算; 在运算结果小于设置的第一阀值时, 判断出模糊短消息为违规短 消息。 这里 , 接口 单元进行加权运算时所采用 的运算公式为 : ^Result. x<w. The operation formula used in the weight operation is: Result w ; where Result is the result of the operation, Result < is the result of each arbitration, and is the arbitration weight of each arbitration module, where n is the number of arbitration modules. Here, for each of the arbitration modules to determine the fuzzy short message according to each arbitration result, each arbitration module is further configured to determine that the fuzzy short message is a violation short message when the arbitration result is greater than the set second threshold value; The second threshold is 1/n times the first threshold. Among them, the result of each arbitration is ReSult '-. Here, for the type of the arbitration module, the arbitration module specifically includes: an arbitration module including a keyword rule, an arbitration module including a user number type information rule, and an arbitration module including a user number segment information rule. The keyword rules include: keyword rules of at least one of the political class, the advertising class, and the security class, that is, the keyword rule can be a keyword rule of any one of the categories, such as a political class. Keyword rules; or keyword rules of combined categories in these categories, such as the keyword rules of the political category plus the security category, the keyword rules using this combination category can express complex logical operation relationships, Therefore, the arbitration of the fuzzy short message can be better realized. The user number type information rule includes: a primary and a called user number type information rule. The user number segment information rule includes: a primary and a called user number segment information rule. Here, for the interface unit to continue to perform the judgment for the fuzzy short message according to the updated arbitration weight value, the interface unit is further configured to determine, in the set system running time, that the probability of the violation short message is greater than the second expected value in the interface unit. In the state of 1 time, the current first threshold is increased; and when the interface unit determines that the probability of the violation short message is less than 1 time of the second expected value, the current first threshold is decreased. Returning the modified first threshold to each arbitration module for updating; by interacting with each arbitration module, and updating the arbitration weight, the first threshold, and implementing the judgment result of the interface unit for the fuzzy short message at the second expected value In the way of convergence in the upper and lower 20% range, the completion of the fuzzy short message is the judgment of the violation short message. In summary, the core component of the system according to the present invention is an arbitration unit, an arbitration module in the case where the arbitration unit includes at least one arbitration module, and an interface unit. For the arbitration module, the arbitration module judges whether the fuzzy short message is a violation short message: based on the configured keyword content, and the judgment of the complex logical relationship between the keywords contained in the fuzzy short message, and adds the fuzzy The short message main and the called user's state judgment are comprehensively considered, and the judgment for the fuzzy short message is automatically realized by the interaction of at least one arbitration module and the interface unit. Further, according to the present invention, the arbitration weight is learned through the interaction between the at least one arbitration module and the interface unit, and after the system runs for a period of time according to the set system running time, the optimal arbiter and the arbitration weight can be finally obtained. combination. That is, the interaction between the at least one arbitration module and the interface unit, and updating the arbitration weight, the first threshold, and the implementation of the interface unit for the fuzzy short message converge within the upper and lower 20% of the second expected value. The way to complete the fuzzy short message is to judge the violation of the short message. FIG. 2 is a schematic diagram of an implementation process of a short message monitoring method according to an embodiment of the present invention. As shown in FIG. 2, the short message monitoring method includes the following steps: Step S101: The interface unit forwards the fuzzy short message to the arbitration unit for arbitration, and the arbitration unit respectively matches the fuzzy short message according to different types of arbitration rules, and The individual arbitration results are returned to the interface unit. Here, the arbitration rules include: a keyword rule, a user number type information rule, and a user number segment information rule. The keyword rule includes: a keyword rule of at least one of a political class, an advertisement class, and a security class; the user number type information rule includes: a primary and a called user number type information rule; the user number segment information rule includes: The called user number segment information rule. Step S102: The interface unit performs a weighting operation according to the matched arbitration results returned by the arbitration unit. When the operation result is less than the set first threshold, it is determined that the fuzzy short message is a violation short message. Here, the operation formula used by the interface unit to perform the weighting operation is: ^Result. x<w.
Result= '=i ; 其中, Result 为运算结果, Result'为各个仲裁结果, Result= '=i ; where Result is the result of the operation and Result ' is the result of each arbitration.
