Information sending method and device, computer equipment and storage medium
By determining the similarity of the initial SMS set in the information sending method and classifying it, generating representative SMS to filter target SMS, the problem of repeated SMS sending is solved and the utilization rate of communication resources is improved.
Patent Information
- Application Number
- CN202410344259.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-25
- Publication Date
- 2025-05-16
AI Technical Summary
The existing information sending method has the lack of collaboration and information sharing among various departments in the enterprise, resulting in repeated sending of similar text messages to users, resulting in wasting of communication resources.
Each SMS group is obtained by determining the first similarity of the initial SMS set and classifying the initial SMS based on the clustering algorithm. Then, the representative text messages for each SMS group are generated based on the preset text summary model, the second similarity between the representative text message and the initial SMS is calculated, the target SMS is selected and sent.
It realizes automatic filtering of similar text messages, avoids repeated sending of similar text messages, and improves the utilization rate of communication resources.
Smart Images
Figure CN120011822A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of big data technology, and in particular to an information sending method, apparatus, computer equipment, storage medium and computer program product. Background Art
[0002] At present, many companies are actively using various modern tools to interact with users and conduct product marketing and information reminders. Among them, SMS outbound calls have become a common and important marketing method. Staff will use the message sending method to send SMS containing marketing content to users.
[0003] The current method of sending information is that the staff edits the SMS to be sent, obtains the pre-sent SMS, and then sends multiple pre-sent SMS to the operator at a scheduled time, and the operator forwards each SMS to the user.
[0004] However, the current method of sending information, due to the lack of coordination and information sharing among departments in the enterprise, causes the enterprise to repeatedly send similar text messages to users, thereby causing a waste of communication resources. Summary of the invention
[0005] Based on this, it is necessary to provide an information sending method, apparatus, computer device, computer readable storage medium and computer program product to address the above technical issues.
[0006] In a first aspect, the present application provides a method for sending information, comprising:
[0007] Determine a first similarity between each initial short message in the initial short message set, and classify each initial short message based on each first similarity and a clustering algorithm to obtain each short message group;
[0008] Generate a representative SMS corresponding to each SMS group according to a preset text summary model, and calculate a second similarity between the representative SMS and each of the initial SMS in the SMS group;
[0009] According to each of the second similarities and each of the first similarities, a target text message is screened from each of the initial text messages included in the text message group, and the target text message is sent to a target object.
[0010] In one embodiment, determining the first similarity between the initial short messages in the initial short message set includes:
[0011] According to a preset time period, each pre-sent initial short message is intercepted, and each of the initial short messages is numbered to obtain an initial short message set;
[0012] For each of the initial short messages in the initial short message set, constructing the initial short message and each of the other initial short messages into a short message pair according to the serial number of the initial short message;
[0013] The similarity of each of the short message pairs is calculated according to a preset semantic similarity model to obtain a first similarity between the initial short messages in the short message pair.
[0014] In one embodiment, the initial short messages are classified based on the first similarities and the clustering algorithm to obtain short message groups, including:
[0015] constructing a similarity matrix according to each of the first similarities, and clustering the similarity matrix based on a clustering algorithm to obtain clusters;
[0016] According to the corresponding relationship between each of the first similarities in each of the clusters and each of the initial short messages, each of the initial short messages is classified to obtain each short message group.
[0017] In one embodiment, generating a representative SMS corresponding to each SMS group according to a preset text summary model, and calculating a second similarity between the representative SMS and each initial SMS in the SMS group, includes:
[0018] For each of the SMS groups, extract text features of the SMS group according to a preset text summary model to obtain a representative SMS corresponding to the SMS group;
[0019] The similarity between the representative text message and each of the initial text messages in the text message group is calculated based on a preset semantic similarity model to obtain a second similarity between each of the initial text messages and the representative text message.
[0020] In one embodiment, the step of screening target text messages from the initial text messages included in the text message group according to the second similarities and the first similarities includes:
[0021] Determining each initial target text message from each initial text message included in the text message group according to a preset second similarity threshold and a second similarity of each initial text message;
[0022] For each of the initial target text messages, determining whether the first similarity between the initial target text message and the remaining initial target text messages is less than a preset first similarity threshold;
[0023] If the first similarity is less than a first similarity threshold, the initial target short message is determined as a target short message.
[0024] In one embodiment, determining each initial target text message from each initial text message included in the text message group according to a preset second similarity threshold and a second similarity of each initial text message includes:
[0025] For each of the initial text messages in the text message group, determining whether the second similarity of the initial text message is less than a preset second similarity threshold;
[0026] If the second similarity of the initial text message is less than the second similarity threshold, the initial text message is determined as the initial target text message.
