Comprehensive short message platform supervision system based on data analysis

Abstract

The invention belongs to the technical field of short message platforms, and particularly relates to a comprehensive short message platform supervision system based on data analysis, which comprises a short message content extraction module, a short message processing module and a short message processing module, the short message content conversion module is used for converting the text information of the sent message content into recognizable text content based on the obtained text information of the sent message content, wherein the recognizable text content is non-punctuation marks and redundant space content; the short message content risk assessment module is used for assessing the familiarity of the sender and the receiver, and the familiarity is a familiarity index of the sender and the receiver; the sensitive keyword quantity is obtained based on the recognizable text content converted by the short message content conversion module, and whether the short message is a harassment short message or not is judged based on the sensitive keyword quantity and the familiarity; and the intelligent interception module is used for intelligently intercepting the harassment short messages based on the short message content risk assessment result.

Description

Technical Field

[0001] The invention belongs to the technical field of SMS platform supervision, and in particular to an integrated SMS platform supervision system based on data analysis. Background Art

[0002] The SMS platform is a platform that connects to the Internet and directly provides SMS ports provided by China Mobile, China Unicom, and China Telecom to realize batch and customized SMS sending to designated numbers. The SMS platform can provide two-way communication between enterprises and customers, internal employees, and suppliers, and become the fastest and most economical communication tool for information exchange. It can be used to send meeting notices, conduct work scheduling and dispatching, realize official document circulation, event notifications, customer contacts, information release and other daily business management work within the enterprise.

[0003] With the widespread use of SMS services, spam and fraudulent SMS are becoming increasingly serious. The current SMS platform supervision system lacks effective data analysis capabilities and is difficult to accurately identify and block spam SMS, resulting in an increased risk of users being harassed and defrauded.

[0004] Moreover, the existing interception system generally directly intercepts any text message that contains sensitive keywords. If the official warning of fraudulent information or some text messages from familiar contacts contain sensitive keywords, they will also be intercepted, affecting the user experience.

[0005] To this end, the present invention provides a comprehensive SMS platform supervision system based on data analysis. Summary of the invention

[0006] In order to make up for the deficiencies of the prior art, at least one technical problem raised in the background technology is solved.

[0007] The technical solution adopted by the present invention to solve the technical problem is: the integrated SMS platform supervision system based on data analysis described in the present invention includes:

[0008] SMS content extraction module: The SMS content extraction module is used to obtain text information of the sent message content;

[0009] SMS content conversion module: The SMS content conversion module converts the acquired text information of the sent message into recognizable text content, wherein the recognizable text content is content without punctuation marks and extra spaces;

[0010] SMS content risk assessment module: The SMS content risk assessment module is used to assess the familiarity between the sender and the receiver, where the familiarity is the familiarity index between the sender and the receiver; and obtain the number of sensitive keywords based on the recognizable text content converted by the SMS content conversion module, and determine whether the SMS is a harassing SMS based on the number of sensitive keywords and the familiarity;

[0011] Intelligent interception module: The intelligent interception module intelligently intercepts harassing text messages based on the risk assessment results of the text message content.

[0012] Preferably, the SMS content risk assessment module assesses whether the SMS has risks, including the following steps:

[0013] S1. After the SMS is converted into recognizable text content by the SMS content conversion module, the recognizable text content is classified and different contents are divided into words to obtain a number of phrases;

[0014] S2. Identify sensitive keywords for the divided phrases, and mark the corresponding SMS as a key SMS based on the number of sensitive keywords. ;

[0015] S3. If the corresponding SMS is a key SMS , then get each key message The corresponding signal code is marked as the sender, and the familiarity is calculated through the sender and the corresponding receiver. and through familiarity To determine the key text messages Whether it can be received, if not, the intelligent interception module outputs an interception SMS signal, otherwise, it is received normally.

[0016] Preferably, the SMS content conversion module converts the SMS into recognizable text content specifically as follows:

[0017] Use regular expressions to convert all the words in the SMS text to lowercase, remove punctuation marks, words and extra spaces. You can use regular expressions to remove punctuation marks and compress extra spaces. Regular expressions are a powerful text processing tool for searching, replacing, checking or parsing strings. They are widely used in text editors to perform complex text operations.

[0018] This is to reduce the noise of feature extraction and make the model more focused on the text content; the continuous text is divided into independent words or vocabulary units through the Chinese word segmentation tool.