^为各个仲裁模块的仲裁权值 , n为仲裁模块的个数。 针对由以上步骤 S101〜步骤 S102 所构成的技术方案而言, 步骤 S101 之前还可以包括: 监控单元获取短消息并统计流量个数, 夺统计的短消息的 流量个数与配置的流量规则匹配; 将流量匹配成功的短消息发送给人机交互 单元。 之后, 人机交互单元将流量匹配成功的短消息的关键字与配置的关键 字规则匹配; 将关键字匹配成功的短消息判断为违规短消息; 将关键字匹配 不成功的短消息判断为模糊短消息并发送给接口单元。 这里, 人机交互单元进行关键字规则匹配之前还可以包括: 初始配置仲 裁规则并同步到仲裁单元。 这里, 当仲裁规则的配置 4爹丈后, 人机交互单元 将爹改后的仲裁规则同步更新到仲裁单元。 由于人机交互单元可以显示包括 模糊短消息、 对模糊短消息仲裁后得到的判断结果在内的监控信息, 因此, 在显示监控信息后, 可以根据显示的监控信息有选择的修改仲裁规则的配置 并同步更新到仲裁单元。 这里需要指出的是, 当仲裁单元包括多个仲裁模块, 且各个仲裁模块中 各自包含的仲裁规则分别属于不同类别情况下, 步骤 S101 的处理过程进一 步包括: 各个仲裁模块分别根据各自包含的仲裁规则与模糊短消息匹配, 并 将匹配后的各个仲裁结果返回接口单元。 这里, 步骤 S 102之后, 还可以包括: 步骤 S103、 各个仲裁模块分别根据各个仲裁结果对模糊短消息进行判 断, 并获得通过各个仲裁模块针对模糊短消息的判断结果。 其中,各个仲裁模块分别根据各个仲裁结果对模糊短消息进行判断的具 体处理过程包括: 当仲裁结果大于设置的第二阀值时, 判断出模糊短消息为 违规短消息; 其中, 第二阀值为第一阀值的 1/n倍, 即第二阀值 =1/η χ第一 阀值, 且 η为仲裁模块的个数。 步骤 S 104、 各个仲裁模块获取通过接口单元针对模糊短消息的判断结 果。 步骤 S105、 在设置的系统运行时间内, 在当前仲裁模块与接口单元判 断结果相同的概率大于第一期望值情况下 , 增加当前仲裁模块的仲裁权值; 在当前仲裁模块与接口单元判断结果相同的概率小于第一期望值情况下 , 减 少当前仲裁模块的仲裁权值; 将 4爹改后的仲裁权值返回至接口单元进行仲裁 权值的更新。 步骤 S106、 接口单元根据返回的更新后的仲裁权值进行力。权运算, 并 继续执行针对模糊短消息的判断。 这里, 步骤 S 106之后, 还可以包括: 步骤 S107、 在设置的系统运行时间内, 在接口单元判断出违规短消息 的¾¾率大于第二期望值 1倍的情况下, 增加当前第一阀值; 在接口单元判断 出违规短消息的概率小于第二期望值 1倍的情况下, 减少当前第一阀值。 步骤 S 108、接口单元将修改后的第一阀值返回各个仲裁模块进行更新; 通过与各个仲裁模块的交互, 并更新仲裁权值、 第一阀值、 和实现接口单元 针对模糊短消息的判断结果在第二期望值的上、 下 20%范围内收敛的方式, 完成模糊短消息为违规短消息的判断。 方法实施例一为: 监控单元包括监控处理模块和监控管理模块 , 仲裁单 元包括多个仲裁模块情况下, 无需仲裁模块与接口单元的多次交互进行仲裁 权值的学习及更新,本方法实施例中, 实现短消息的监控流程包括以下步骤: 步骤 S201、 监控处理模块从短消息中心接收用户发送的短消息。 这里, 短消息中封装有主、 被叫用户号码; 包含关键字的短消息具体内 容; 主、 被叫用户号码类型; 主、 被叫用户号段信息等信息。 步骤 S202、 监控处理模块对接收到的短消息进行统计计数, 并与系统 配置的流量规则进行匹配 , 将流量匹配成功的短消息发送给监控管理模块 , 监控管理模块将流量匹配成功的短消息导入数据库存储, 并将流量匹配成功 的短消息发送给人机交互单元。 这里, 人机交互单元也可以称为控制台, 流 量匹配成功的短消息即为违规短消息, 其所对应的用户为违规用户。 步骤 S203、 人机交互单元显示违规用户, 并对流量匹配成功的短消息 的内容进行过滤, 将流量匹配成功的短消息的关键字与人机交互单元中所配 置的关键字规则匹配, 以实现对短消息内容的过滤; 在匹配成功的^!犬态下, 判断出关键字匹配成功的短消息为违规短消息; 匹配不成功的状态下 , 判断 出关键字匹配不成功的短消息为模糊短消息并发送给接口单元。 这里 , 人机交互单元中所配置的关键字规则与现有的关键字规则相同 , 也可以称为特定关键字规则。 对含有特定关键字的违规短消息所对应的违规 用户直接加入黑名单, 后续该违规用户不能再发送短消息, 对不含有特定关 键字即主要包括政治类 , 安全类关键字的模糊短消息发送至接口单元处理。 步骤 S204、 接口单元将需要仲裁的模糊短消息分别发送给 n个仲裁模 块, 并配置好各个仲裁模块初始的仲裁权值, 仲裁模块超时处理, 第一阀值 等信息后, 等待各个仲裁模块的响应。 步骤 S205、 各个仲裁模块对需仲裁的模糊短消息进行仲裁。 步骤 S206、 各个仲裁器将仲裁结果返回至接口单元。 步骤 S207、接口单元根据各个仲裁模块返回的仲裁结果进行加权运算, 与设置好的第一阀值进行比较及判断, 并对模糊短消息进行定性。 这里 , 所谓对模糊短消息进行定性指: 判断出模糊短消息的性质到底是 违规短消息还是正常短消息。 步骤 S208、 结束当前短消息监控流程。 方法实施例二为: 仲裁单元包括多个仲裁模块情况下, 需要仲裁模块与 接口单元的多次交互进行仲裁权值的学习及更新, 本方法实施例中, 实现短 消息的监控流程包括以下步骤: 步骤 S301、 在短消息监控的人机交互单元完成系统基本属性配置、 监 控规则配置和仲裁模块的相关配置等 , 并进行监控规则的同步。 这里, 举例来说, 比如配置相关的流量和关键字规则; 比如配置仲裁模 块的号码, 可以初始化为 3个; 配置各个仲裁模块的仲裁权值, 可以初始化 仲裁权值都为 1 ; 配置链路信息; 配置超时处理信息, 可以初始化为 3秒, 以判断发送短消息的用户是否为违规用户。 步骤 S302、 各个仲裁模块导入针对仲裁模块的相关配置。 这里, 导入的仲裁模块的相关配置中包括为仲裁模块配置的监控规则, 监控规则包括关键字规则、 用户类型信息规则、 号段信息规则在内的至少一 个类别的仲裁规则。 针对关键字规则而言, 这里仲裁模块中所配置的关键字 规则与人机交互单元中人机交互单元中所配置的关键字规则不同 , 仲裁模块 中所配置的关键字规则, 其信息量很大, 关键字基本包含: 所有的政治类、 广告类和安全类违规字符以及它们之间复杂的逻辑运算关系。 针对用户类型 信息规则而言, 用户类型信息包含: 所有的黑名单和白名单信息, 在线计费 系统( OCS ) 欠费用户信息, 以及与营帐接口的数据包格式协议( H2 )属性 信息等用户类型信息。 针对号段信息规则而言, 号段信息包含: 所有的预付 费号段, 本省号段, 本网号段, 特殊用户号段即白名单号段等。 其中, 用户 类型信息规则主、 被叫用户号码类型信息规则; 用户号段信息规则包括: 主、 被叫用户号段信息规则。 步骤 S303、 监控单元接收短消息中心的短消息, 对短消息进行流量计 数并与系统配置的流量规则进行匹配; 将流量匹配成功的短消息确定为违规 短消息 , 将达到该流量监控规则的违规短消息对应的违规用户通过人机交互 单元进行显示。 步骤 S304、 人机交互单元对违规短消息的内容进行过滤, 也就是说, 将流量匹配成功的短消息的关键字与人机交互单元中所配置的关键字规则匹 配, 以实现对短消息内容的过滤。 步骤 S305、 过滤后产生的模糊短消息通过接口单元发送给各个仲裁模 块进行仲裁处理, 每个仲裁模块对需仲裁的模糊短消息进行仲裁。 这里, 针对仲裁模块而言, 初始化每个仲裁模块的仲裁结果为 20 , 则 仲裁模块可以分为以下五类, 如下所示: 第一类仲裁模块为: 包含关键字规则的仲裁模块 , 用于根据关键字规则 对模糊短消息进行分类及仲裁, 不同关键字根据其重要性不同划分为几个等 级, 分别为 10分, 8分, 5分, 2分四个等级。 其中, 政治类: 如法轮功、 天安门事件等配置为 1类关键字; 安全类: 如自焚、 静坐、 手枪等配置为 2 类关键字; 色情类: 如成人、 情色等配置为 3类关键字; 广告类: 如卫星安 装、 发票等 配置为 4 类关键字, 并对各类逻辑关系关键字进行组合配置, 如反对&共产党, 色情 &卫星安装, 政府& (静坐 I游行) 等组合配置为高级 别的关键字。 对模糊短消息的内容进行所有关键字的匹配,对匹配到的关键 字进行统计, 对匹配得到的关键字进行扣分的仲裁处理, 分数扣到 0以后作 为 0分处理。 第二类仲裁模块为:包含主叫用户号码类型信息规则的主叫用户类型仲 裁模块, 用于根据主叫用户号码类型信息规则对主叫用户进行号码分类及仲 裁。 号码分为以下几个等级: 白名单用户加 10分、 黑名单用户减 20分、 OCS 欠费用户扣 10分、 H2属性为钻石卡用户的加 10分、 金卡用户加 5、 银卡用 户力。 3分、 普通卡用户不加分, 对主被叫用户进行匹配后得到仲裁结果, 分 数扣到 0以后作为 0分处理。 其中, 钻石卡用户、 金卡用户、 4艮卡用户、 普 通卡用户是运营商在营帐系统中对用户所作的不同分级。 第三类仲裁模块为:包含被叫用户号码类型信息规则的被叫用户类型仲 裁模块, 用于根据被叫用户号码类型信息规则对被叫用户进行号码分类及仲 裁。 号码分为以下几个等级: 白名单用户加 10分、 黑名单用户减 20分、 OCS 欠费用户扣 10分、 H2属性为钻石卡用户的加 10分、 金卡用户加 5、 银卡用 户力。 3分、 普通卡不加分, 对被叫用户进行匹配后得到仲裁结果, 分数扣到 0以后作为 0分处理。 第四类仲裁模块为:包含主叫用户号段信息规则的主叫号段信息仲裁模 块, 用于根据主叫用户号段信息规则对主叫用户进行号段分类及仲裁。 号段 分为以下几个等级: VIP集团号段加 10分、 全球通号段加 5分、 梦网网关加 5分、 行业网关不加分、 预付费号段不加分、 外省号段减 2分、 外网号段减 5分等, 对主被叫用户进行匹配后得到仲裁结果, 分数扣到 0以后作为 0分 处理。 上述这些号段信息全部是运营商配置的。 其中, VIP集团号段是运营 商对大的集团用户所提供的号段, 号段内用户拨打电话资费低, 且可以直接 拨打短号; 梦网网关是运营商直接经营的服务供应商; 行业网关是非运营商 经营的服务供应商; 预付费号段是运营商设置的号段, 号段中所有号码全部 为预付费用户。 第五类仲裁模块为:包含被叫用户号段信息规则的被叫号段信息仲裁模 块, 用于根据被叫用户号段信息规则对被叫用户进行号段分类及仲裁。 号段 分为以下几个等级: VIP集团号段加 10分、 全球通号段加 5分、 梦网网关加 5分、 行业网关不加分、 预付费号段不加分、 外省号段减 2分、 外网号段减 5分等, 对被被叫用户进行匹配后得到仲裁结果, 分数扣到 0以后作为 0分 处理。 步骤 S306、 各个仲裁模块将仲裁结果发送至接口单元, 接口单元对仲 裁结果进行加权运算。 在包括以上五类仲裁模块 , 仲裁模块个数为 5情况下 , 加权运算采用的 公式为: Result=Resulti X ω 1+Result2 χ ω 2+Result3 χ ω 3+Result4 χ ω 4+Result5 χ ω 5 , 其中, 各个仲裁器的仲裁权值即 0^、 ω 2、 ω 3. ω 4、 ω 5皆初始化为 1 , 运算结果与设置的第一阈值进行比较, 这里的第一阈值可以初始化为 80。 当 运算结果小于设置的第一阀值时 , 判断出模糊短消息为违规短消息 , 否则为 正常短消息。 步骤 S307、 接口单元将运算结果返回至各个仲裁模块, 仲裁模块首先 将运算结果与设置的第二阀值比较 , 这里的各个仲裁模块的第二阀值是相同 的, 皆为接口单元设置的第一阈值的 1/5倍。 当运算结果小于设置的第二阀 值时, 判断出模糊短消息为正常短消息; 当运算结果大于设置的第二阀值时, 判断出模糊短消息为违规短消息。 步骤 S308、 比较通过各个仲裁模块针对模糊短消息的判断结果与通过 接口单元针对模糊短消息的判断结果是否相同 , 如果相同 , 则确定通过仲裁 模块针对模糊短消息的判断结果正确 , 否则确定为错误。 步骤 S309、 仲裁模块运行一段时间后进行权值学习, 这段时间是指默 认设置的系统运行时间, 可以为 1小时, 将计算出的这段时间内当前仲裁模 块判断正确的 4既率与第一期望值比较, 这里的第一期望值可以设置为 50%; 如果当前仲裁模块判断正确的概率大于 50% , 则增加当前仲裁模块的仲裁权 值, 比如将下一阶段的仲裁权值增加 0.1 ; 如果当前仲裁模块判断正确的相无 率小于 50%, 则减少当前仲裁模块的仲裁权值, 比如将下一阶段的仲裁权值 减少 0.1 ; 将修改后的仲裁权值发送至接口单元, 接口机进行仲裁权值的更 新。 这里, 计算当前仲裁模块判断正确的 ^既率具体为: 当前仲裁模块判断正 确的概率 =通过各个仲裁模块针对模糊短消息的判断结果正确的消息 /判断的 消息总量。 步骤 S310、 接口机运行一段时间后, 这段时间是指默认设置的系统运 行时间, 可以为 1小时, 将计算出的这段时间内违规短消息的概率与第二期 望值比较, 这里的第二期望值可以初始设置为 40%, 但不大于 50%; 如果违 规短消息的概率小于第二期望值 1倍的情况下 , 则减少当前第一阀值, 比如 将第一阀值减小 5; 如果违规短消息的¾¾率大于第二期望值 1倍的情况下, 则增加当前第一阀值, 比如将第一阀值增加 5; 将 4爹改后的第一阀值发送至 各个仲裁模块, 各个仲裁模块进行第一阀值的更新。 步骤 S311、 每天将 24小时内违规短消息占总消息的百分比的结果写成 文件, 该文件即为包含接口单元针对模糊短消息的判断结果的文件, 观察判 断结果是否收敛在一个范围内, 如果收敛的范围很小,且在第二期望值周围, 比如在第二期望值的上、下 20%范围内收敛,则认为系统的仲裁权值和第一、 第二阈值训练结果正常; 如果结果发散, 则修改第二期望值后再进行训练, 一天后再观察判断结果并进行相应的处理。 综上所述, 采用本发明, 在系统运行初期需要耗费一定时间进行训练, 但是在训练出最优的期望值和仲裁权值后 , 系统对模糊短消息的自动仲裁结 果相比较于现有的人工判断有了很大的提高 , 包括判断的准确性和自动化程 度。 对现有的短消息监控系统是一个很好的补充, 同时可以分析用户发送垃 圾短消息的一些行为习惯, 对改进现有系统提供了一些经 -险。 由于现有的短 消息监控系统只能基于流量规则或筒单的关键字规则, 即上面所涉及到的人 机交互单元中所配置的关键字规则进行监控, 因此, 艮容易使监控失效, 监 控出违规短消息越来越少。 如果一味降低流量的门限值会导致监控出来的违 规短消息呈数量级增加, 人工操作维护的工作量将会很大。 新的违规短消息 的内容层出不穷,仅仅增加有限的、个别的关键字必然无法满足监控的需求, 特别是一些新的内容违规短消息的发送模式艮难仅仅配置关键字进行监控。 本发明正是增加了关键字之间的复杂运算关系后大大改善了系统的性能; 而 且本发明的监控系统是基于仲裁模块与接口单元之间的交互来通过机器自动 地学习仲裁权值 , 从而能得到可靠的仲裁结果 , 对现有的监控系统是一个很 大的改进。 另外, 本发明对违规短消息进行分类及仲裁, 对发送数量仅仅达 到流量规则 , 消息内容中不含有特定关键字的短消息即模糊短消息进行仲裁 处理, 由仲裁模块自动地对这类模糊短消息的消息内容, 主、 被叫号码类型, 主、 被叫号段类型等信息进行全面分析, 最终对这类模糊短消息是否为垃圾 短消息进行判断 , 采用本发明 , 比现有监控系统采用单一维护人员进行人工 的、 主观的判断方案, 准确性和效率都有 4艮大的提高。 