[0027] In one embodiment, sending the target SMS to the target object includes:
[0028] Determine whether there are multiple target text messages in the same text message group;
[0029] If there are multiple target SMS messages in the same SMS group, the sending time of each target SMS message is set according to a preset time interval; the time interval is greater than the time interval threshold;
[0030] The target short message is sent to the target object according to the sending time of the target short message.
[0031] In a second aspect, the present application also provides an information sending device, including:
[0032] A classification module, used for determining a first similarity between each initial short message in the initial short message set, and classifying each initial short message based on each first similarity and a clustering algorithm to obtain each short message group;
[0033] A calculation module, used for generating a representative SMS corresponding to each SMS group according to a preset text summary model, and calculating a second similarity between the representative SMS and each of the initial SMS in the SMS group;
[0034] The screening module is used for screening target short messages from the initial short messages included in the short message group according to the second similarities and the first similarities, and sending the target short messages to the target object.
[0035] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:
[0036] Determine a first similarity between each initial short message in the initial short message set, and classify each initial short message based on each first similarity and a clustering algorithm to obtain each short message group;
[0037] Generate a representative SMS corresponding to each SMS group according to a preset text summary model, and calculate a second similarity between the representative SMS and each of the initial SMS in the SMS group;
[0038] According to each of the second similarities and each of the first similarities, a target text message is screened from each of the initial text messages included in the text message group, and the target text message is sent to a target object.
[0039] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:
[0040] Determine a first similarity between each initial short message in the initial short message set, and classify each initial short message based on each first similarity and a clustering algorithm to obtain each short message group;
[0041] Generate a representative SMS corresponding to each SMS group according to a preset text summary model, and calculate a second similarity between the representative SMS and each of the initial SMS in the SMS group;
[0042] According to each of the second similarities and each of the first similarities, a target text message is screened from each of the initial text messages included in the text message group, and the target text message is sent to a target object.
[0043] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:
[0044] Determine a first similarity between each initial short message in the initial short message set, and classify each initial short message based on each first similarity and a clustering algorithm to obtain each short message group;
[0045] Generate a representative SMS corresponding to each SMS group according to a preset text summary model, and calculate a second similarity between the representative SMS and each of the initial SMS in the SMS group;
[0046] According to each of the second similarities and each of the first similarities, a target text message is screened from each of the initial text messages included in the text message group, and the target text message is sent to a target object.
[0047] The above-mentioned information sending method, device, computer equipment, storage medium and computer program product determine the first similarity between each initial short message in the initial short message set, and classify each initial short message based on each first similarity and clustering algorithm to obtain each short message group; generate a representative short message corresponding to each short message group according to a preset text summary model, and calculate the second similarity between the representative short message and each initial short message in the short message group; according to each second similarity and each first similarity, select the target short message from each initial short message included in the short message group, and send the target short message to the target object. By adopting this method, each initial short message is classified based on each first similarity to obtain short message groups of various types, and by determining the second similarity between the representative short message in each short message group and each initial short message, the similarity of the short message content is clarified. Then, the target short message is selected based on the first similarity and the second similarity, and the target short message is sent to the target object, so as to realize automatic filtering of similar short messages, avoid repeated sending of similar short messages, and improve the utilization rate of communication resources. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0049] Figure 1 A schematic diagram of a flow chart of a method for sending information in an embodiment;
[0050] Figure 2 A schematic diagram of a flow chart of a step of determining a first similarity in one embodiment;
[0051] Figure 3 is a workflow diagram of similarity analysis in one embodiment;
[0052] Figure 4 A schematic diagram of a process for determining a text message group in one embodiment;
[0053] Figure 5 A workflow diagram of cluster analysis in one embodiment;
[0054] Figure 6 is a schematic diagram of a process of determining a second similarity in one embodiment;
[0055] Figure 7 A schematic diagram of a process of screening target text messages in one embodiment;
[0056] Figure 8 A flowchart of the initial SMS screening process in one embodiment;
[0057] Fig. 9 A schematic diagram of a process of screening initial target text messages in one embodiment;
[0058] Fig.10 A schematic diagram of a process of sending a target SMS in one embodiment;
[0059] Fig.11 This is an application environment diagram of a marketing SMS screening system in one embodiment;
[0060] Fig.12 A schematic diagram of the structure of a marketing SMS screening system in one embodiment;
[0061] Fig.13 is a structural block diagram of an information sending device in an embodiment;
[0062] Fig.14 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0063] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0064] In one embodiment, Figure 1 As shown, a method for sending information is provided. The embodiment of the present application takes the method applied to a computer device as an example for explanation. The embodiment of the present application does not limit the execution device of the method for sending information, and includes the following steps 102 to 106:
[0065] Step 102, determining the first similarities between the initial short messages in the initial short message set, and classifying the initial short messages based on the first similarities and a clustering algorithm to obtain short message groups.