[0019] Preferably, based on step S2, the divided phrases are identified as sensitive keywords, and it is determined whether the text message is a key text message. Specifically:

[0020] Step 1: Establish a database: The database is used to store the vocabulary corresponding to various sensitive keywords;

[0021] Step 2: Extract sensitive keywords based on recognizable text content: and obtain the number of sensitive keywords in the text message ;

[0022] Step 3: If the number of sensitive keywords in the text message If the value is greater than the set threshold of sensitive keywords, the corresponding SMS will be marked as a key SMS. If not, the corresponding SMS is marked as a normal receivable SMS; the threshold of sensitive keywords is determined according to the number of words in the SMS content. For example, if there are three sensitive keywords in a 20-word SMS content, it can be marked as a key SMS;

[0023] Step 4: Based on the familiarity between the sender and the receiver Determine the key text messages Whether it can be received, each important text message Familiarity in Compare with the set familiarity threshold. For important SMS with a familiarity level less than the set familiarity threshold, the corresponding SMS is marked as a harassing SMS, and the intelligent interception module outputs an interception SMS signal; otherwise, a signal is generated and transmitted to the SMS content risk assessment module; the familiarity threshold is set based on the number of times the receiver and the sender receive the SMS, and based on the number of times the receiver actively sends the corresponding SMS to the sender.

[0024] Preferably, the word library establishment types corresponding to the various sensitive keywords are as follows:

[0025] The database divides sensitive keywords into three types:

[0026] Political sensitive keywords, commercial advertising sensitive keywords, illegal and irregular sensitive keywords, and user sensitive keywords:

[0027] Based on the type of sensitive keywords, the severity of sensitive keywords is ranked. The severity of political sensitive keywords is greater than that of illegal and irregular sensitive keywords, and the severity of illegal and irregular sensitive keywords is greater than that of commercial advertising sensitive keywords.

[0028] If familiarity When the key SMS is less than the set familiarity threshold and the corresponding SMS is marked as a harassing SMS, the intelligent interception module intercepts it according to the priority of political sensitive keywords, illegal and irregular sensitive keywords, and commercial advertising sensitive keywords;

[0029] If familiarity When the number of key SMS messages is greater than the set familiarity threshold, and there are political sensitive keywords in the SMS, or there are no political sensitive keywords, but the number of illegal and irregular sensitive keywords is greater than the set illegal and irregular sensitive keyword threshold, the intelligent interception module outputs an interception signal;

[0030] If familiarity For SMS messages with a familiarity level greater than the set threshold, and containing only sensitive keywords such as commercial advertisements, the corresponding SMS messages are marked as acceptable for receipt;

[0031] The user-sensitive keywords are: based on the user's search content on the website, if the SMS content contains the same keywords as the user's search content, and the familiarity For an important SMS with a familiarity threshold lower than the set one, when the user-related sensitive keywords are greater than the set threshold, the intelligent interception module outputs an interception signal; specifically, the user's search content on the website is: for example, the user searches for "mobile phone", "home appliances", etc. on the shopping software. If these words exist in the SMS, and there are no political sensitive keywords, commercial advertising sensitive keywords, or illegal and irregular sensitive keywords, and the receiver and the sender are not familiar with each other, then the corresponding SMS is also a harassing SMS.

[0032] Among them, if the sender and the receiver are familiar with each other more, that is, the familiarity is greater than the set threshold, but the text message also contains political sensitive keywords, the text message will be directly intercepted; if the number of illegal and irregular sensitive keywords is greater than the set sensitive keyword threshold, the text message will also be intercepted; if only commercial advertising sensitive keywords exist, the corresponding text message will be marked as receivable.

[0033] Preferably, based on the familiarity in step S3 The specific calculation method is as follows:

[0034] Step 1: Periodically obtain the call record number between the receiver and the sender of the key SMS ; Among them, the periodic acquisition can be once a week;

[0035] Step 2: Periodically obtain the SMS frequency between the receiver and the sender of the key SMS ; Among them, the periodic acquisition can be once a week;

[0036] Step 3: Determine whether the sender is in the recipient's contact list. If so, the familiarity is calculated using the following formula: ;in Number of call records, Frequency of SMS communication, Whether the recipient and sender exist in the contact list, , and is the preset weight coefficient, The value can be 0.135, The value can be 0.069. The value can be 0.796; if the recipient and the sender do not have any contacts, the familiarity is calculated using the following formula: ; and With the preset weight coefficient, The value can be 0.469. The value can be 0.531;

[0037] If familiarity When the number of key SMS messages is greater than the set familiarity threshold and there are no politically sensitive keywords in the SMS messages, the periodic SMS content between the sender and the receiver is extracted. If the number of illegal and irregular sensitive keywords in the SMS content is greater than the set illegal and irregular sensitive keyword threshold, the intelligent interception module outputs an interception signal.