显然, 本领域的技术人员应该明白, 上述的本发明的各模块或各步骤可 以用通用的计算装置来实现, 它们可以集中在单个的计算装置上, 或者分布 在多个计算装置所组成的网络上, 可选地, 它们可以用计算装置可执行的程 序代码来实现, 从而, 可以将它们存储在存储装置中由计算装置来执行, 或 者将它们分别制作成各个集成电路模块, 或者将它们中的多个模块或步骤制 作成单个集成电路模块来实现。 这样, 本发明不限制于任何特定的硬件和软 件结合。 以上所述, 仅为本发明的较佳实施例而已, 并非用于限定本发明的保护 范围。 对于本领域的技术人员来说, 本发明可以有各种更改和变化。 凡在本 发明的精神和原则之内, 所作的任何修改、 等同替换、 改进等, 均应包含在 本发明的保护范围之内。 ^ is the arbitration weight of each arbitration module, where n is the number of arbitration modules. For the technical solution of the above steps S101 to S102, the step S101 may further include: the monitoring unit acquires the short message and counts the number of the traffic, and the number of the traffic of the short message is matched with the configured traffic rule; The short message with successful traffic matching is sent to the human interaction unit. After that, the human interaction unit matches the keyword of the short message with successful traffic matching with the configured keyword rule; the short message with the keyword matching success is determined as the violation short message; The unsuccessful short message is judged as a fuzzy short message and sent to the interface unit. Here, before the human-computer interaction unit performs keyword rule matching, the method may further include: initially configuring the arbitration rule and synchronizing to the arbitration unit. Here, after the configuration of the arbitration rule is 4, the human interaction unit synchronously updates the arbitrated arbitration rule to the arbitration unit. The human-computer interaction unit can display the monitoring information including the fuzzy short message and the judgment result obtained after the fuzzy short message is arbitrated. Therefore, after the monitoring information is displayed, the configuration of the arbitration rule can be selectively modified according to the displayed monitoring information. And update to the arbitration unit synchronously. It should be noted that, when the arbitration unit includes multiple arbitration modules, and the arbitration rules respectively included in the respective arbitration modules belong to different categories, the processing of step S101 further includes: each arbitration module separately according to the arbitration rules respectively included Matches with the fuzzy short message, and returns the matched arbitration results to the interface unit. Here, after step S102, the method further includes: Step S103: Each arbitration module determines the fuzzy short message according to each arbitration result, and obtains a determination result for the fuzzy short message by each arbitration module. The specific processing process for determining, by each arbitration module, the fuzzy short message according to each arbitration result includes: when the arbitration result is greater than the set second threshold, determining that the fuzzy short message is a violation short message; wherein, the second threshold It is 1/n times the first threshold, that is, the second threshold = 1 / η χ the first threshold, and η is the number of arbitration modules. Step S104: Each arbitration module acquires a determination result of the fuzzy short message by the interface unit. Step S105: In the set system running time, when the probability that the current arbitration module and the interface unit determine the same result is greater than the first expected value, increase the arbitration weight of the current arbitration module; and the current arbitration module and the interface unit determine the same result. When the probability is less than the first expected value, the arbitration weight of the current arbitration module is reduced; and the 4 modified arbitration weight is returned to the interface unit for updating the arbitration weight. Step S106: The interface unit performs a force according to the returned updated arbitration weight value. The weight operation, and continue to perform the judgment for the fuzzy short message. Here, after step S106, the method may further include: Step S107: In the set system running time, when the interface unit determines that the 3⁄4⁄4 rate of the violation short message is greater than the second expected value, the current first threshold is increased; and the probability of the violation short message is determined by the interface unit is less than In the case where the second expected value is 1 time, the current first threshold value is decreased. Step S108: The interface unit returns the modified first threshold value to each arbitration module for updating; and interacts with each arbitration module, and updates the arbitration weight, the first threshold, and the interface unit to determine the fuzzy short message. As a result, the manner in which the fuzzy short message is completed is a violation of the short message in a manner of convergence in the upper and lower 20% of the second expected value. The first embodiment of the method is as follows: The monitoring unit includes a monitoring processing module and a monitoring management module. When the arbitration unit includes multiple arbitration modules, the arbitration function and the interface unit do not need to perform multiple interactions of the arbitration module to learn and update the arbitration weight. The monitoring process for implementing the short message includes the following steps: Step S201: The monitoring processing module receives the short message sent by the user from the short message center. Here, the short message encapsulates the primary and called user numbers; the short message specific content including the keyword; the primary and the called user number type; the primary and the called user number segment information and the like. Step S202: The monitoring processing module performs statistical counting on the received short message, and matches the traffic rule configured by the system, and sends a short message with successful traffic matching to the monitoring management module, and the monitoring management module imports the short message with successful traffic matching. The database stores and sends a short message with successful traffic matching to the human interaction unit. Here, the human-computer interaction unit may also be referred to as a console, and the short message