Phishing email detection method and system

By using the ecological law characteristic database of enterprise emails and preset phishing email types, we automatically detect the characteristic values ​​and scores of the mail to be detected, solving the problems of low accuracy and high workload in the existing technology, and achieving high efficiency, Accurate phishing email detection.

CN113240297BActive Publication Date: 2024-11-29TSINGHUA UNIVERSITY +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202110548189.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-05-19
Publication Date
2024-11-29
Estimated Expiration
2041-05-19

AI Technical Summary

Technical Problem

The existing phishing email detection methods have problems such as missed and false alarms, resulting in large workloads from email security administrators and low detection accuracy.

Method used

By obtaining the ecological law characteristic database of enterprise emails and the preset phishing email types, obtain the characteristic values ​​of the mail to be detected in the target email ecological law group, calculate the score and sort it to achieve automatic detection of phishing emails.

Benefits of technology

It realizes automated detection of possible phishing emails in the enterprise email system, reduces the workload of email security administrators and improves the accuracy of detection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113240297B_ABST
    Figure CN113240297B_ABST
Patent Text Reader

Abstract

The present invention provides a phishing email detection method and system, the method comprising: obtaining emails to be detected in a preset period; obtaining each feature value of the email to be detected in a target email ecological law group according to an ecological law feature database of corporate emails and a preset phishing email type, the ecological law feature database comprising an incoming email ecological feature database and a outgoing email ecological feature database, the target email ecological law group being determined according to the preset phishing email type; obtaining the score of the email to be detected according to the feature value, sorting the score, and obtaining the phishing email detection result of the email to be detected. The present invention realizes automatic detection of phishing emails that may exist in the corporate email system, reduces the workload of email security administrators, and improves the accuracy of the enterprise in detecting phishing emails.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of Internet mail transmission and processing, and in particular to a phishing mail detection method and system. Background Art

[0002] As one of the most common applications of the Internet, email is an indispensable part of modern people's work and life.

[0003] However, the security of emails is currently facing security challenges that cannot be ignored. Among them, phishing email attacks against enterprises have a great impact on the normal operation of enterprises and caused certain economic losses. Existing phishing email detection methods have defects in different aspects. Some methods are prone to miss phishing emails, posing a threat to enterprise security; some methods will generate a large number of false positive phishing emails, consuming the energy of enterprise security administrators. Existing technologies make the workload of email security administrators heavy, and the monitoring methods for phishing emails are less accurate.

[0004] Therefore, there is an urgent need for a phishing email detection method and system to solve the above problems. Summary of the invention

[0005] In view of the problems existing in the prior art, the present invention provides a phishing email detection method and system.

[0006] The present invention provides a phishing email detection method, comprising:

[0007] Get the emails to be tested in the preset period;

[0008] According to the ecological law feature database of corporate emails and the preset phishing email type, each feature value of the email to be detected in the target email ecological law group is obtained, the ecological law feature database includes a received email ecological feature database and a sent email ecological feature database, and the target email ecological law group is determined according to the preset phishing email type;

[0009] According to the characteristic value, the score of the email to be detected is obtained, and the score is sorted to obtain the phishing email detection result of the email to be detected.

[0010] According to a phishing email detection method provided by the present invention, before obtaining each feature value of the email to be detected in the target email ecological law group according to the ecological law feature database of corporate emails and the preset phishing email type, the method further includes:

[0011] Get sample corporate emails for the target period;

[0012] Based on the email logs of sample corporate emails, statistics are collected on the sample corporate emails to obtain ecological law key-value pairs, and an ecological law feature database is constructed based on the ecological law key-value pairs.

[0013] According to a phishing email detection method provided by the present invention, before the email log based on the sample corporate emails is used to count the sample corporate emails and obtain the ecological law characteristic value, the method further includes:

[0014] According to the field classification record of each email, the email log of the sample corporate email is obtained. The field classification record includes the recipient's email address, the sender's email address, the sender's name, the corporate user's login IP, the region where the corporate user's IP address is located, the timestamp, the URL that appears in the email body, the sending and receiving flag field, and the email classification algorithm classification field.

