13 results about "Sms spam" patented technology

The application discloses a method and device for preventing and treating junk messages. The method comprises the following steps: receiving a target message, determining a target number of the target message and a first use behavior of the target number, wherein the first use behavior comprises multiple types of use behaviors; determining a use behavior value corresponding to the first use behavior of the target number, and judging whether the target number is added to a monitoring number library based on the use behavior value, wherein the use behavior value is used to reflect the probability of the target number existing the first use behavior; in the case of judging that the target number is added to the monitoring number library, when the content of the target message includes a first preset keyword, intercepting the target message and limiting the message sending function of the target number; when the content of the target message does not include the first preset keyword, intercepting the target message. The application solves the technical problem that the use behavior of the message sender is not considered in the related art, so that the junk messages cannot be found and intercepted in time and effectively.

The present application relates to the technical field of short message processing, in particular to a method and device for intelligent analysis and processing of big data of spam short messages, the method comprising: judging feature difference of short message content; according to the judging result of feature difference, giving the short message corresponding spam short message feature value, and accumulating the spam short message feature values together to obtain a first spam short message feature value; if the first spam short message feature value exceeds a first preset threshold, intercepting the short message, if the first spam short message feature value does not exceed the first preset threshold, inputting the short message content into a short message content intelligent analysis strategy module to obtain a second spam short message feature value; performing addition accumulation calculation on the first spam short message feature value and the second spam short message feature value to obtain a third spam short message feature value, if the third spam short message feature value exceeds a second preset threshold, intercepting the short message, if the third spam short message feature value does not exceed the second preset threshold, passing the short message. The present application improves short message quality and enhances the image of a telecom operator enterprise.

The invention discloses a spam short message quasi-real-time classification method and system based on a large language model workflow, and relates to the technical field of artificial intelligence and mobile communication big data, and the method comprises the steps: extracting to-be-classified suspected spam short message data in a previous time period from a database; grouping identifiers are constructed for the suspected junk short message data, data compression is carried out, and a data pair combination of the suspected junk short messages and the corresponding grouping identifiers is constructed; pre-arranging a junk short message classification workflow based on load limitation of a large language model application development platform, and inputting the data pair combination into the junk short message classification workflow to generate reply information; extracting a classification result according to the reply information; and mapping all suspected junk short messages in the suspected junk short message data in the last time period to corresponding classification results according to the grouping identifiers, and returning the classification results to the database. According to the invention, efficient, accurate and automatic classification of massive suspected junk short messages can be realized.

The application provides a short message sending method and related device to improve the safety of short message receiving and sending and to govern the behavior of junk short messages. A short message sending method is applied to a first client, and the method comprises: in response to a short message sending instruction, generating a first authentication data packet, wherein the first authentication data packet comprises a number corresponding to the first client, a target short message, a number corresponding to a second client, historical sending behavior information of the first client, and the second client is a receiving client of the target short message; and sending the first authentication data packet.

The system according to this embodiment aims to efficiently analyze the content of received SMS messages, identify fraud and spam, automatically copy authentication codes, and classify them according to their content. [Solution] The system according to the embodiment comprises an analysis unit, a discrimination unit, a copy unit, and a classification unit. The analysis unit analyzes the content of the received SMS. The discrimination unit identifies fraud and spam based on the content analyzed by the analysis unit. The copy unit automatically copies the authentication code based on the content analyzed by the analysis unit. The classification unit classifies the SMS into conversation, authentication, spam, etc., based on the content analyzed by the analysis unit.

The invention discloses a short message issuing congestion intelligent processing method and device, and relates to the technical field of communication network flow scheduling and congestion control. An emergency scene is intelligently recognized through a dynamic threshold model, and a processing mode is switched; dynamically setting a short message priority according to the mode and the calling number; a resource allocation method of a fixed dedicated channel and an adjustable public channel is adopted, and a flow rule is predicted by utilizing machine learning so as to dynamically optimize public resource scheduling; differential time window management based on multi-dimensional evaluation and junk short message linkage punishment are carried out on the short messages; designing an intelligent cache library with priority perception capability to perform congestion buffering; in the emergency mode, the retransmission strategy of the short message center platform is actively intervened to cooperate with flow control, fine identification, self-adaptive scheduling and intelligent congestion control of the short message flow are realized, and the guarantee capability of high-priority services, the utilization efficiency of system resources and the overall service reliability are remarkably improved.

The invention discloses a short message management and control method and device, a storage medium and electronic equipment, and relates to the technical field of data processing, and the method comprises the steps: firstly obtaining a sending terminal address of a short message marked as a junk short message; then determining a plurality of target areas corresponding to the junk short message through a clustering algorithm based on the geographic position characteristics of the sending terminal address; in a preset time period, counting the frequency of sending the short message by the sending terminal in the target area; and controlling the short messages sent by the target area based on the frequency of sending the short messages by the sending terminal. Compared with the prior art, the method has the advantages that the address geographic position of the junk short message sending terminal is screened, classified and aggregated, the sending frequency in the high-incidence area is monitored after the high-incidence area is identified, and the short messages are monitored and effectively filtered, so that novel or variant junk short messages are actively and accurately prevented and controlled; and the identification flexibility and accuracy of the junk short messages are improved.