with successful traffic matching is a violation short message, and the corresponding user is a violation user. Step S203: The human-machine interaction unit displays the offending user, and filters the content of the short message with the matching of the traffic, and matches the keyword of the short message with the matching of the traffic to the keyword rule configured in the human-computer interaction unit. Filtering the content of the short message; In the ^^ dog state with successful matching, the short message that the keyword is successfully matched is determined to be a violation short message; if the matching is unsuccessful, the short message with the unsuccessful keyword matching is determined to be fuzzy The short message is sent to the interface unit. Here, the keyword rules configured in the human-computer interaction unit are the same as the existing keyword rules, and may also be referred to as specific keyword rules. The violating user corresponding to the violation short message containing the specific keyword is directly added to the blacklist, and the subsequent violating user can no longer send the short message, and the fuzzy short message that does not contain the specific keyword, mainly including the political category and the security keyword, is sent. To the interface unit processing. Step S204: The interface unit sends the fuzzy short message that needs to be arbitrated to the n arbitration modules, and configures the initial arbitration weight of each arbitration module, and the arbitration module times out, the first threshold After waiting for the information, wait for the response of each arbitration module. Step S205: Each arbitration module arbitrates the fuzzy short message to be arbitrated. Step S206: Each arbitrator returns the arbitration result to the interface unit. Step S207: The interface unit performs a weighting operation according to the arbitration result returned by each arbitration module, compares and determines with the set first threshold, and characterizes the fuzzy short message. Here, the so-called qualitative short message refers to: Determine whether the nature of the fuzzy short message is a violation short message or a normal short message. Step S208, ending the current short message monitoring process. The second embodiment of the method is as follows: In the case that the arbitration unit includes multiple arbitration modules, the interaction between the arbitration module and the interface unit is required to learn and update the arbitration weight. In the embodiment of the method, the monitoring process of the short message includes the following steps. Step S301: The human-machine interaction unit of the short message monitoring completes the basic attribute configuration of the system, the configuration of the monitoring rule, and the related configuration of the arbitration module, and synchronizes the monitoring rules. Here, for example, configuration related traffic and keyword rules; for example, the number of the arbitration module can be initialized to three; the arbitration weight of each arbitration module can be configured, and the arbitration weight can be initialized to 1; Information; Configure timeout processing information, which can be initialized to 3 seconds to determine whether the user who sent the short message is a violating user. Step S302: Each arbitration module imports a related configuration for the arbitration module. Here, the related configuration of the imported arbitration module includes a monitoring rule configured for the arbitration module, and the monitoring rule includes at least one category of arbitration rules including a keyword rule, a user type information rule, and a segment information rule. For the keyword rule, the keyword rule configured in the arbitration module is different from the keyword rule configured in the human-computer interaction unit in the human-computer interaction unit, and the keyword rule configured in the arbitration module has a very large amount of information. Large, keywords basically include: all political, advertising, and security-type violation characters and complex logical operations between them. For the user type information rule, the user type information includes: all blacklist and whitelist information, online charging system (OCS) arrears user information, and packet format protocol (H2) attribute information with the camp interface. Type information. For the segment information rule, the segment information includes: all prepaid number segments, the province number segment, the network segment segment, and the special user segment segment, that is, the whitelist segment segment. Among them, the user Type information rule master and called user number type information rules; User number segment information rules include: primary and called user number segment information rules. Step S303: The monitoring unit receives the short message of the short message center, performs traffic counting on the short message, and matches the traffic rule configured by the system. The short message with the successfully matched traffic is determined as the violation short message, and the violation of the traffic monitoring rule is reached. The violating user corresponding to the short message is displayed by the human-computer interaction unit. Step S304: The human-machine interaction unit filters the content of the violation short message, that is, matches the keyword of the short message with successful traffic matching with the keyword rule configured in the human-computer interaction unit to implement the short message content. Filtering. Step S305: The fuzzy short message generated after filtering is sent to each arbitration module through an interface unit for arbitration, and each arbitration module arbitrates the fuzzy short message to be arbitrated. Here, for the arbitration module, the arbitration result of initializing each arbitration module is 20, and the arbitration module can be divided into the following five categories, as follows: The first type of arbitration module is: an arbitration module containing a keyword rule, used for The fuzzy short messages are classified and arbitrated according to the keyword rules. Different keywords are divided into several