[0015] According to a phishing email detection method provided by the present invention, the email log based on the sample enterprise emails is used to count the sample enterprise emails, obtain ecological law key-value pairs, and construct an ecological law feature database based on the ecological law key-value pairs, including:

[0016] Obtain sender and receiver information according to the email log of the sample enterprise email;

[0017] The sample corporate email is judged according to the sender and receiver information. If it is judged that the sample corporate email is an email sent by a corporate mailbox, according to the preset rules corresponding to each type of key-value pair, the key-value pairs corresponding to the sender information of the sample corporate email, the key-value pairs corresponding to the URL appearing in the email body, and the key-value pairs corresponding to the number of emails sent at the target time are counted to obtain a first ecological law feature value, and a sent email ecological law feature library is constructed according to the first ecological law feature value;

[0018] If it is determined that the sample corporate email is an email received by a corporate mailbox, the key-value pairs corresponding to all field classification records are counted according to the preset rules corresponding to each type of key-value pair, a second ecological law characteristic value is obtained, and a received email ecological law characteristic database is constructed according to the second ecological law characteristic value;

[0019] An ecological law feature database is obtained based on the ecological law feature database of sent emails and the ecological law feature database of received emails.

[0020] According to a phishing email detection method provided by the present invention, obtaining the score of the email to be detected according to the characteristic value, and sorting the score to obtain the phishing email detection result of the email to be detected, including:

[0021] S1. Obtain each feature value of the first mail to be detected in the target ecological law group, and each feature value of each second mail to be detected in the target ecological law group; the first mail to be detected is any mail other than the scored mails in the mails to be detected in the current scoring round; the second mail to be detected is any mail other than the first mail to be detected in the mails to be detected in the current scoring round;

[0022] S2, comparing each feature value of the first mail to be detected with each feature value of the second mail to be detected in sequence, and after the first mail to be detected is compared with all the second mails to be detected, obtaining a score of the first mail to be detected according to the number of feature value items in the first mail to be detected that meet a preset comparison condition;

[0023] S3, repeating steps S1 to S3 to obtain the score of each email in the email to be detected;

[0024] S4. Sort the score of each email in the emails to be detected, and obtain the phishing email detection result of the emails to be detected according to the sorting result.

[0025] According to a phishing email detection method provided by the present invention, after obtaining the emails to be detected in a preset period, the method further includes:

[0026] The mail to be detected is subjected to data cleaning processing, wherein the data cleaning includes type exclusion processing, garbled code processing and URL field cleaning processing.

[0027] The present invention also provides a phishing email detection system, comprising:

[0028] The module for acquiring mails to be detected is used to acquire mails to be detected in a preset period;

[0029] An ecological law characteristic value acquisition module is used to acquire each characteristic value of the to-be-detected email in the target email ecological law group according to the ecological law characteristic database of the corporate email and the preset phishing email type, wherein the ecological law characteristic database includes the ecological characteristic database of the received email and the ecological characteristic database of the sent email;

[0030] The scoring and sorting module is used to obtain the scores of the emails to be detected according to the characteristic values, and sort the scores to obtain the phishing email detection results of the emails to be detected.

[0031] According to a phishing email detection system provided by the present invention, the system further includes:

[0032] Sample enterprise email acquisition module, used to acquire sample enterprise emails in the target period;

[0033] The ecological law characteristic database construction module is used to collect statistics on the sample enterprise emails based on the email logs of the sample enterprise emails, obtain ecological law key-value pairs, and construct an ecological law characteristic database based on the ecological law key-value pairs.

[0034] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any of the above-mentioned phishing email detection methods are implemented.

[0035] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above-mentioned phishing email detection methods are implemented.