[0038] Preferably, when the intelligent interception module receives an interception signal;

[0039] The integrated SMS platform supervision system further ranks harassing SMS into: severe harassment, moderate harassment, and low harassment based on the number and type of sensitive keywords in the SMS, and based on the fact that among the sensitive keywords, political sensitive keywords are more serious than illegal and irregular sensitive keywords, and illegal and irregular sensitive keywords are more serious than commercial advertising sensitive keywords;

[0040] The intelligent interception module processes the severe disturbance, the moderate disturbance and the low disturbance in sequence according to their priorities.

[0041] Preferably, the harassment level ranking based on SMS is specifically to set up a scoring mechanism to calculate the harassment level score of each SMS according to the number and type of sensitive keywords in the SMS. ;

[0042] The minimum and maximum thresholds for harassment levels are pre-set; the thresholds are set based on the severity of politically sensitive keywords, commercial advertising sensitive keywords, and illegal and irregular sensitive keywords;

[0043] If the harassment level score If the number of SMS messages is less than the minimum threshold of the harassment level, the corresponding SMS messages will be marked as low-level harassment.

[0044] If the harassment level score If the message is between the minimum harassment threshold and the maximum harassment threshold, the corresponding message is marked as moderate harassment;

[0045] If the harassment level score If the value is greater than the maximum threshold of the harassment level, the corresponding text message will be marked as severe harassment.

[0046] Preferably, the harassment level score The specific calculation method is: through the SMS content risk assessment module, the number of political sensitive keywords, commercial advertising sensitive keywords, and illegal and irregular sensitive keywords are obtained respectively, and marked as , and ;

[0047] Harassment level score It can be calculated by the following formula:

[0048] ;

[0049] in , and They are the preset weight coefficients of political sensitive keywords, commercial advertising sensitive keywords, and illegal and irregular sensitive keywords, among which the preset weight coefficient of political sensitive keywords is greater than that of illegal and irregular sensitive keywords, and the preset weight coefficient of illegal and irregular sensitive keywords is greater than that of commercial advertising sensitive keywords.

[0050] Preferably, the SMS content extraction module is electrically connected to the SMS content conversion module, the SMS content conversion module is electrically connected to the SMS content risk assessment module, and the SMS content risk assessment module is electrically connected to the intelligent interception module.

[0051] The beneficial effects of the present invention are as follows:

[0052] 1. The integrated SMS platform supervision system based on data analysis described in the present invention, through the identification of harassing SMS, is not simply to directly determine it as harassing SMS when sensitive keywords are detected, but to determine it based on the familiarity index between the sender and the receiver. If the sender and the receiver are relatively familiar with each other, that is, greater than a set threshold, the SMS is determined to be a normal SMS; if the sender and the receiver are not familiar with each other, that is, less than a set threshold, the SMS is determined to be a harassing SMS; in this way, harassing SMS can be distinguished to the greatest extent, avoiding the problem that the sender cannot receive the SMS due to the presence of sensitive keywords.

[0053] 2. The integrated SMS platform supervision system based on data analysis described in the present invention improves the efficiency and accuracy of SMS supervision by sorting and grading harassing SMS; improves processing efficiency: by sorting harassing SMS, high-risk harassing SMS can be processed first to reduce the interference to users; optimizes resource allocation: operators and service providers can reasonably allocate resources and energy according to the level of harassment and concentrate on dealing with serious harassment problems; enhances user experience: users can identify and block high-risk harassing SMS more quickly, thereby improving the overall communication experience; data-driven decision-making: by analyzing the characteristics and patterns of harassing SMS, data support can be provided for subsequent anti-harassment strategies to help formulate more effective protective measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The present invention will be further described below in conjunction with the accompanying drawings.

[0055] Figure 1 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0056] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0057] The existing interception system generally intercepts text messages directly as long as sensitive keywords are present. If the official warnings are about fraudulent information or some text messages from familiar contacts contain sensitive keywords, they will also be intercepted, affecting the user experience.

[0058] like Figure 1 As shown, an integrated SMS platform supervision system based on data analysis according to an embodiment of the present invention includes an SMS content extraction module, an SMS content conversion module, an SMS content risk assessment module and an intelligent interception module;

[0059] SMS content extraction module: The SMS content extraction module is used to obtain text information of the sent message content;

[0060] SMS content conversion module: The SMS content conversion module converts the acquired text information of the sent message into recognizable text content, wherein the recognizable text content is content without punctuation marks and extra spaces;

[0061] SMS content risk assessment module: The SMS content risk assessment module is used to assess the familiarity between the sender and the receiver, where the familiarity is the familiarity index between the sender and the receiver; and obtain the number of sensitive keywords based on the recognizable text content converted by the SMS content conversion module, and determine whether the SMS is a harassing SMS based on the number of sensitive keywords and the familiarity;

[0062] Intelligent interception module: The intelligent interception module intelligently intercepts harassing text messages based on the risk assessment results of the text message content.