levels according to their importance, which are 10 points, 8 points, 5 points, 2 points and 4 levels. Among them, political categories: such as Falun Gong, Tiananmen incidents, etc. are configured as type 1 keywords; security categories: such as self-immolation, sit-down, pistol, etc. are configured as type 2 keywords; pornography: such as adult, erotic, etc. are configured as three types of keywords; Advertising categories: such as satellite installation, invoices, etc. are configured as 4 types of keywords, and combined configuration of various logical relationship keywords, such as opposition & Communist, porn & satellite installation, government & (quiet sit I parade), etc. Level of keywords. All the keywords are matched to the content of the fuzzy short message, the matched keywords are counted, and the matching keywords are deducted for arbitration, and the score is deducted to 0 and processed as 0. The second type of arbitration module is: a calling user type arbitration module including a calling party number type information rule, configured to perform number classification and arbitration on the calling user according to the calling party number type information rule. The number is divided into the following levels: whitelist users plus 10 points, blacklist users minus 20 points, OCS arrears users deduct 10 points, H2 attributes are diamond card users plus 10 points, gold card users plus 5, silver card users force. 3 points, ordinary card users do not add points, the main and called users are matched to get the arbitration result, the score is deducted to 0 and then treated as 0 points. Among them, the diamond card user, the gold card user, the 4 Leica user, and the ordinary card user are different ratings of the operator in the camping system. The third type of arbitration module is: a called user type arbitration module that includes a called user number type information rule, configured to perform number classification and arbitration on the called user according to the called user number type information rule. The number is divided into the following levels: whitelist users plus 10 points, blacklist users minus 20 points, OCS arrears users deduct 10 points, H2 attributes are diamond card users plus 10 points, gold card users plus 5, silver card users force. 3 points, the ordinary card does not add points, the matching user is matched to get the arbitration result, and the score is deducted to 0 and then treated as 0. The fourth type of arbitration module is: a calling number segment information arbitration module including a calling party number segment information rule, configured to classify and arbitrate the calling party according to the calling party number segment information rule. The number is divided into the following levels: VIP group number plus 10 points, global number section plus 5 points, Monternet gateway plus 5 points, industry gateway does not add points, prepaid number paragraph does not add points, foreign province number minus 2 The sub-group and the external network number segment are reduced by 5 points, etc., and the matching result is obtained after matching the calling and called users, and the score is deducted to 0 and then treated as 0 points. All of the above segment information is configured by the operator. Among them, the VIP group number segment is the number provided by the operator to the large group users. The number of calls made by the users in the number segment is low, and the short number can be dialed directly; the Monternet Gateway is a service provider directly operated by the operator; The gateway is a service provider operated by a non-operator; the prepaid number segment is a number segment set by the operator, and all numbers in the number segment are all prepaid users. The fifth type of arbitration module is: a called number segment information arbitration module that includes a called user number segment information rule, and is used for classifying and arbitrating the called user according to the called user number segment information rule. The number is divided into the following levels: VIP group number plus 10 points, global number section plus 5 points, Monternet gateway plus 5 points, industry gateway does not add points, prepaid number paragraph does not add points, foreign province number minus 2 The sub-network and the external network number segment are deducted by 5 points, etc., and the matching result is obtained after matching the called user, and the score is deducted to 0 and then treated as 0. Step S306: Each arbitration module sends the arbitration result to the interface unit, and the interface unit performs a weighting operation on the arbitration result. In the case of the above five types of arbitration modules, the number of arbitration modules is 5, the formula used for the weighting operation is: Result=Resulti X ω 1 +Result 2 χ ω 2 +Result 3 χ ω 3 +Result 4 χ ω 4 +Result 5 χ ω 5 , wherein the arbitration weights of the arbiters are 0^, ω 2 , ω 3 . ω 4 , ω 5 are all initialized to 1, and the operation result is compared with the set first threshold, where the first threshold is Can be initialized to 80. When the operation result is less than the set first threshold, it is determined that the fuzzy short message is a violation short message, otherwise it is a normal short message. Step S307, the interface unit returns the operation result to each arbitration module, and the arbitration module first The operation result is compared with the set second threshold, where the second threshold of each arbitration module is the same, which is 1/5 times the first threshold set by the interface unit. When the operation result is less than the set second threshold, it is determined that the fuzzy short message is a normal short message; when the operation result is greater than the set second threshold, it is determined that the fuzzy short message is a violation short message. Step S308, comparing whether the judgment result of the fuzzy short message by each arbitration module is the same as the judgment result of the fuzzy short message by the interface unit, if the same, determining that the judgment result by the arbitration module for the fuzzy short message is correct, otherwise determining that the error is incorrect . Step