[0036] The phishing email detection method and system provided by the present invention determine the target email ecological law group by pre-setting the phishing email type, obtain each feature value of the email to be detected in the target email ecological law group according to the ecological law feature database of the corporate email, and then obtain the score of the email to be detected according to the feature value, and perform score sorting to obtain the detection result of the phishing email, thereby realizing automatic detection of phishing emails that may exist in the corporate email system, reducing the workload of email security administrators, and improving the accuracy of the company's detection of phishing emails. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 A schematic diagram of the process of the phishing email detection method provided by the present invention;

[0039] Figure 2 A schematic diagram of the flow of the ecological law algorithm for extracting corporate emails provided by the present invention;

[0040] Figure 3 A schematic diagram of the flow chart of the real-time to-be-detected email alarm algorithm provided by the present invention;

[0041] Figure 4 A schematic diagram of the structure of the phishing email detection system provided by the present invention;

[0042] Figure 5 This is a schematic structural diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] For an enterprise, its business content, operation mode, inter-enterprise relationship network, employee work mode and internal relationship tend to remain stable over a period of time. These factors constitute the unique corporate ecological law of an enterprise and have a profound impact on the relevant laws of corporate emails.

[0045] Specifically, the business content, corporate network, and internal relationships of employees will affect the email sending and receiving addresses, sender name, etc., and will also affect the URL contained in the body of the email; the company's operating model and employee work model will affect the email sending time, client login time, client login IP address and login location, etc. Once the ecological rules of the company over a period of time can be extracted, when a new batch of emails is received in real time, the threat level of this batch of emails can be judged.

[0046] Figure 1 A schematic diagram of the process of the phishing email detection method provided by the present invention is shown in FIG. Figure 1 As shown, the present invention provides a phishing email detection method, comprising:

[0047] Step 101: Obtain emails to be detected within a preset period.

[0048] In the present invention, the preset period can be designated as a period value t, and the mails to be detected in the preset period are obtained, that is, at the end of each period, all the real-time mails to be detected within the period value t are obtained.

[0049] Optionally, the period value t may be 1 hour or 5 hours, which is not specifically limited here.

[0050] Step 102, based on the ecological law feature database of corporate emails and the preset phishing email type, obtain each feature value of the email to be detected in the target email ecological law group, the ecological law feature database includes an ecological feature database for received emails and an ecological feature database for sent emails, and the target email ecological law group is determined based on the preset phishing email type.

[0051] In the present invention, the preset phishing email types include lateral penetration phishing emails, forged phishing emails, obfuscated phishing emails, and unknown type phishing emails. According to the preset phishing email types, the target email ecological law group can be determined. Accordingly, the target email ecological law group includes four categories: A, B, C, and D, and the corresponding categories are:

[0052] A: Lateral penetration phishing emails;

[0053] B: Forged phishing email;

[0054] C: Obfuscated phishing email;

[0055] D: Unknown type of phishing email.

[0056] According to the categories of incoming and outgoing mails of enterprise mails, two databases, RCPT_DATABASE and SEND_DATABASE, are obtained respectively. Among them, RCPT_DATABASE is the ecological characteristic database of incoming mails; SEND_DATABASE is the ecological characteristic database of outgoing mails. According to the ecological characteristic database of incoming mails and the ecological characteristic database of outgoing mails, the ecological regularity characteristic database of enterprise mails is constructed.

[0057] Furthermore, by querying the ecological law feature database of corporate emails, the feature value of each type of key-value pair of the email to be detected in the target email ecological law group can be obtained.

[0058] Step 103: Obtain the scores of the emails to be detected according to the characteristic values, sort the scores, and obtain the phishing email detection results of the emails to be detected.

[0059] In the present invention, a mail ecological rule group threat level score for a mail is defined: the initial score is 0, and the mail is compared with every other mail in this batch of real-time received mails. If the mail meets the preset comparison conditions, the score of the mail is +1.