[0063] Specifically, the SMS content conversion module can process a variety of complex information and situations such as inconsistent punctuation, typos, extra words, etc., which are used to circumvent security system screening, so that the subsequent SMS content evaluation module can determine whether the SMS is a harassing SMS, and can quickly and accurately identify and intercept harassing SMS to avoid the risk of users being harassed and defrauded; and the SMS content risk assessment module is used to assess whether the sender is familiar with the message and the number of sensitive keywords, so that the subsequent intelligent interception module can intercept according to the SMS; comparison can help reduce the situation where normal SMS are mistakenly identified as spam, improve the accuracy of the SMS filtering system, and avoid important information being intercepted by mistake.

[0064] The SMS content risk assessment module assesses whether a SMS message has risks by including the following steps:

[0065] S1. After the SMS is converted into recognizable text content through the SMS content conversion module, the recognizable text content is classified and different contents are divided into words to obtain a number of phrases;

[0066] S2. Identify sensitive keywords for the divided phrases, and mark the corresponding SMS as a key SMS based on the number of sensitive keywords. ;

[0067] S3. If the corresponding SMS is a key SMS , then get each key message The corresponding signal code is marked as the sender, and the familiarity is calculated through the sender and the corresponding receiver. and through familiarity To determine the key messages Whether it can be received, if not, the intelligent interception module outputs an interception SMS signal, otherwise, it is received normally;

[0068] Specifically, in the embodiment of the present invention, the identification of harassing text messages is not simply to directly determine it as a harassing text message when sensitive keywords are detected, but to determine it based on the familiarity index between the sender and the receiver. If the sender and the receiver are relatively familiar with each other, that is, greater than a set threshold, the text message is determined to be a normal text message; if the sender and the receiver are not familiar with each other, that is, less than a set threshold, the text message is determined to be a harassing text message; in this way, harassing text messages can be distinguished to the greatest extent, avoiding the problem that the sender cannot receive the text message due to the presence of sensitive keywords.

[0069] The SMS content conversion module converts the SMS into recognizable text content specifically as follows:

[0070] Use regular expressions to convert all the words in the SMS text to lowercase, remove punctuation marks, words and extra spaces. You can use regular expressions to remove punctuation marks and compress extra spaces. Regular expressions are a powerful text processing tool for searching, replacing, checking or parsing strings. They are widely used in text editors to perform complex text operations.

[0071] This is to reduce the noise of feature extraction and make the model more focused on the text content; the continuous text is segmented into independent words or vocabulary units through the Chinese word segmentation tool; this step can be achieved by using Chinese word segmentation tools such as Jieba Word Segmentation, among which Jieba Word Segmentation supports three word segmentation modes: precise mode, full mode and search engine mode. The precise mode attempts to segment the sentence most accurately and is suitable for text analysis; the full mode scans all the words in the sentence that can be formed into words, which is very fast, but cannot resolve ambiguity; the search engine mode re-segments long words on the basis of the precise mode to improve the recall rate, which is suitable for search engine word segmentation.

[0072] Based on step S2, the divided phrases are identified as sensitive keywords, and it is determined whether the text message is a key text message. Specifically:

[0073] Step 1: Establish a database: The database is used to store the vocabulary corresponding to various sensitive keywords;

[0074] Step 2: Extract sensitive keywords based on recognizable text content: and obtain the number of sensitive keywords in the text message ;

[0075] Step 3: If the number of sensitive keywords in the text message If the value is greater than the set threshold of sensitive keywords, the corresponding SMS will be marked as a key SMS. If not, the corresponding SMS is marked as a normal receivable SMS; the threshold of sensitive keywords is determined according to the number of words in the SMS content. For example, if there are three sensitive keywords in a 20-word SMS content, it can be marked as a key SMS;

[0076] Step 4: Based on the familiarity between the sender and the receiver Determine the key text messages Whether it can be received, each important text message Familiarity in Compare with the set familiarity threshold. If the key SMS is less than the set familiarity threshold, the corresponding SMS is marked as a harassing SMS, and the intelligent interception module outputs an interception SMS signal; otherwise, a signal is generated and transmitted to the SMS content risk assessment module; wherein the familiarity threshold is set based on the number of times the receiver and the sender receive the SMS, and based on the number of times the receiver actively sends the corresponding message to the sender;

[0077] Among them, the threshold of sensitive keywords is set based on the type and quantity of sensitive keywords; such setting can prevent the intelligent interception module from intercepting some official warning fraud or activity information;

[0078] Some official warning scams or activity information can be service notification text messages: such as "your account balance change reminder", the text message may contain the bank's official domain name, such as " ", such SMS is related to specific bank services; verification code SMS: such as " The verification code is 123456", such SMS messages are related to specific online services, etc. There are also other SMS messages such as "For more information, visit ” is related to a specific activity.