S309: After the arbitration module runs for a period of time, the weight learning is performed. The period is the default system running time, which can be 1 hour, and the current arbitration module determines the correct rate and the first time during the calculated period. For a comparison of expected values, the first expected value here may be set to 50%; if the probability of the current arbitration module determining that the correctness is greater than 50%, the arbitration weight of the current arbitration module is increased, for example, the arbitration weight of the next stage is increased by 0.1; If the current arbitration module judges that the correct phase ratio is less than 50%, the arbitration weight of the current arbitration module is reduced, for example, the arbitration weight of the next stage is reduced by 0.1; the modified arbitration weight is sent to the interface unit, and the interface machine performs The renewal of the arbitration weight. Here, the calculation of the current arbitration module determines that the correct rate is: the probability that the current arbitration module judges correctly = the total number of messages/judged messages that are determined by the respective arbitration modules for the fuzzy short message. Step S310: After the interface machine runs for a period of time, the period of time refers to the default system running time, which may be 1 hour, and compares the calculated probability of the violation short message with the second expected value during the period, where the second The expected value can be initially set to 40%, but not more than 50%; if the probability of the violation short message is less than 1 times of the second expected value, then the current first threshold is reduced, such as reducing the first threshold by 5; If the 3⁄4⁄4 rate of the short message is greater than the second expected value by one time, the current first threshold is increased, for example, the first threshold is increased by 5; and the fourth modified threshold is sent to each arbitration module, and each arbitration is performed. The module updates the first threshold. Step S311: Write a result of the percentage of the violation short message to the total message within 24 hours every day, and the file is a file including the judgment result of the interface unit for the fuzzy short message, and observe whether the judgment result converges within a range, if the convergence The range is small, and around the second expected value, such as convergence in the upper and lower 20% of the second expected value, the system's arbitration weight and the first and second threshold training results are considered normal; if the result is divergent, then After the second expected value is modified, the training is performed, and the judgment result is observed one day later and the corresponding processing is performed. In summary, according to the present invention, it takes a certain time to perform training in the initial stage of the system operation, but after training the optimal expected value and the arbitration weight, the automatic arbitration result of the fuzzy short message is compared with the existing manual. Judgment has been greatly improved, including the accuracy and automation of judgment. It is a good complement to the existing short message monitoring system. At the same time, it can analyze some behavior habits of users sending spam messages, and provides some risks for improving existing systems. Since the existing short message monitoring system can only monitor based on the traffic rules or the keyword rules of the single ticket, that is, the keyword rules configured in the human-computer interaction unit mentioned above, it is easy to disable the monitoring and monitor There are fewer and fewer short messages. If the threshold of traffic reduction is reduced, the number of monitored short messages will increase by an order of magnitude, and the workload of manual operation and maintenance will be large. The content of new violation short messages is endless. Only adding limited and individual keywords will not meet the monitoring requirements. In particular, some new content violation short message delivery modes are difficult to configure only keywords for monitoring. The invention greatly improves the performance of the system by increasing the complex operation relationship between the keywords; and the monitoring system of the present invention automatically learns the arbitration weight through the machine based on the interaction between the arbitration module and the interface unit, thereby A reliable arbitration result can be obtained, which is a great improvement for the existing monitoring system. In addition, the present invention classifies and arbitrates the short message of the violation, and arbitrates the short message that does not contain the specific keyword in the message content, and the fuzzy short message is automatically arbitrated by the arbitration module. The message content of the message, the type of the main and called numbers, the type of the main and the called number, and the like, are comprehensively analyzed. Finally, whether the fuzzy short message is a junk short message is judged, and the present invention is adopted than the existing monitoring system. The single maintenance personnel carry out manual and subjective judgment programs, and the accuracy and efficiency are greatly improved. Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general-purpose computing device, which can be concentrated on a single computing device or distributed over a network composed of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device, such that they may be stored in the storage device by the computing device, or they may be separately fabricated into individual integrated circuit modules, or they may be Multiple modules or steps are made into a single integrated circuit module. Thus, the invention is not limited to any specific combination of hardware and software. The above is only the preferred embodiment of the present invention and is not intended to limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications and changes can be made in the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Claims