[0060] Specifically, if a sender address or sender name that has never appeared in the enterprise ecology appears in an email, or if the login location or sending time of an enterprise email user violates the enterprise ecology rules when sending emails, these are all factors that increase the threat level of the email. According to the characteristic value of each type of key-value pair in the target email ecology rule group of the email to be detected, the threat level score of the email to be detected is obtained, and then the threat level scores of the emails to be detected are sorted by a sorting method, and the batch of emails with the highest threat level in the real-time emails to be detected are screened out to obtain the phishing email detection results of the emails to be detected.

[0061] The phishing email detection method provided by the present invention determines the target email ecological law group by pre-setting the phishing email type, obtains each feature value of the email to be detected in the target email ecological law group according to the ecological law feature database of the corporate email, and then obtains the score of the email to be detected according to the feature value, and performs score sorting to obtain the detection result of the phishing email, thereby realizing automatic detection of phishing emails that may exist in the corporate email system, reducing the workload of email security administrators, and improving the accuracy of the company's detection of phishing emails.

[0062] On the basis of the above embodiment, before obtaining each feature value of the to-be-detected email in the target email ecological law group according to the corporate email ecological law feature database and the preset phishing email type, the method further includes:

[0063] Get sample corporate emails for the target period;

[0064] Based on the email logs of sample corporate emails, statistics are collected on the sample corporate emails to obtain ecological law key-value pairs, and an ecological law feature database is constructed based on the ecological law key-value pairs.

[0065] In the present invention, the target period can be specified as a period value T (eg, 30 days). At the end of each period, all historical corporate emails within the period value T are obtained as sample corporate emails, thereby extracting the ecological regularity characteristics of corporate emails.

[0066] Furthermore, according to the email log record information of the sample enterprise emails, various features of the sample enterprise emails are counted and statistically analyzed to obtain the ecological law key-value pair type represented by {key:value}, which is specifically represented as follows:

[0067] Type 1: {region:count}, which indicates the number of emails sent from the region in the historical emails with a period value of T;

[0068] Where region represents the region and is of string type; count represents the number of emails sent from the region and is of integer type, with an initial value of 0;

[0069] Type 2: {ip_prefix_16:count}, which indicates the number of emails sent from the client IP address with a 16-bit prefix in the historical emails with a period value of T;

[0070] Among them, ip_prefix_16 represents the IP address prefix, which is 16 bits long and belongs to the string type; count represents the number of emails sent by the IP prefix, with an initial value of 0 and belongs to the integer type;

[0071] Type 3: {ip_prefix_24:count}, which indicates the number of emails sent from the client IP address with a 24-bit prefix in the historical emails with a period value of T;

[0072] Among them, ip_prefix_24 represents the IP address prefix, which is 24 bits long and belongs to the string type; count represents the number of emails sent by the IP prefix, with an initial value of 0 and belongs to the integer type;

[0073] Type 4: {ip:count}, which indicates the number of emails sent by the client IP address corresponding to the complete address in the historical emails with a period value of T;

[0074] Among them, ip represents the IP address, which is a string type; count represents the number of emails sent by the IP address, with an initial value of 0, which is an integer type;

[0075] Type 5: {[fromname,sender_address]:count}, which indicates the number of emails with the same sender name and sender mailbox in the historical emails with a period value of T;

[0076] Among them, [fromname,sender_address] represents the combination of sender name and sender email address. An email can only be matched if both conditions are met at the same time. It is of array type. fromname represents the sender name and is of string type. sender_address represents the sender email address and is of string type. count represents the number of emails sent by the combination of sender name and sender email address. The initial value is 0 and it is of integer type.