[0066] Each SMS group includes initial SMS messages of the same type.
[0067] In implementation, the computer device intercepts each pre-sent initial text message and constructs an initial text message set based on each initial text message. Then, for each initial text message in the initial text message set, the computer device constructs the initial text message and other initial text messages in the initial text message set into a text message pair to obtain each text message pair. The computer device determines the first similarity between the initial text messages in each text message pair based on a preset semantic similarity model. Then, the computer device constructs each first similarity into a similarity matrix, and classifies each initial text message based on the similarity matrix and a clustering algorithm to obtain each text message group.
[0068] Step 104: Generate a representative SMS corresponding to each SMS group according to a preset text summary model, and calculate a second similarity between the representative SMS and each initial SMS in the SMS group.
[0069] In implementation, a text summary model is pre-set in the computer device. For each SMS group, the computer device extracts text features in the SMS group according to the preset text summary model to obtain a representative SMS corresponding to the SMS group. Then, the computer device calculates the second similarity between the representative SMS and each initial SMS in the SMS group according to the preset semantic similarity model.
[0070] Step 106 , screening target text messages from the initial text messages included in the text message group according to the second similarities and the first similarities, and sending the target text messages to the target object.
[0071] In implementation, a first similarity threshold and a second similarity threshold are pre-set in the computer device. The computer device selects the initial target SMS from the initial SMS included in the SMS group according to the second similarities and the second similarity threshold. Then, the computer device selects the target SMS from the initial target SMS based on the first similarity threshold and the first similarity corresponding to the initial target SMS. Then, the computer device sends the target SMS to the target object.
[0072] In the above information sending method, each initial short message is classified based on each first similarity to obtain each type of short message group, and by determining the second similarity between the representative short message in each short message group and each initial short message, the similarity of the short message content is clarified. Then, the target short message is screened based on the first similarity and the second similarity, and the target short message is sent to the target object, so that the automatic filtering of similar short messages is realized, the repeated sending of similar short messages is avoided, and the utilization rate of communication resources is improved.
[0073] In an exemplary embodiment, Figure 2 As shown, the specific process of determining the first similarity between the initial short messages in the initial short message set in step 102 includes steps 202 to 206. Among them:
[0074] Step 202: intercept each pre-sent initial short message according to a preset time period, and number each initial short message to obtain an initial short message set.
[0075] In implementation, a time period is pre-set in the computer device. The computer device intercepts each initial SMS pre-sent by the marketing system to the target object according to the preset time period. Then, the computer device numbers the initial SMS according to the trigger time of the initial SMS (the sending time of the marketing system), and constructs an initial SMS set according to each numbered initial SMS. The SMS numbers are 1, 2, ..., m (m is the number of initial SMS in the initial SMS set).
[0076] Optionally, the frequency of sending the initial SMS varies due to different business stages, business scenarios, etc. Therefore, the option of setting a common time range setting in the computer device supports customizing the time period of SMS interception according to business needs.
[0077] Step 204: for each initial short message in the initial short message set, the initial short message and each of the other initial short messages are constructed into a short message pair according to the serial number of the initial short message.
[0078] In implementation, the computer device constructs a message pair for each initial message in the initial message set according to the initial message number, the initial message and each initial message in other initial messages, wherein the other initial messages are the remaining initial messages in the initial message set except the initial message.
[0079] In an exemplary embodiment, the computer device pairs the initial text messages in the initial text message set according to the text message numbers to form corresponding numbers 1-2, 1-3, ..., 1-m; 2-3, 2-4, ..., 2-m; ...; (m-1)-m. Group SMS pairs. Where m is the number of initial SMS messages.
[0080] Step 206 , performing similarity calculation on each SMS pair according to a preset semantic similarity model to obtain a first similarity between initial SMS messages in the SMS pair.