权 利 要 求 书 Claim
1. 一种短消息监控系统, 其特征在于, 所述系统包括: 接口单元和仲裁 单元; 其中, A short message monitoring system, the system comprising: an interface unit and an arbitration unit;
所述接口单元 , 用于将模糊短消息转发给仲裁单元进行仲裁; 根 据所述仲裁单元返回的匹配后的各个仲裁结果进行加权运算; 在运算 结果小于设置的第一阀值状态下, 判断出模糊短消息为违规短消息; 所述仲裁单元, 用于经由所述接口单元的转发获取到模糊短消 息; 才艮据不同类别的仲裁规则分别与模糊短消息匹配, 并将匹配后的 各个仲裁结果返回接口单元。  The interface unit is configured to forward the fuzzy short message to the arbitration unit for arbitration; perform weighting operation according to the matched arbitration results returned by the arbitration unit; and determine that the operation result is less than the set first threshold value The fuzzy short message is a violation short message; the arbitration unit is configured to obtain a fuzzy short message by forwarding by the interface unit; respectively, matching the fuzzy short message according to different categories of arbitration rules, and matching each of the arbitrated short messages The result is returned to the interface unit.
2. 根据权利要求 1所述的系统, 其特征在于, 所述仲裁规则包括: 关键 字规则、 用户号码类型信息规则、 用户号段信息规则。 The system according to claim 1, wherein the arbitration rule comprises: a keyword rule, a user number type information rule, and a user number segment information rule.
3. 根据权利要求 2所述的系统, 其特征在于, 所述仲裁单元, 进一步包 括至少两个仲裁模块, 且各个仲裁模块中各自包含的所述仲裁规则分 别属于不同类别; 各个仲裁模块, 用于分别根据各自包含的所述仲裁 规则与模糊短消息匹配 , 并将匹配后的各个仲裁结果返回所述接口单 元。 The system according to claim 2, wherein the arbitration unit further comprises at least two arbitration modules, and the arbitration rules respectively included in each arbitration module belong to different categories; And matching the fuzzy short message according to the arbitration rules respectively included, and returning each of the matched arbitration results to the interface unit.
4. 根据权利要求 3所述的系统, 其特征在于, 所述各个仲裁模块, 进一 步用于分别根据所述各个仲裁结果对所述模糊短消息进行判断 , 并获 得通过各个仲裁模块针对所述模糊短消息的判断结果; The system according to claim 3, wherein each of the arbitration modules is further configured to determine the fuzzy short message according to the respective arbitration results, and obtain the fuzzy by each arbitration module The result of the short message;
获取通过所述接口单元针对所述模糊短消息的判断结果, 在设置 的系统运行时间内, 在当前仲裁模块与接口单元判断结果相同的概率 大于第一期望值状态下, 增加当前仲裁模块的仲裁权值; 在当前仲裁 模块与接口单元判断结果相同的概率小于第一期望值状态下, 减少当 前仲裁模块的仲裁权值; 将 4爹改后的仲裁权值返回接口单元进行仲裁 权值的更新。  Acquiring the judgment result of the fuzzy short message by using the interface unit, in the set system running time, increasing the arbitration right of the current arbitration module when the probability that the current arbitration module and the interface unit determine the same result is greater than the first expected value If the probability that the current arbitration module and the interface unit determine the same result is less than the first expected value, the arbitration weight of the current arbitration module is reduced; and the 4 modified arbitration weight is returned to the interface unit to update the arbitration weight.
5. 根据权利要求 3或 4所述的系统 , 其特征在于 , 所述接口单元 , 进一 步用 于进行所述加权运算 时所采用 的运算公式为 : ^Result. x<w. The system according to claim 3 or 4, wherein the interface unit is further configured to perform the weighting operation as: ^Result. x<w.
Result= '=i ; 其中, Result 为所述运算结果, Result'为所述 各个仲裁结果, ^为所述各个仲裁模块的仲裁权值, n 为仲裁模块的 个数。 Result= '=i ; where Result is the result of the operation, Result ' is the result of each arbitration, ^ is the arbitration weight of each arbitration module, and n is the number of arbitration modules.
6. 根据权利要求 4所述的系统, 其特征在于, 所述各个仲裁模块, 进一 步用于在所述仲裁结果大于设置的第二阀值状态下, 判断出所述模糊 短消息为违规短消息; 其中, 所述第二阀值为所述第一阀值的 1/n倍; n为仲裁模块的个数。 The system according to claim 4, wherein each of the arbitration modules is further configured to determine that the fuzzy short message is a violation short message when the arbitration result is greater than a set second threshold value. Wherein the second threshold is 1/n times the first threshold; n is the number of arbitration modules.
7. 根据权利要求 5所述的系统, 其特征在于, 所述接口单元, 进一步用 于在设置的系统运行时间内 , 在接口单元判断出违规短消息的相无率大 于第二期望值 1倍的状态下, 增加当前第一阀值; 在接口单元判断出 违规短消息的概率小于第二期望值 1倍的状态下,减少当前第一阀值; 将修改后的第一阀值返回所述各个仲裁模块进行更新。 The system according to claim 5, wherein the interface unit is further configured to: during the set system running time, determine, by the interface unit, that the phase ratio of the violation short message is greater than the second expected value by one time. In the state, the current first threshold is increased; when the interface unit determines that the probability of the violation short message is less than 1 time of the second expected value, the current first threshold is decreased; and the modified first threshold is returned to the respective arbitration The module is updated.
8. 一种短消息监控方法, 其特征在于, 所述方法包括: A short message monitoring method, the method comprising:
仲裁单元根据不同类别的仲裁规则分别与所述模糊短消息匹配 , 并将匹配后的各个仲裁结果返回接口单元;  The arbitration unit respectively matches the fuzzy short message according to different types of arbitration rules, and returns the matched arbitration results to the interface unit;
接口单元才艮据匹配后的各个仲裁结果进行加权运算; 当运算结果 'J、于设置的第一阀值时, 判断出模糊短消息为违规短消息。  The interface unit performs a weighting operation according to each of the matched arbitration results; when the operation result 'J, at the set first threshold, it is determined that the fuzzy short message is a violation short message.
9. 根据权利要求 8所述的方法, 其特征在于, 所述仲裁规则包括: 关键 字规则、 用户号码类型信息规则、 用户号段信息规则。 The method according to claim 8, wherein the arbitration rule comprises: a keyword rule, a user number type information rule, and a user number segment information rule.
10. 根据权利要求 9所述的方法 , 其特征在于 , 当所述仲裁单元包括至少 两个仲裁模块, 且各个仲裁模块中各自包含的所述仲裁规则分别属于 不同类别情况下, 所述仲裁单元进行所述匹配进一步包括: 各个仲裁 模块分别根据各自包含的所述仲裁规则与模糊短消息匹配 , 并将匹配 后的各个仲裁结果返回所述接口单元。 The method according to claim 9, wherein when the arbitration unit includes at least two arbitration modules, and the arbitration rules respectively included in each arbitration module belong to different categories, the arbitration unit Performing the matching further includes: each arbitration module respectively matching the fuzzy short message according to the arbitration rule respectively included, and returning each of the matched arbitration results to the interface unit.
11. 根据权利要求 10所述的方法, 其特征在于, 在所述接口单元根据匹配 后的各个仲裁结果进行加权运算; 当运算结果小于设置的第一阀值时, 判断出模糊短消息为违规短消息之后 , 所述方法进一步包括: The method according to claim 10, wherein the interface unit performs a weighting operation according to each of the matched arbitration results; when the operation result is less than the set first threshold, determining that the fuzzy short message is a violation After the short message, the method further includes:
A、 所述各个仲裁模块分别根据所述各个仲裁结果对所述模糊短 消息进行判断, 并获得通过各个仲裁模块针对所述模糊短消息的判断 结果; A. Each of the arbitration modules respectively shorts the blur according to the respective arbitration results The message is judged, and the judgment result of the fuzzy short message by each arbitration module is obtained;
B、 获取通过所述接口单元针对所述模糊短消息的判断结果; 在 设置的系统运行时间内, 在当前仲裁模块与接口单元判断结果相同的 率大于第一期望值情况下, 增加当前仲裁模块的仲裁权值; 在当前 仲裁模块与接口单元判断结果相同的概率小于第一期望值情况下 , 减 少当前仲裁模块的仲裁权值; 将 4爹改后的仲裁权值返回至所述接口单 元进行仲裁权值的更新;  B. Obtain a judgment result of the fuzzy short message by using the interface unit. In a set system running time, if the rate at which the current arbitration module and the interface unit determine the same result is greater than the first expected value, increase the current arbitration module. Arbitration weight; when the probability that the current arbitration module and the interface unit determine the same result is less than the first expected value, reduce the arbitration weight of the current arbitration module; and return the 4 modified arbitration weight to the interface unit for arbitration Update of value;
C、 所述接口单元才艮据返回的更新后的仲裁权值进行加权运算 , 并继续执行针对所述模糊短消息的判断。  C. The interface unit performs a weighting operation according to the returned updated arbitration weight value, and continues to perform the determination for the fuzzy short message.
12. 根据权利要求 10或 11所述的方法, 其特征在于, 所述接口单元进行 The method according to claim 10 or 11, wherein the interface unit performs
^Result. x<w. ^Result. x<w.
所述加权运算时所采用的运算公式为: Result= '=i ; 其中,  The operation formula used in the weighting operation is: Result= '=i ;
Result 为所述运算结果, Result<为所述各个仲裁结果, 为所述各个 仲裁模块的仲裁权值, n为仲裁模块的个数。 Result is the result of the operation, and Result < is the arbitration result, which is the arbitration weight of each arbitration module, and n is the number of arbitration modules.
13. 根据权利要求 11所述的方法, 其特征在于, 所述 A中, 所述各个仲裁 模块分别根据所述各个仲裁结果对所述模糊短消息进行判断进一步包 括: The method according to claim 11, wherein, in the A, the determining, by the respective arbitration modules, the fuzzy short message according to the respective arbitration results, further comprising:
当所述仲裁结果大于设置的第二阀值时, 判断出所述模糊短消息 为违规短消息; 其中, 所述第二阀值为所述第一阀值的 1/n倍; n为仲 裁模块的个数。  When the arbitration result is greater than the set second threshold, determining that the fuzzy short message is a violation short message; wherein, the second threshold is 1/n times the first threshold; n is arbitration The number of modules.
14. 根据权利要求 12所述的方法, 其特征在于, 所述 C之后, 所述方法还 包括: The method according to claim 12, wherein after the C, the method further comprises:
在设置的系统运行时间内 , 在所述接口单元判断出违规短消息的 率大于第二期望值 1倍的情况下, 增加当前第一阀值; 在所述接口 单元判断出违规短消息的 4既率小于第二期望值 1倍的情况下, 减少当 前第一阀值; 将修改后的第一阀值返回各个仲裁模块进行更新。  In the set system running time, when the interface unit determines that the rate of the violation short message is greater than the second expected value by one time, the current first threshold is increased; and the interface unit determines that the violation short message is 4 When the rate is less than 1 time of the second expected value, the current first threshold is decreased; the modified first threshold is returned to each arbitration module for updating.
PCT/CN2009/074516 2009-05-20 2009-10-19 System and method for short message monitoring WO2010133063A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200910084138.3 2009-05-20
CN2009100841383A CN101895828B (en) 2009-05-20 2009-05-20 Short message monitoring system and method