[0079] The types of word libraries corresponding to the various sensitive keywords are as follows:

[0080] The database divides sensitive keywords into three types:

[0081] Political sensitive keywords, commercial advertising sensitive keywords, illegal and irregular sensitive keywords, and user sensitive keywords:

[0082] Based on the type of sensitive keywords, the severity of sensitive keywords is ranked. The severity of political sensitive keywords is greater than that of illegal and irregular sensitive keywords, and the severity of illegal and irregular sensitive keywords is greater than that of commercial advertising sensitive keywords.

[0083] If familiarity When the key SMS is less than the set familiarity threshold and the corresponding SMS is marked as a harassing SMS, the intelligent interception module intercepts it according to the priority of political sensitive keywords, illegal and irregular sensitive keywords, and commercial advertising sensitive keywords;

[0084] If familiarity When the number of key SMS messages is greater than the set familiarity threshold, and there are political sensitive keywords in the SMS, or there are no political sensitive keywords, but the number of illegal and irregular sensitive keywords is greater than the set illegal and irregular sensitive keyword threshold, the intelligent interception module outputs an interception signal;

[0085] If familiarity For SMS messages with a familiarity level greater than the set threshold, and containing only sensitive keywords such as commercial advertisements, the corresponding SMS messages are marked as acceptable for receipt;

[0086] The user-sensitive keywords are: based on the user's search content on the website, if the SMS content contains the same keywords as the user's search content, and the familiarity For a key SMS with a familiarity level lower than the set threshold, when the user-related sensitive keywords are greater than the set threshold, the intelligent interception module outputs an interception signal; specifically, the user's search content on the website is: for example, the user searches for "mobile phone" or "home appliance" on the shopping software. If these words exist in the SMS, and there are no political sensitive keywords, commercial advertising sensitive keywords, or illegal and irregular sensitive keywords, and the receiver and the sender are not familiar with each other, then the corresponding SMS is also a harassing SMS;

[0087] Among them, sensitive keywords in the political category include "separatism" and "cult", as well as text messages involving illegal activities and "guns". Sensitive keywords in commercial advertisements include "purchase", "promotion", "discount", "free" and "discount". These words often appear in notifications, marketing and verification codes, and are suspected of fraudulent information.

[0088] SPs illegally send SMS messages: These SMS service providers may illegally send SMS messages to mislead users into subscribing to SMS services. The numbers from which these SMS messages are sent are usually SP access codes.

[0089] Sensitive keywords related to illegal activities: may include sensitive keywords such as "black car", and real-time classification and identification of sensitive keywords in text messages, such as "Mark Six Lottery" and "Taobao store"; the domain name in the fraudulent text message may be a mixture of uppercase and lowercase letters and numbers, the domain name ending is incomplete, or it is not the standard domain name ending of the government website (such as .gov.cn); punctuation and typos: the fraudulent text messages may contain inconsistent punctuation, typos, and extra words, which are the means used by scammers to evade security system screening;

[0090] Among them, if the sender and the receiver are familiar with each other more, that is, the familiarity is greater than the set threshold, but the text message also contains political sensitive keywords, the text message will be directly intercepted; if the number of illegal and irregular sensitive keywords is greater than the set sensitive keyword threshold, the text message will also be intercepted; if only commercial advertising sensitive keywords exist, the corresponding text message will be marked as receivable.

[0091] Based on the familiarity in step S3 The specific calculation method is as follows:

[0092] Step 1: Periodically obtain the call record number between the receiver and the sender of the key SMS ; Among them, the periodic acquisition can be once a week;

[0093] Step 2: Periodically obtain the SMS frequency between the receiver and the sender of the key SMS ; Among them, the periodic acquisition can be once a week;

[0094] Step 3: Determine whether the sender is in the recipient's contact list. If so, the familiarity is calculated using the following formula: ;in Number of call records, Frequency of SMS communication, Whether the recipient and sender exist in the contact list, , and is the preset weight coefficient, The value can be 0.135, The value can be 0.069. The value can be 0.796; if the recipient and the sender do not have any contacts, the familiarity is calculated using the following formula: ; and With the preset weight coefficient, The value can be 0.469. The value can be 0.531;

[0095] If familiarity When the number of key SMS messages is greater than the set familiarity threshold, and there are no political sensitive keywords in the SMS messages, the periodic SMS content between the sender and the receiver is extracted. If the number of illegal and irregular sensitive keywords in the SMS content is greater than the set illegal and irregular sensitive keyword threshold, the intelligent interception module outputs an interception signal;