[0081] The semantic similarity model is the SBERT model (Siamese BERT, a semantic similarity model for natural language processing). The SBERT model is a generalization of BERT (Bidirectional Encoder Representation) on sentence vectors. The pre-trained BERT is modified to use the Siamese and Triplet network structures to obtain semantically meaningful sentence embeddings, thereby obtaining fixed-length sentence embeddings. Each SMS pair includes a first initial SMS and a second initial SMS.
[0082] In the implementation, the computer device inputs each SMS pair into the SBERT model, and performs vector conversion on the SMS pair through the SBERT model to obtain a first initial vector corresponding to the first initial SMS and a second initial vector corresponding to the second initial SMS. Then, the computer device performs similarity calculation on the first initial vector and the second initial vector according to a preset similarity algorithm to obtain a first similarity between the initial SMS in the SMS pair. The similarity algorithm is shown in the following formula (1):
[0083] (1)
[0084] In the above formula (1), is the first similarity, for function, which normalizes the numerical vector into a probability distribution vector, and the sum of each probability is 1; is the weight parameter, n represents the dimension of the vector, and k represents the number of labels. is the first initial vector, is the second initial vector.
[0085] In an exemplary embodiment, Figure 3 The workflow diagram for similarity analysis. Figure 3 A method for performing similarity analysis on an initial short message is provided, comprising:
[0086] Step 301, setting a time period.
[0087] Step 302, intercepting the initial short messages triggered by the system within the selected time period, and numbering them, and then collecting the numbered initial short messages into an initial short message set.
[0088] Step 303, pairing the initial SMS messages in pairs according to their serial numbers to obtain SMS pairs.
[0089] Step 304: select SMS pairs from the initial SMS set and input them into a preset semantic similarity model.
[0090] Step 305: pre-train the first BERT model in the semantic similarity model to obtain a first pre-trained BERT model.
[0091] Step 306: pre-train the second BERT model in the semantic similarity model to obtain a second pre-trained BERT model.
[0092] Step 307, modify the first pre-trained BERT model and perform pooling.
[0093] Step 308: Modify the second pre-trained BERT model and perform pooling.
[0094] Step 309 : Convert the first initial SMS in the SMS pair into a first initial vector through the pooled first pre-trained BERT model.
[0095] Step 310: Convert the second initial SMS in the SMS pair into a second initial vector through the pooled second pre-trained BERT model.
[0096] Step 311, performing similarity calculation on the first initial vector and the second initial vector, thereby obtaining the similarity of the short message pair.
[0097] Step 312, similarity calculation is performed on the remaining SMS pairs in the initial SMS set, and this process is repeated to obtain a similarity matrix.
[0098] In this embodiment, by combining the initial text messages in the initial text message set into text message pairs and calculating the first similarity of each text message pair, the similarity of the text message contents between the initial text messages in the text message pair is clarified, which facilitates subsequent screening of the initial text messages.
[0099] In an exemplary embodiment, Figure 4 As shown, the specific processing process of classifying each initial short message based on each first similarity and clustering algorithm in step 102 to obtain each short message group includes steps 402 to 404. Among them:
[0100] Step 402: construct a similarity matrix according to each first similarity, and cluster the similarity matrix based on a clustering algorithm to obtain clusters.
[0101] The clustering algorithm is a K-value clustering algorithm. The larger the K value, the more detailed the classification of the initial SMS set, and the smaller the K value, the coarser the classification of the initial SMS set.
[0102] In implementation, the computer device constructs a similarity matrix based on each first similarity. Then, the computer device inputs the similarity matrix into a K-value clustering algorithm, and performs clustering processing on the similarity matrix through the K-value clustering algorithm to obtain each cluster cluster.
[0103] In an exemplary embodiment, a computer device obtains a K value. The computer device inputs a similarity matrix into a K-value clustering algorithm to calculate the distance between data points. Then, the computer device determines the similarity information between each pair of variables based on the distance between the data points, and performs cluster division. Specifically, the computer device determines the number of clusters K by a preset elbow method, and then randomly selects K cluster centers to assign each data point to a cluster corresponding to the cluster center closest to it. The information in the similarity matrix can be used to evaluate the change in the cluster center. If the numerical value in the similarity matrix does not change much, it can be considered that the cluster allocation has stabilized. At the same time, the computer device allows the user to set the maximum number of iterations by himself, and the iterative process of K-means clustering stops when the maximum number of iterations is reached.
[0104] Optionally, the computer device performs clustering processing on the similarity matrix according to a K-value clustering algorithm that automatically determines the K value to obtain each cluster cluster.