Publications (1)

Publication Number Publication Date
WO2010133063A1 true WO2010133063A1 (en) 2010-11-25

Family

ID=43104863

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2009/074516 WO2010133063A1 (en) 2009-05-20 2009-10-19 System and method for short message monitoring

Country Status (2)

Country Link
CN (1) CN101895828B (en)
WO (1) WO2010133063A1 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113965899A (en) * 2021-12-21 2022-01-21 杭州云在线科技有限公司 Short message deduction detection server and method
CN115623485A (en) * 2022-12-20 2023-01-17 杭州孝道科技有限公司 Short message bombing detection method, system, server and storage medium

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102612007B (en) * 2011-01-19 2015-06-24 中国电信股份有限公司 Flow control method and device for short messages
CN103067896B (en) * 2013-01-17 2015-08-19 中国联合网络通信集团有限公司 Method for filtering spam short messages and device
US20180197099A1 (en) * 2017-01-11 2018-07-12 Google Inc. User state predictions for presenting information

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060168031A1 (en) * 2004-12-21 2006-07-27 Lucent Technologies, Inc. Detection of unwanted messages (spam)
CN101136874A (en) * 2007-07-25 2008-03-05 华南理工大学 Compound decision based anti-rubbish E-mail error filtering method and system
CN101257671A (en) * 2007-07-06 2008-09-03 浙江大学 Method for real time filtering large scale rubbish SMS based on content

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7711779B2 (en) * 2003-06-20 2010-05-04 Microsoft Corporation Prevention of outgoing spam
CN1741526A (en) * 2005-09-05 2006-03-01 北京启明星辰信息技术有限公司 Method and system for detecting exception flow of network
CN101335920B (en) * 2008-07-15 2011-04-13 中国联合网络通信集团有限公司 Rubbish short message recognition system and method based on calling number location and transmitted content

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060168031A1 (en) * 2004-12-21 2006-07-27 Lucent Technologies, Inc. Detection of unwanted messages (spam)
CN101257671A (en) * 2007-07-06 2008-09-03 浙江大学 Method for real time filtering large scale rubbish SMS based on content
CN101136874A (en) * 2007-07-25 2008-03-05 华南理工大学 Compound decision based anti-rubbish E-mail error filtering method and system

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113965899A (en) * 2021-12-21 2022-01-21 杭州云在线科技有限公司 Short message deduction detection server and method
CN113965899B (en) * 2021-12-21 2022-04-01 杭州云在线科技有限公司 Short message deduction detection server and method
CN115623485A (en) * 2022-12-20 2023-01-17 杭州孝道科技有限公司 Short message bombing detection method, system, server and storage medium
CN115623485B (en) * 2022-12-20 2023-04-07 杭州孝道科技有限公司 Short message bombing detection method, system, server and storage medium

Also Published As

Publication number Publication date
CN101895828B (en) 2013-01-16
CN101895828A (en) 2010-11-24

Similar Documents

Publication Publication Date Title
CN103404193B (en) The connection that adjustment data transmission is established with the transmission being optimized for through wireless network
EP2830044B1 (en) Instruction processing method, apparatus, and system
CN108156265B (en) A kind of application control method and mobile device
CN101257671B (en) Method for real time filtering large scale rubbish SMS based on content
Hassani et al. Context-as-a-Service Platform: exchange and share context in an IoT ecosystem
WO2010133063A1 (en) System and method for short message monitoring
CN107005597A (en) The wireless flow management system cached based on user characteristics in mobile device
WO2013097714A1 (en) Statistical analysis and prompting method and system for mobile terminal internet traffic
CN106961384A (en) A kind of message treatment method and electronic equipment
CN107846295A (en) Micro services configuration device and method
CN107193836B (en) Identification method and device
Saadat Survey on spam filtering techniques
CN107832132B (en) Application control method and device, storage medium and electronic equipment
CN1809821A (en) Feedback loop for spam prevention
CN107147724A (en) A kind of information push method, server and computer-readable recording medium
CN104348974A (en) Keyword-verification-based specific message prompting method for communication group
CN107181664B (en) Automatic fusing message sending method, device and system
CN110297621A (en) Method, apparatus, equipment and the storage medium of application message notice management
CN112131036A (en) Overload protection method, device, equipment and computer readable storage medium
CN110072251B (en) Method and device for analyzing user communication behavior and managing user
CN101114907B (en) Method and system for managing and filtering black list
CN108009944A (en) A kind of information technology consultative service system based on internet
CN107704734A (en) A kind of recognition methods of user account and its equipment
CN107171948A (en) A kind of method, device and the mail server of filtering spam mail
WO2013078798A1 (en) Method and system for monitoring a short message

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 09844815

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 09844815

Country of ref document: EP

Kind code of ref document: A1