[0096] Specific, familiarity It can be an index based on the call records between the user and the number, the frequency of SMS exchanges, whether the number exists in the contact list of the recipient and the sender, etc. The largest proportion is that if the sender is in the recipient's contact list, it may mean that the recipient and the sender know each other, that is, they are friends or relatives, so the text message may not be a harassing text message. The second largest proportion is that if the sender does not exist in the recipient's contact list, the frequency of SMS communication And the number of call records The familiarity can be calculated, so that specific harassing text messages can be identified more accurately;

[0097] The comparison can help reduce the situation where normal text messages are mistakenly identified as spam text messages, improve the accuracy of the text message filtering system, and avoid important information from being erroneously intercepted from the annotation characterization parameters of the sending signal code corresponding to each important text message; users are usually more willing to receive text messages from familiar contacts. By comparing the sending signal code with the user's familiar parameters, it can ensure that users receive text messages from friends and relatives and frequently used contacts first, thereby improving the user experience.

[0098] When the intelligent interception module receives an interception signal;

[0099] The integrated SMS platform supervision system further ranks harassing SMS into: severe harassment, moderate harassment, and low harassment based on the number and type of sensitive keywords in the SMS, and based on the fact that among the sensitive keywords, political sensitive keywords are more serious than illegal and irregular sensitive keywords, and illegal and irregular sensitive keywords are more serious than commercial advertising sensitive keywords;

[0100] The intelligent interception module processes the severe harassment, moderate harassment and low harassment in order of priority;

[0101] By sorting and grading harassing text messages, the efficiency and accuracy of text message supervision can be improved;

[0102] Improve processing efficiency: By sorting harassing text messages, high-risk harassing text messages can be processed first to reduce the interference to users; optimize resource allocation: operators and service providers can reasonably allocate resources and energy according to the level of harassment and focus on handling serious harassment problems.

[0103] Enhanced user experience: Users can more quickly identify and block high-risk harassment text messages, thereby improving the overall communication experience; Data-driven decision-making: By analyzing the characteristics and patterns of harassment text messages, data support can be provided for subsequent anti-harassment strategies to help formulate more effective protective measures. Legal compliance: Through effective harassment text message management, legal disputes caused by harassment text messages can be reduced and the legitimate rights and interests of users can be protected.

[0104] The specific method of sorting the harassment level based on SMS is to set up a scoring mechanism to calculate the harassment level score of each SMS according to the number and type of sensitive keywords in the SMS. ;

[0105] The minimum and maximum thresholds for harassment levels are pre-set; the thresholds are set based on the severity of politically sensitive keywords, commercial advertising sensitive keywords, and illegal and irregular sensitive keywords;

[0106] If the harassment level score If the number of SMS messages is less than the minimum threshold of the harassment level, the corresponding SMS messages will be marked as low-level harassment.

[0107] If the harassment level score If the message is between the minimum harassment threshold and the maximum harassment threshold, the corresponding message is marked as moderate harassment;

[0108] If the harassment level score If the value is greater than the maximum threshold of the harassment level, the corresponding text message will be marked as severe harassment.

[0109] The harassment rating score The specific calculation method is: through the SMS content risk assessment module, the number of political sensitive keywords, commercial advertising sensitive keywords, and illegal and irregular sensitive keywords are obtained respectively, and marked as , and ;

[0110] Harassment level score It can be calculated by the following formula:

[0111] ;

[0112] in , and They are respectively the preset weight coefficients of political sensitive keywords, commercial advertising sensitive keywords, and illegal and irregular sensitive keywords, among which the preset weight coefficient of political sensitive keywords is greater than that of illegal and irregular sensitive keywords, and the preset weight coefficient of illegal and irregular sensitive keywords is greater than that of commercial advertising sensitive keywords; The value can be 0.805; The value can be 0.123; The value can be 0.072

[0113] Sensitive keyword quantity scoring: For each sensitive keyword in the text message, a score is given based on the number of times it appears in the text message. For example, if each sensitive keyword appears once, a certain score is added.

[0114] Sensitive keyword type weighting: Different types of sensitive keywords may represent different degrees of harassment. Different weights can be set for different types of sensitive keywords. For example, politically sensitive keywords may add 10 points, commercially sensitive keywords add 3 points, and illegal and illegal words add 15 points. With this design, users can more quickly identify and block high-risk harassing text messages, thereby improving the overall communication experience. Data-driven decision-making: By analyzing the characteristics and patterns of harassing text messages, data support can be provided for subsequent anti-harassment strategies to help formulate more effective protective measures. Legal compliance: Through effective harassing text message management, legal disputes caused by harassing text messages can be reduced and the legitimate rights and interests of users can be protected.

[0115] The SMS content extraction module is electrically connected to the SMS content conversion module, the SMS content conversion module is electrically connected to the SMS content risk assessment module, and the SMS content risk assessment module is electrically connected to the intelligent interception module.