[0105] Optionally, select a suitable K value according to the number of SMS messages in the SMS set and the actual business scenario. The computer device provides options for commonly used K value settings, supporting users to select K values according to business needs.
[0106] Step 404: Classify the initial short messages according to the corresponding relationship between the first similarities in the clusters and the initial short messages to obtain short message groups.
[0107] In implementation, the computer device determines, for each cluster, a text message group corresponding to the cluster according to the corresponding relationship between each first similarity and each first initial text message in the cluster.
[0108] In an exemplary embodiment, Figure 5 Workflow diagram for cluster analysis. Figure 5 A method for performing cluster analysis on a similarity matrix is provided, comprising:
[0109] Step 501, determine the K value according to the number of short messages in the short message set and the requirements of the actual scenario.
[0110] Step 502: Input the similarity matrix into the K-value clustering algorithm.
[0111] Step 503: clustering the similarity matrix using a K-value clustering algorithm to obtain a clustering result.
[0112] Step 504: classify the initial short messages according to the clustering result to obtain short message groups.
[0113] In this embodiment, each initial text message is classified by the first similarity of each text message pair and the clustering method to obtain text message groups of various types, which facilitates the subsequent screening of initial text messages based on the text message groups, reduces processing time, and improves the efficiency of the information sending method.
[0114] In an exemplary embodiment, Figure 6 As shown, the specific processing process of step 104 includes steps 602 to 604. Among them:
[0115] Step 602: for each SMS group, extract text features of the SMS group according to a preset text summary model to obtain a representative SMS corresponding to the SMS group.
[0116] The representative SMS may represent text contents in most of the initial SMS in the SMS group.
[0117] In implementation, a text summary model is pre-set in the computer device. For each SMS group, the computer device extracts the common text features of each initial SMS in the SMS group according to the preset text summary model to obtain a representative SMS corresponding to the SMS group.
[0118] Optionally, the text summary model may be, but is not limited to, GPT-3 (Generative Pre-trained Transformer 3). The embodiments of the present application do not limit the text summary model.
[0119] Step 604: Calculate the similarity between the representative text message and each initial text message in the text message group based on a preset semantic similarity model to obtain a second similarity between each initial text message and the representative text message.
[0120] In implementation, a semantic similarity model is pre-set in the computer device. The semantic similarity model is the same as the semantic similarity model in step 206. The computer device calculates the similarity between each initial SMS in the SMS group and the representative SMS based on the preset semantic similarity model to obtain a second similarity between the initial SMS and the representative SMS.
[0121] In this embodiment, the representative SMS of the SMS group is generated by the text summary model, which can eliminate the uncertainty and subjectivity of randomly selecting the representative SMS. Then, the second similarity between the representative SMS and the initial SMS is determined, and K second similarity matrices are obtained. The second similarity matrix contains the similarity information between all the initial SMS and the representative SMS in the SMS group, which is convenient for subsequent screening of the initial SMS based on the second similarity.
[0122] In an exemplary embodiment, Figure 7 As shown, the specific processing process of selecting the target SMS from the initial SMS included in the SMS group according to the second similarities and the first similarities in step 106 includes steps 702 to 706. Among them:
[0123] Step 702: determining each initial target text message from each initial text message included in the text message group according to a preset second similarity threshold and a second similarity of each initial text message.
[0124] In implementation, a second similarity threshold is preset in the computer device. The computer device determines, from among the initial short messages, each initial target short message whose second similarity is less than the second similarity threshold.
[0125] Step 704: for each initial target text message, determine whether the first similarity between the initial target text message and the remaining initial target text messages is less than a preset first similarity threshold.
[0126] In implementation, a first similarity threshold is preset in the computer device. The computer device determines, for each of the initial target text messages, whether the first similarity between the initial target text message and the remaining initial target text messages in the text message group is less than the preset first similarity threshold.
[0127] In an optional embodiment, if there is a first similarity greater than or equal to a first similarity threshold, the computer device determines that the content of the initial target text message is similar to that of the remaining initial target text messages in the text message group, and the computer device screens out the initial target text message.
[0128] Optionally, the first similarity threshold may be set to 80% or 75%, which is determined according to business requirements. The first similarity threshold is not limited in the embodiment of the present application.
[0129] Step 706: If the first similarity is less than the first similarity threshold, the initial target short message is determined as the target short message.
[0130] In implementation, if the first similarity is less than a first similarity threshold, the computer device determines that the content of the initial target text message is not similar to the content of the remaining initial target text messages in the text message group, and determines the initial target text message as the target text message.