[0077] Type 6: {fromname:count}, which indicates the number of weeks in which the sender name appears at least 5 days a week in the historical emails with a period value of T;

[0078] Among them, fromname indicates the sender name, which is a string type; count indicates the number of weeks with at least 5 days in a week, with an initial value of 0, which is an integer type;

[0079] Type 7: {fromname:count}, which indicates the number of days that the sender has sent emails in the historical emails with a period value of T;

[0080] Among them, fromname indicates the sender name, which is a string type; count indicates the number of days that the email has been sent, with an initial value of 0, which is an integer type;

[0081] Type 8: {sender_address:count}, which indicates the number of days that the sender's mailbox has sent emails in the historical emails with a period value of T;

[0082] Among them, sender_address represents the sender's email address, which is a string type; count represents the number of days that emails have been sent, with an initial value of 0, which is an integer type;

[0083] Type 9: {fqdn:count}, which indicates the number of occurrences of the full domain name corresponding to the URL in the email body in the historical emails with a period value of T. When querying, if there are multiple full domain names at the same time, their minimum value is returned;

[0084] Among them, fqdn represents the full domain name, which is a string type; count represents the minimum number of occurrences, the initial value is 0, and it is an integer type;

[0085] Type 10: {fqdn:offset}, which indicates the time difference between the first appearance of the full domain name corresponding to the URL in the email body and the present in the historical emails with a period value of T. When querying, if there are multiple full domain names at the same time, the minimum value of them is returned;

[0086] Among them, fqdn represents the full domain name, which is a string type; offset represents the time difference, the initial value is 0, which is a time type;

[0087] Type 11: {hour, count}, which indicates the number of emails sent or received in the hour of the day in which the emails were sent or received in the historical emails with a period value of T;

[0088] Among them, hour represents an integer from 1 to 24, corresponding to the hour of the day, and is of integer type; count represents the number of occurrences of the corresponding day of the week, with an initial value of 0, and is of integer type;

[0089] Type 12: {weekday, count}, which indicates the number of emails sent or received on the day of the week in which the emails were sent or received in the historical emails with a period value of T;

[0090] Among them, weekday represents an integer 1-7, corresponding to Monday to Sunday, and is of integer type; count represents the number of occurrences of the corresponding day of the week, with an initial value of 0, and is of integer type.

[0091] Furthermore, according to the above 12 types of key-value pairs corresponding to 12 types of ecological law characteristics, an ecological law characteristic database is constructed.

[0092] On the basis of the above embodiment, before collecting statistics on the sample enterprise emails based on the email log of the sample enterprise emails to obtain the ecological law characteristic value, the method further includes:

[0093] According to the field classification record of each email, the email log of the sample corporate email is obtained. The field classification record includes the recipient's email address, the sender's email address, the sender's name, the corporate user's login IP, the region where the corporate user's IP address is located, the timestamp, the URL that appears in the email body, the sending and receiving flag field, and the email classification algorithm classification field.

[0094] In the present invention, for each sample enterprise email e, the email log records the fields of the following 9 elements, including e.rcpt_address, e.sender_address, e.fromname, e.client_ip, e.client_region, e.timestamp, e.urls, e.is_sender, and e.tag, to obtain the email log of the sample enterprise email.

[0095] Among them, rcpt_address represents the recipient's email address, which is a string type; sender_address represents the sender's email address, which is a string type; fromname represents the sender's name, which is a string type; client_ip represents the enterprise user's login IP, which is a string type; client_region represents the region where the enterprise user's IP address is located, which is a string type; timestamp represents the timestamp, which is a time type; urls represents the url in the email body, which stores all the urls that appear in the email, which is an array type; url represents the element in urls, which stores a url that appears in the email body, which is a string type; is_sender represents the sending and receiving flag field, which indicates whether the email is sent from or received by the enterprise mailbox, which is a Boolean type; tag represents the classification field of other email classification algorithms (such as the naive Bayes algorithm and the support vector algorithm), and different values ​​represent different types of emails, which is an integer value type.

[0096] Optionally, the phishing email detection method provided by the present invention can operate together with other email classification methods (such as Naive Bayes algorithm, support vector algorithm), using existing mature spam identification algorithms to exclude spam in the tag optional field and reduce false positives caused by spam.

[0097] Based on the above embodiment, after obtaining the emails to be detected in the target period, the method further includes:

[0098] The sample email is subjected to data cleaning processing, wherein the data cleaning includes type exclusion processing, garbled code processing and URL field cleaning processing.