The invention relates to the technical field of communication security, and relates to a junk short message identification and filtering method and system fused with artificial intelligence. The method is realized through software and hardware cooperation of a mobile device or a server system, when a short message arrives, a receiving module analyzes a protocol to extract plain text content, and then the plain text content is input into a pre-trained deep learning classification model for analysis. The model adopts an advanced neural network architecture, comprises structures such as a convolutional neural network, a recurrent neural network or Transform, and can automatically learn text features and output a junk short message probability value; comparative judgment is carried out through a configurable threshold value, various filtering operations are carried out on the identified spam short messages, and meanwhile, user feedback and a self-adaptive adjustment mechanism are supported. According to the method, the deep semantic features of the junk short messages are automatically captured by using the deep learning model, the disguised complex junk short messages are effectively identified, meanwhile, the user privacy and the equipment performance are guaranteed, and a complete technical solution is provided for mobile communication security.

This invention relates to the field of communication technology, and in particular to an industry-specific SMS sending and verification method for preventing spam messages. The method includes: an operator providing a template application platform and dedicated accounts to enterprises holding fixed industry SMS access numbers; after logging in, enterprises submit SMS template applications with variable content annotations; upon approval, they receive a two-digit template number, which, along with the access number, template number, and annotated content, is stored in the industry SMS gateway's storage medium; the enterprise constructs a receiving request using the target mobile number as the receiving address, appending the two-digit template number, the variable content annotations, and the sending end using the access number; the gateway receives the request, intercepts key information, retrieves the corresponding template, removes both curly braces and internal characters, and compares the results; if they match, the SMS is sent; otherwise, it is rejected and a content mismatch message is returned. This invention achieves source control of spam messages and strengthens compliance to improve governance efficiency through template approval and storage, standardized request verification, interception of violations, and code-based log storage.

The invention discloses an adversarial short message dichotomy method based on deep semantic comprehension and multi-mode gating. The method comprises the following steps: acquiring a short message to be identified; the method comprises the following steps: extracting a semantic embedding feature vector from a short message to be identified through a pre-trained chine-roberta-wwm-ext-large model encoder, and extracting a semantic embedding feature vector from the short message to be identified through a pre-trained chine-roberta-wwm-ext model encoder; extracting pinyin embedding feature vectors from the pinyin sequence of the short message to be recognized through a pre-trained pinyin denoising auto-encoder; extracting a font embedding feature vector from the font structure of the short message to be identified through a pre-trained font denoising auto-encoder; performing weighted fusion on the semantic embedding feature vector, the pinyin embedding feature vector and the font embedding feature vector through a pre-trained gating fusion module to obtain weighted fusion features; and using a pre-trained short message classifier to classify the short messages to be identified according to the weighted fusion features. Through multi-modal fusion and an effective attention mechanism, high-precision and interpretable classification of resistant short messages is realized, and reliable technical support is provided for real-time detection of spam short messages in an actual scene.

The invention discloses a spam short message real-time detection method and system, electronic equipment and a storage medium, and belongs to the technical field of communication security and artificial intelligence. The method comprises the steps of obtaining to-be-detected short message content and determining a short message length L; extracting the number n of non-standard characters and the number m of unusual Chinese characters in the short message content; fusing the short message content, the pinyin representation of the short message content and the stroke sequence of each Chinese character in the short message content into a unified embedded representation; inputting the embedded representation into a pre-trained dichotomy large language model to obtain a probability p that the short message is a junk short message; calculating a score according to the probability p, the number n of non-standard characters, the number m of unseen Chinese characters and the length L of the short message, and carrying out normalization processing on the score; and judging whether the short message is a junk short message based on a comparison result of the normalized score and a preset threshold value. The detection accuracy is improved, the false alarm rate is effectively reduced, the user experience is remarkably improved, and the real-time detection requirement of the mobile terminal is met.

The invention discloses a junk short message identification method and system based on multi-source data feature feedback, and relates to the technical field of junk short message identification, and the method comprises the steps: obtaining a sending number of a target short message, short message content, a risk label of the sending number, historical reported times, a malicious domain name library and user interaction behavior data, and according to the user interaction behavior data, carrying out the identification of the junk short message. And obtaining a trust scoring coefficient, and obtaining a second risk score according to the short message content. According to the method, the identification precision is improved by comprehensively sending the number risk tag, the short message content, the user interaction behavior and the link risk, the deformed sensitive words are identified based on initial consonant, vowel and tone digital codes and the weighted editing distance, the risk judgment threshold is optimized according to misjudgment feedback and correct interception feedback, and the method adapts to communication habits of different users; a time decay factor and a user reply proportion are introduced, a real relationship between a sending number and a target user is evaluated, and the misjudgment rate of short message interception is reduced.

The application discloses a kind of high-efficient and accurate junk message interception method, equipment and system, belong to communication technical field.Method includes: synchronously recent year's mobile phone number of entering network as blacklist and storage;Length screening is carried out to be issued short message first, and the short message of length≤40 bytes is directly issued;The verification whether the sending number of the short message of length>40 bytes is in blacklist is not in then, is issued, and is carried out short message content similarity matching in then;Similarity is greater than preset threshold X then, intercepts and updates junk message content library, otherwise, is issued;Similarity is calculated by the ratio of set intersection and union.This application further discloses the equipment and system for realizing this method.This application significantly improves the efficiency and accuracy of junk message interception, reduces the processing pressure of short message center, and has low deployment cost and strong sustainability, and can be widely applied to junk message management of operator short message center.