[0131] In an exemplary embodiment, Figure 8 Diagram for the initial SMS screening workflow. Figure 8 Provide a method for initial SMS screening, including:
[0132] Step 801 , based on the generative text summarization technology, generate summary texts that can cover the main points of all initial text messages in the class for each of the K text message groups, as representative text messages of each class.
[0133] Step 802, using the SBERT model again, calculates the similarity between all the initial SMS messages and the representative SMS messages in the K SMS groups.
[0134] Step 803: Set a different threshold (second similarity threshold) according to actual scenario requirements.
[0135] Step 804, determine whether the second similarity between the initial SMS and the representative SMS is less than the second similarity threshold. If the second similarity is less than the second similarity threshold, execute step 805. If the second similarity is greater than or equal to the second similarity threshold, execute step 806.
[0136] Step 805: Send the initial SMS.
[0137] Step 806, delete the initial SMS.
[0138] In this embodiment, the target SMS is determined by each second similarity and each first similarity, thereby realizing the screening of SMS without manual intervention. And by personalizing the first similarity threshold and the second similarity threshold, the needs of users in different scenarios are met, and the scope of use of the information sending method is expanded. In addition, by including the target SMS in the initial SMS, the problem of duplication or redundancy of SMS content, which affects user experience and marketing effect, is solved, communication costs are reduced, and communication efficiency is improved.
[0139] In an exemplary embodiment, Fig. 9As shown, the specific processing process of step 702 includes steps 902 to 904. Among them:
[0140] Step 902: for each initial SMS in the SMS group, determine whether the second similarity of the initial SMS is less than a preset second similarity threshold.
[0141] In implementation, a second similarity threshold is preset in the computer device. For each initial SMS in the SMS group, the computer device determines whether the second similarity of the initial SMS is less than the preset second similarity threshold.
[0142] In an optional embodiment, if the second similarity of the initial text message is greater than the second similarity threshold, the computer device determines that the text message content of the initial text message is similar to the text message content of most of the initial text messages in the text message group, and therefore, the computer device deletes the initial text message.
[0143] Optionally, the second similarity threshold may be set to 80% or 75%, which is determined according to business requirements. The embodiment of the present application does not limit the second similarity threshold.
[0144] Step 904: If the second similarity of the initial text message is less than the second similarity threshold, the initial text message is determined as the initial target text message.
[0145] In implementation, if the second similarity of the initial text message is less than the second similarity threshold, the computer device determines that the text message content of the initial text message is different from the text message content of most of the initial text messages in the text message group, that is, the initial text message contains text message content that other initial text messages in the text message group do not have. The computer device determines the initial text message as the initial target text message.
[0146] In this embodiment, the initial target SMS is screened from the initial SMS by using the second similarity and the second similarity threshold, and an initial target SMS different from the representative SMS of the SMS group is obtained, thereby achieving automatic screening of the initial SMS.
[0147] In an exemplary embodiment, Fig.10 As shown, the specific process of sending the target short message to the target object in step 106 includes steps 1002 to 1006. Among them:
[0148] Step 1002, determining whether there are multiple target text messages in the same text message group.
[0149] In implementation, the computer device determines whether the number of target text messages in the same text message group is greater than or equal to 2. If the number of target text messages in the same text message group is greater than or equal to 2, the computer device determines that there are multiple target text messages in the text message group.
[0150] In an optional embodiment, if the number of target SMS messages in the same SMS group is 1, the computer device directly sends the target SMS message to the target object.
[0151] Step 1004: If there are multiple target SMS messages in the same SMS group, set the sending time of each target SMS message according to a preset time interval.
[0152] The time interval is greater than the time interval threshold.
[0153] In implementation, a time interval is preset in the computer device. If there are multiple target SMS messages in the same SMS group, the computer device sets the sending time of each target SMS message in the SMS group according to the preset time interval, so that the interval between the sending times of the target SMS messages in the SMS group is greater than or equal to the time interval.
[0154] Optionally, the time interval threshold is determined according to business requirements, and the embodiment of the present application does not limit the time interval threshold.
[0155] Step 1006, sending the target SMS to the target object according to the sending time of the target SMS.
[0156] In implementation, the computer device sends the target SMS to the target object through the communication connection according to the sending time of the target SMS.
[0157] In an exemplary embodiment, the computer device sends the target SMS to the operator according to the sending time of the target SMS, and forwards the target SMS to the target object through the operator.