[0099] In the present invention, each sample email is subjected to data cleaning processing, including:

[0100] Perform type exclusion processing on each sample email (optional step). If the phishing email detection method provided by the present invention works together with other email classification methods, if it is determined to be a spam or promotion type email based on the tag field, the email will be deleted;

[0101] Perform garbled processing on each sample email, check each character in all fields, and delete the character if the Unicode code is not in the visible or meaningful character range (including common languages ​​such as Chinese and English. Other language code ranges can also be added according to the company's business. For example, if the company has Thailand-related business, the Thai Unicode code range can be added, which is 0E00-0E7F).

[0102] The url field of each sample email is cleaned. For each url in the urls field, image link cleaning, unsubscribe link cleaning and rendering file link cleaning are performed respectively. Among them, image link cleaning includes removing regular matching image file suffixes and regular matching base64 represented images; unsubscribe link cleaning includes removing keyword matching results; rendering file link cleaning includes removing regular matching css and js file suffixes.

[0103] On the basis of the above embodiment, the email log based on the sample enterprise emails is used to collect statistics on the sample enterprise emails, obtain ecological law key-value pairs, and construct an ecological law feature database based on the ecological law key-value pairs, including:

[0104] Obtain sender and receiver information according to the email log of the sample enterprise email;

[0105] The sample corporate email is judged according to the sender and receiver information. If it is judged that the sample corporate email is an email sent by a corporate mailbox, according to the preset rules corresponding to each type of key-value pair, the key-value pairs corresponding to the sender information of the sample corporate email, the key-value pairs corresponding to the URL appearing in the email body, and the key-value pairs corresponding to the number of emails sent at the target time are counted to obtain a first ecological law feature value, and a sent email ecological law feature library is constructed according to the first ecological law feature value;

[0106] If it is determined that the sample corporate email is an email received by a corporate mailbox, the key-value pairs corresponding to all field classification records are counted according to the preset rules corresponding to each type of key-value pair, a second ecological law characteristic value is obtained, and a received email ecological law characteristic database is constructed according to the second ecological law characteristic value;

[0107] An ecological law feature database is obtained based on the ecological law feature database of sent emails and the ecological law feature database of received emails.

[0108] In the present invention, the sender and receiver information is obtained from the email log of the sample enterprise email. If the sample enterprise email is judged to be an email sent by the enterprise mailbox, according to the preset rules corresponding to each type of key-value pair, the key-value pairs of type 5, type 6, type 7, type 8, type 9, type 10, type 11, and type 12 are counted and statistically analyzed to obtain the first ecological law characteristic value. Among them, the first ecological law characteristic value represents the ecological law characteristic value composed of the second column (i.e., cout or offset) in the key-value pairs type 5 ({[fromname,sender_address]:count}), type 6 ({fromname:count}), type 7 ({fromname:count}), type 8 ({sender_address:count}), type 9 ({fqdn:count}), type 10 ({fqdn:offset}), type 11 ({hour:count}), and type 12 ({weekday:count}). According to the above-mentioned first ecological law characteristic value and the key-value pair expressed in the form of {key:value}, an ecological law characteristic library for sending emails is constructed.

[0109] Furthermore, if the sample corporate email is judged to be an email received by a corporate mailbox, according to the preset rules corresponding to each type of key-value pair, the key-value pairs of type 1, type 2, type 3, type 4, type 5, type 6, type 7, type 8, type 9, type 10, type 11, and type 12 are counted and statistically analyzed to obtain the second ecological law characteristic value. Among them, the second ecological law characteristic value represents the ecological law characteristic value of the second column (i.e., cout or offset) in the key-value pairs type 1 ({region:count}), type 2 ({ip_prefix_16:count}), type 3 ({ip_prefix_24:count}), type 4 ({ip:count}), type 5 ({[fromname,sender_address]:count}), type 6 ({fromname:count}), type 7 ({fromname:count}), type 8 ({sender_address:count}), type 9 ({fqdn:count}), type 10 ({fqdn:offset}), type 11 ({hour:count}), and type 12 ({weekday:count}). According to the above second ecological law characteristic value and the key-value pairs expressed in the form of {key:value}, a received email ecological law characteristic library is constructed.

