CN-121985304-A - Method and equipment for processing short message in real time

CN121985304ACN 121985304 ACN121985304 ACN 121985304ACN-121985304-A

Abstract

The application provides a real-time processing method of a short message and electronic equipment, the real-time processing method of the short message comprises the steps of receiving a short message sending request containing short message content to be sent, initiating a real-time classification request of the short message content, enabling a classification module to carry out text matching on the short message content and classified short messages to obtain text matching degree, determining the short message type of the short message content based on the text matching degree, feeding back the short message type to a supervision platform, enabling the supervision platform to process the short message sending request according to the short message type, carrying out real-time processing of the short message without depending on a large number of labeling samples, solving the cold starting problem, adapting to the dynamic change of the short message content by depending on text matching logic, needing not to manually update samples frequently, meeting real-time identification requirements, reducing the dependence on sample labeling accuracy, improving classification consistency by direct matching of the classified short message, and meeting the classification accuracy requirements.

Inventors

WANG YUEHUA
LI HAO
GUO WEIHENG

Assignees

北京神州泰岳软件股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260311

Claims (10)

1. The method is characterized by being applied to a short message gateway, and comprises the following steps: receiving a short message sending request, wherein the short message sending request comprises short message content to be sent; Initiating a real-time classification request of the short message content so that a classification module carries out text matching on the short message content and the classified short message to obtain text matching degree; Determining the short message type of the short message content based on the text matching degree; And feeding the short message type back to a supervision platform so that the supervision platform processes the short message sending request according to the short message type.
2. The method for processing the short message in real time according to claim 1, wherein the determining the type of the short message content based on the text matching degree comprises: if the text matching degree is greater than or equal to a preset threshold value, determining a first short message type of the short message content based on the short message type corresponding to the classified short message; And if the text matching degree is smaller than the preset threshold, performing type identification on the short message content based on a large model to obtain a second short message type of the short message content.
3. The method for processing the short message in real time according to claim 2, wherein the determining the short message type of the short message content based on the short message type corresponding to the classified short message comprises: Sorting the classified short messages according to the text matching degree to obtain a classified short message list; taking the classified short messages in the preset order in the classified short message list as target classified short messages; counting the number of the short message types of the target classified short messages, and taking the short message type with the number of the short message types being larger than a number threshold value as a first short message type of the short message content.
4. The method for processing the short message in real time according to claim 1, wherein the initiating the real-time classification request of the short message content to enable the classification module to perform text matching on the short message content and the classified short message, and obtaining the text matching degree comprises: initiating a real-time classification request of the short message content so that the classification module extracts a first text feature of the short message content; Performing text matching based on the first text feature and second text features of the classified short messages extracted in advance to obtain matching feature items; and carrying out average value calculation based on the short message quantity of the classified short messages and the matching characteristic items to obtain the text matching degree.
5. The method for processing the short message in real time according to claim 2, wherein the method further comprises: Labeling the short message content based on a second short message type of the short message content to obtain a labeled short message; And writing the marked short message into a short message library storing the classified short messages so as to determine the short message types of other short messages contained in other short message sending requests based on the marked short message stored in the short message library and the classified short message under the condition that other short message sending requests are received.
6. The method for processing a short message in real time according to claim 1, wherein the method further comprises: acquiring a short message sample; Performing type identification on the short message sample through a large model to obtain the short message type of the short message sample; And labeling the short message sample based on the short message type of the short message sample to obtain the classified short message.
7. The method for processing the short message in real time according to claim 1, wherein the method comprises the following steps: Judging whether the short message information of the short message content accords with the report category of the short message content or not; if not, the sender corresponding to the short message sending request is subjected to sending restriction.
8. A device for processing a short message in real time, the device comprising: The request acquisition module is used for receiving a short message sending request, wherein the short message sending request comprises short message content to be sent; the text matching module is used for initiating a real-time classification request of the short message content so that the classification module carries out text matching on the short message content and the classified short messages to obtain text matching degree; the type determining module is used for determining the short message type of the short message content based on the text matching degree; and the short message type feedback module is used for feeding the short message type back to the supervision platform so that the supervision platform processes the short message sending request according to the short message type.
9. A computer readable storage medium storing a computer program which, when executed by a processor, implements the method of real-time processing of a short message according to any of the preceding claims 1-7.
10. An electronic device, comprising: One or more processors, and A memory configured to store one or more programs; wherein the one or more processors implement the method for real-time processing of a short message according to any one of claims 1-7 when the one or more programs are executed by the one or more processors.

Description

Method and equipment for processing short message in real time Technical Field The embodiment of the application relates to the field of data processing, in particular to a method and equipment for processing a short message in real time. Background With the development of mobile communication technology, industry short messages become important channels for enterprises to communicate with users, and various scenes such as business popularization, notification reminding and the like are covered. According to the supervision requirement, the industry short messages need to be reported and prepared according to the category, but in practice, a large amount of short message contents are not in accordance with the reporting and preparing category, so that the supervision work is challenged. The existing short message type identification is based on a traditional machine learning algorithm, classification is achieved through a labeling sample training model, but obvious short plates exist in the scheme, such as high dependence on sample labeling accuracy, insufficient classification accuracy can be caused by labeling deviation or incomplete labeling, when the short message content dynamically changes, the model needs to manually update a sample for retraining, real-time identification requirements are difficult to adapt, in addition, the system is lack of the labeling sample during initial deployment, cold starting problems exist, and manual labeling cost is high and period is long. These problems make the prior art unable to efficiently and accurately meet the requirements of real-time processing of short messages, and therefore, a real-time processing method of short messages with high efficiency and high accuracy is needed. Disclosure of Invention The embodiment of the application provides a method and equipment for processing a short message in real time, which can improve the accuracy and efficiency of short message type identification on the basis of real-time processing of the short message in real time. In order to achieve the above purpose, the embodiment of the application adopts the following technical scheme: The embodiment of the application provides a method for processing a short message in real time, which comprises the steps of receiving a short message sending request, initiating a real-time classification request of the short message content to enable a classification module to carry out text matching on the short message content and classified short messages to obtain text matching degree, determining the short message type of the short message content based on the text matching degree, and feeding the short message type back to a supervision platform to enable the supervision platform to process the short message sending request according to the short message type. In another possible implementation manner, the determining the text message type of the text message content based on the text matching degree includes determining a first text message type of the text message content based on the text message type corresponding to the classified text message if the text matching degree is greater than or equal to a preset threshold value, and performing type recognition on the text message content based on a big model if the text matching degree is less than the preset threshold value to obtain a second text message type of the text message content. In another possible implementation manner, the determining the short message type of the short message content based on the short message type corresponding to the classified short message includes sorting the classified short messages according to the text matching degree to obtain a classified short message list, taking the classified short messages in a preset order in the classified short message list as target classified short messages, counting the number of the short message types of the target classified short messages, and taking the short message type with the number of the short message types being greater than a number threshold as a first short message type of the short message content. In yet another possible implementation manner, the step of initiating the real-time classification request of the short message content to enable the classification module to perform text matching on the short message content and the classified short messages to obtain text matching degree includes the steps of initiating the real-time classification request of the short message content to enable the classification module to extract first text features of the short message content, performing text matching on the basis of the first text features and a plurality of pre-extracted second text features of the classified short messages to obtain matching feature items, and performing average calculation on the basis of the number of the short messages of the classified short messages and the matching feature items to obtain the text matching degree. In a furthe