[0158] In this embodiment, the sending time of the target SMS is set according to the time interval, which further avoids the problem of poor user experience due to similar SMS contents, thereby improving the user experience.
[0159] In an exemplary embodiment, a marketing SMS screening system corresponding to the SMS sending method is provided in a computer device. Fig.11 This is the application environment diagram of the marketing SMS screening system. Fig.11 As shown, the marketing SMS screening system is used to screen the SMS to be sent by the bank. That is, the marketing SMS screening system obtains the initial SMS set from the bank customer marketing system, and screens the initial SMS set to obtain the target SMS. Then, the marketing SMS sending system sends the target SMS to the operator. Fig.12 This is a schematic diagram of the marketing SMS screening system. Fig.12As shown, the marketing SMS screening system includes a similarity analysis module, a cluster analysis module and an SMS screening module. The similarity analysis module is used to obtain an initial SMS set, and determine the first similarity between the initial SMS in the initial SMS set based on the SBERT model. Then, the similarity analysis module constructs a similarity matrix based on each first similarity. The cluster analysis module is used to classify each initial SMS based on the similarity matrix and the clustering algorithm to obtain each SMS group. The SMS screening module is used to generate a representative SMS corresponding to each SMS group based on a generative text summary, and calculate the second similarity between the representative SMS and the initial SMS based on the SBERT model. The SMS screening module is also used to determine whether the initial SMS is sent based on a similarity threshold.
[0160] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0161] Based on the same inventive concept, the embodiment of the present application also provides an information sending device for implementing the information sending method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more information sending device embodiments provided below can refer to the limitations on the information sending method above, and will not be repeated here.
[0162] In an exemplary embodiment, Fig.13 As shown, an information sending device 1300 is provided, comprising: a classification module 1301, a calculation module 1302 and a screening module 1303, wherein:
[0163] The classification module 1301 is used to determine the first similarities between the initial short messages in the initial short message set, and classify the initial short messages based on the first similarities and the clustering algorithm to obtain short message groups.
[0164] The calculation module 1302 is used to generate a representative SMS corresponding to each SMS group according to a preset text summary model, and calculate a second similarity between the representative SMS and each initial SMS in the SMS group.
[0165] The screening module 1303 is used to screen target short messages from the initial short messages included in the short message group according to the second similarities and the first similarities, and send the target short messages to the target object.
[0166] In an exemplary embodiment, the classification module 1301 includes a first determination submodule and a first classification submodule. The first determination submodule includes:
[0167] The interception submodule is used to intercept each pre-sent initial short message according to a preset time period, and number each initial short message to obtain an initial short message set.
[0168] The first construction submodule is used for constructing each initial SMS in the initial SMS set into an SMS pair according to the serial number of the initial SMS.
[0169] The first calculation submodule is used to perform similarity calculation on each SMS pair according to a preset semantic similarity model to obtain a first similarity between initial SMS messages in the SMS pair.
[0170] In an exemplary embodiment, the classification module 1301 includes a first determination submodule and a first classification submodule. The first classification submodule includes:
[0171] The clustering submodule is used to construct a similarity matrix according to each first similarity, and cluster the similarity matrix based on a clustering algorithm to obtain each cluster cluster.
[0172] The second classification submodule is used to classify each initial short message according to the corresponding relationship between each first similarity in each cluster and each initial short message to obtain each short message group.
[0173] In an exemplary embodiment, the calculation module 1302 includes:
[0174] The extraction submodule is used to extract text features of each SMS group according to a preset text summary model to obtain a representative SMS corresponding to the SMS group.
[0175] The second calculation submodule is used to calculate the similarity between the representative text message and each initial text message in the text message group based on a preset semantic similarity model to obtain a second similarity between each initial text message and the representative text message.
[0176] In an exemplary embodiment, the screening module 1303 includes a first screening submodule and a first sending submodule. The first screening submodule includes:
[0177] The second determination submodule is used to determine each initial target SMS message among each initial SMS message included in the SMS group according to a preset second similarity threshold value and a second similarity of each initial SMS message.
[0178] The first judgment submodule is used to judge, for each initial target text message, whether a first similarity between the initial target text message and the remaining initial target text messages is less than a preset first similarity threshold.
[0179] The third determination submodule is used to determine the initial target text message as the target text message if the first similarity is less than a first similarity threshold.
[0180] In an exemplary embodiment, the second determining submodule includes:
[0181] The second judgment submodule is used to judge, for each initial SMS in the SMS group, whether the second similarity of the initial SMS is less than a preset second similarity threshold.