[0116] The threshold is set to facilitate comparison. The threshold depends on the amount of sample data and the number of bases set by technicians in this field for each set of sample data. It does not affect the proportional relationship between the parameter and the quantized value.

[0117] The size of the coefficient is to quantify each parameter to obtain a specific value, which is convenient for subsequent comparison. The size of the coefficient depends on the amount of sample data and the preliminary setting of the corresponding operating coefficient for each set of sample data by the technicians in this field; as long as it does not affect the proportional relationship between the parameter and the quantified value. The above formulas are obtained by collecting a large amount of data for software simulation and selecting a formula close to the true value. The coefficients in the formula are set by technicians in this field according to actual conditions. The above is only a preferred specific implementation of the present invention, but the protection scope of the present invention is not limited to this. Any technician familiar with the technical field within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, shall be equivalently replaced or changed, which should be covered within the protection scope of the present invention.

Claims

1. A comprehensive SMS platform supervision system based on data analysis, characterized by: include SMS content extraction module: The SMS content extraction module is used to obtain text information of the sent message content; SMS content conversion module: The SMS content conversion module converts the acquired text information of the sent message into recognizable text content, wherein the recognizable text content is content without punctuation marks and extra spaces; SMS content risk assessment module: The SMS content risk assessment module is used to assess the familiarity between the sender and the receiver, where the familiarity is the familiarity index between the sender and the receiver; and obtain the number of sensitive keywords based on the recognizable text content converted by the SMS content conversion module, and determine whether the SMS is a harassing SMS based on the number of sensitive keywords and the familiarity; Intelligent interception module: The intelligent interception module intelligently intercepts harassing text messages based on the risk assessment results of the text message content.

2. According to claim 1, a comprehensive SMS platform supervision system based on data analysis is characterized in that: The SMS content risk assessment module assesses whether a SMS message has risks by including the following steps: S1. After the SMS is converted into recognizable text content by the SMS content conversion module, the recognizable text content is classified and different contents are divided into words to obtain a number of phrases; S2. Identify sensitive keywords for the divided phrases, and mark the corresponding SMS as a key SMS based on the number of sensitive keywords. ; S3. If the corresponding SMS is a key SMS , then get each key message The corresponding signal code is marked as the sender, and the familiarity is calculated through the sender and the corresponding receiver. and through familiarity To determine the key text messages Whether it can be received, if not, the intelligent interception module outputs an interception SMS signal, otherwise, it is received normally.

3. The integrated SMS platform monitoring system based on data analysis according to claim 2 is characterized by: The SMS content conversion module converts the SMS into recognizable text content specifically as follows: Use regular expressions to convert all the words in the SMS text into lowercase, remove punctuation marks, words and extra spaces; use the Chinese word segmentation tool to split the continuous text into independent words or vocabulary units.

4. The integrated SMS platform monitoring system based on data analysis according to claim 2 is characterized by: Based on step S2, the divided phrases are identified as sensitive keywords, and it is determined whether the text message is a key text message. Specifically: Step 1: Establish a database: The database is used to store the vocabulary corresponding to various sensitive keywords; Step 2: Extract sensitive keywords based on recognizable text content: and obtain the number of sensitive keywords in the text message ; Step 3: If the number of sensitive keywords in the text message If the value is greater than the set threshold of sensitive keywords, the corresponding SMS will be marked as a key SMS. If not, the corresponding SMS is marked as a normal receivable SMS; the threshold of sensitive keywords is determined according to the number of words in the SMS content. For example, if there are three sensitive keywords in a 20-word SMS content, it can be marked as a key SMS; Step 4: Based on the familiarity between the sender and the receiver Determine the key text messages Whether it can be received, each important text message Familiarity in Compare with the set familiarity threshold. For important text messages with a familiarity level less than the set threshold, the corresponding text messages are marked as harassing text messages, and the intelligent interception module outputs an interception signal; otherwise, a signal is generated and transmitted to the text message content risk assessment module.

5. The integrated SMS platform monitoring system based on data analysis according to claim 4 is characterized by: The types of word libraries corresponding to the various sensitive keywords are as follows: The database divides sensitive keywords into three types: Political sensitive keywords, commercial advertising sensitive keywords, illegal and irregular sensitive keywords, and user sensitive keywords: Based on the type of sensitive keywords, the severity of sensitive keywords is ranked. The severity of political sensitive keywords is greater than that of illegal and irregular sensitive keywords, and the severity of illegal and irregular sensitive keywords is greater than that of commercial advertising sensitive keywords. If familiarity When the key SMS is less than the set familiarity threshold and the corresponding SMS is marked as a harassing SMS, the intelligent interception module intercepts it according to the priority of political sensitive keywords, illegal and irregular sensitive keywords, and commercial advertising sensitive keywords; If familiarity When the number of key SMS messages is greater than the set familiarity threshold, and there are political sensitive keywords in the SMS, or there are no political sensitive keywords, but the number of illegal and irregular sensitive keywords is greater than the set illegal and irregular sensitive keyword threshold, the intelligent interception module outputs an interception signal; If familiarity For SMS messages with a familiarity level greater than the set threshold, and containing only sensitive keywords such as commercial advertisements, the corresponding SMS messages are marked as acceptable for receipt; The user-sensitive keywords are: based on the user's search content on the website, if the SMS content contains the same keywords as the user's search content, and the familiarity For important SMS messages with a familiarity level lower than the set threshold, when the user-type sensitive keywords are greater than the set threshold, the intelligent interception module outputs an interception signal; If the sender and the receiver are familiar with each other, that is, the familiarity is greater than the set threshold, but the text message also contains political sensitive keywords, the text message will be directly blocked; if the number of illegal and irregular sensitive keywords is greater than the set sensitive keyword threshold, the text message will also be blocked; If only commercial advertising sensitive keywords exist, the corresponding SMS will be marked as receivable.

6. The integrated SMS platform monitoring system based on data analysis according to claim 5 is characterized by: Based on the familiarity in step S3 The specific calculation method is as follows: Step 1: Periodically obtain the call record number between the receiver and the sender of the key SMS ; Step 2: Periodically obtain the SMS frequency between the receiver and the sender of the key SMS ; Step 3: Determine whether the sender is in the recipient's contact list. If so, the familiarity is calculated using the following formula: ;in Number of call records, Frequency of SMS communication, Whether the recipient and sender exist in the contact list, , and is the preset weight coefficient; if the recipient and the sender do not have any contacts, the familiarity is calculated using the following formula: ; and With preset weight coefficient; If familiarity When the number of key SMS messages is greater than the set familiarity threshold and there are no politically sensitive keywords in the SMS messages, the periodic SMS content between the sender and the receiver is extracted. If the number of illegal and irregular sensitive keywords in the SMS content is greater than the set illegal and irregular sensitive keyword threshold, the intelligent interception module outputs an interception signal.

7. The integrated SMS platform monitoring system based on data analysis according to claim 1 is characterized by: When the intelligent interception module receives an interception signal; The integrated SMS platform supervision system further ranks harassing SMS into: severe harassment, moderate harassment, and low harassment based on the number and type of sensitive keywords in the SMS, and based on the fact that among the sensitive keywords, political sensitive keywords are more serious than illegal and irregular sensitive keywords, and illegal and irregular sensitive keywords are more serious than commercial advertising sensitive keywords; The intelligent interception module processes the severe disturbance, the moderate disturbance and the low disturbance in sequence according to their priorities.

8. The integrated SMS platform monitoring system based on data analysis according to claim 7 is characterized by: The specific method of sorting the harassment level based on SMS is to set up a scoring mechanism to calculate the harassment level score of each SMS according to the number and type of sensitive keywords in the SMS. ; Pre-set the minimum and maximum harassment thresholds; The threshold is set based on the severity of political sensitive keywords, commercial advertising sensitive keywords, and illegal and irregular sensitive keywords; If the harassment level score If the number of SMS messages is less than the minimum threshold of the harassment level, the corresponding SMS messages will be marked as low-level harassment. If the harassment level score If the message is between the minimum harassment threshold and the maximum harassment threshold, the corresponding message is marked as moderate harassment; If the harassment level score If the value is greater than the maximum threshold of the harassment level, the corresponding text message will be marked as severe harassment.

9. The integrated SMS platform monitoring system based on data analysis according to claim 8 is characterized by: The harassment rating score The specific calculation method is: through the SMS content risk assessment module, the number of political sensitive keywords, commercial advertising sensitive keywords, and illegal and irregular sensitive keywords are obtained respectively, and marked as , and ; Harassment level score It can be calculated by the following formula: ； in , and They are the preset weight coefficients of political sensitive keywords, commercial advertising sensitive keywords, and illegal and irregular sensitive keywords, among which the preset weight coefficient of political sensitive keywords is greater than that of illegal and irregular sensitive keywords, and the preset weight coefficient of illegal and irregular sensitive keywords is greater than that of commercial advertising sensitive keywords.

10. The integrated SMS platform monitoring system based on data analysis according to claim 1 is characterized by: The SMS content extraction module is electrically connected to the SMS content conversion module, the SMS content conversion module is electrically connected to the SMS content risk assessment module, and the SMS content risk assessment module is electrically connected to the intelligent interception module.

Images