[0182] The fourth determination submodule is configured to determine the initial text message as the initial target text message if the second similarity of the initial text message is less than a second similarity threshold.
[0183] In an exemplary embodiment, the screening module 1303 includes a first screening submodule and a first sending submodule. The first sending submodule includes:
[0184] The third judgment submodule is used to judge whether there are multiple target short messages in the same short message group.
[0185] A setting submodule is used to set the sending time of each target SMS according to a preset time interval if there are multiple target SMS in the same SMS group; the time interval is greater than the time interval threshold;
[0186] The second sending submodule is used to send the target SMS to the target object according to the sending time of the target SMS.
[0187] Each module in the above-mentioned information sending device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module above.
[0188] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Fig.14As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for sending information is realized. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.
[0189] Those skilled in the art will understand that Fig.14 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0190] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.
[0191] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.
[0192] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.
[0193] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.
[0194] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0195] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A method for sending information, characterized in that: The method comprises: Determine a first similarity between each initial short message in the initial short message set, and classify each initial short message based on each first similarity and a clustering algorithm to obtain each short message group; Generate a representative SMS corresponding to each SMS group according to a preset text summary model, and calculate a second similarity between the representative SMS and each of the initial SMS in the SMS group; According to each of the second similarities and each of the first similarities, a target text message is screened from each of the initial text messages included in the text message group, and the target text message is sent to a target object.
2. The method according to claim 1, characterized in that: The determining of the first similarity between the initial short messages in the initial short message set includes: According to a preset time period, each pre-sent initial short message is intercepted, and each of the initial short messages is numbered to obtain an initial short message set; For each of the initial short messages in the initial short message set, constructing the initial short message and each of the other initial short messages into a short message pair according to the serial number of the initial short message; The similarity of each of the short message pairs is calculated according to a preset semantic similarity model to obtain a first similarity between the initial short messages in the short message pair.
3. The method according to claim 1, characterized in that The method of classifying the initial short messages based on the first similarities and the clustering algorithm to obtain short message groups includes: constructing a similarity matrix according to each of the first similarities, and clustering the similarity matrix based on a clustering algorithm to obtain clusters; According to the corresponding relationship between each of the first similarities in each of the clusters and each of the initial short messages, each of the initial short messages is classified to obtain each short message group.
4. The method according to claim 1, characterized in that: Generating a representative SMS corresponding to each SMS group according to a preset text summary model, and calculating a second similarity between the representative SMS and each initial SMS in the SMS group, includes: For each of the SMS groups, extract text features of the SMS group according to a preset text summary model to obtain a representative SMS corresponding to the SMS group; The similarity between the representative text message and each of the initial text messages in the text message group is calculated based on a preset semantic similarity model to obtain a second similarity between each of the initial text messages and the representative text message.
5. The method according to claim 1, characterized in that The step of screening target short messages from the initial short messages included in the short message group according to the second similarities and the first similarities includes: Determining each initial target text message from each initial text message included in the text message group according to a preset second similarity threshold and a second similarity of each initial text message; For each of the initial target text messages, determining whether the first similarity between the initial target text message and the remaining initial target text messages is less than a preset first similarity threshold; If the first similarity is less than a first similarity threshold, the initial target short message is determined as a target short message.
6. The method according to claim 5, characterized in that The determining of each initial target text message from each initial text message included in the text message group according to a preset second similarity threshold and a second similarity of each initial text message includes: For each of the initial text messages in the text message group, determining whether the second similarity of the initial text message is less than a preset second similarity threshold; If the second similarity of the initial text message is less than the second similarity threshold, the initial text message is determined as the initial target text message.
7. The method according to claim 1, characterized in that The sending the target short message to the target object includes: Determine whether there are multiple target text messages in the same text message group; If there are multiple target SMS messages in the same SMS group, the sending time of each target SMS message is set according to a preset time interval; the time interval is greater than the time interval threshold; The target short message is sent to the target object according to the sending time of the target short message.
8. An information sending device, characterized in that: The device comprises: A classification module, used for determining a first similarity between each initial short message in the initial short message set, and classifying each initial short message based on each first similarity and a clustering algorithm to obtain each short message group; A calculation module, used for generating a representative SMS corresponding to each SMS group according to a preset text summary model, and calculating a second similarity between the representative SMS and each of the initial SMS in the SMS group; The screening module is used for screening target short messages from the initial short messages included in the short message group according to the second similarities and the first similarities, and sending the target short messages to the target object.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.