[0110] Furthermore, an ecological law characteristic database is constructed based on the ecological law characteristic database of sent emails and the ecological law characteristic database of received emails, so as to facilitate obtaining the target ecological law characteristic value when querying the ecological law characteristic database later.

[0111] Optionally, in the sixth type of key-value pairs, the sender name is counted for no less than 5 days in a week. Once the 5 counts are met, the sender name is marked and the sender name is no longer counted in this week.

[0112] Optionally, in the seventh category of key-value pairs, the sender name is counted no less than once in a day. Once the count is met, the sender name is marked and the sender name is no longer counted within the day.

[0113] Optionally, in the 8th category of key-value pairs, the sender's email address is counted no less than once in a day. Once the count is met, the sender's email address is marked and will not be counted again within the day.

[0114] It should be noted that the preset rules corresponding to each type of key-value pair may be different, and are not specifically limited here.

[0115] In one embodiment, at the end of each period, at a specified period value T (e.g., 30 days), for all emails E sent and received by the enterprise mailbox in this period, the emails are judged. If the email is judged to be sent by the enterprise mailbox, the sent email ecological law feature library is used for query; if the email is judged to be received by the enterprise mailbox, the received email ecological law feature library is used for query, each email is counted, and the enterprise ecological database is updated. The algorithm code is as follows:

[0116] For each email e in E:

[0117] if e.is_sender == False:

[0118] use RCPT_DATABASE; (use means using this database in the following query)

[0119] if e.is_sender == True:

[0120] use SEND_DATABASE;

[0121] Calculate the key-value pairs of type 5, type 6, type 7, type 8, type 9, type 10, type 11, and type 12 of the database respectively:

[0122] Type 5: {[e.fromname,e.sender_address]:count}->{[e.fromname,e.sender_address]:count+1};

[0123] Type 6: If e.fromname appears for at least 5 days in this week, {e.fromname:count}->{e.fromname:count+1}, marking the fromname and not counting it in this week;

[0124] Type 7: If e.fromname appears at least once in the day, {e.fromname:count}->{e.fromname:count+1}, mark the fromname and do not count it in the day;

[0125] Type 8: If e.sender_address appears at least once in the day, {e.sender_address:count}->{e.sender_address:count+1}, mark the sender_address and no longer count it in the day;

[0126] According to e.timestamp;

[0127] Calculate hour, which is the hour in the day corresponding to e.timestamp;

[0128] Calculate weekday, which is the day of the week corresponding to e.timestamp;

[0129] Type 11: {hour, count}->{hour, count+1};

[0130] Type 12: {weekday, count}->{weekday, count+1};

[0131] For each url in e.urls:

[0132] Calculate fqdn, which is the FQDN of the url;

[0133] Calculate new_offset, which is the difference between the timestamp of the first appearance of the fqdn and e.timestamp. If it has never appeared, it is 0;

[0134] Type 9: {fqdn:count}->{fqdn:count+1};

[0135] Type 10: {fqdn:offset}->{fqdn:max(offset,new_offset)};

[0136] if e.is_sender == False:

[0137] use RCPT_DATABASE;

[0138] Update key-value pairs of types 1, 2, 3, and 4 in the database:

[0139] {e.region:count}->{e.region:count+1};

[0140] According to e.ip:

[0141] Calculate ip_prefix_16, which is the first 16 bits of e.ip;

[0142] Calculate ip_prefix_24, which is the first 24 bits of e.ip;

[0143] Type 2: {ip_prefix_16:count}->{ip_prefix_16:count+1};

[0144] Type 3: {ip_prefix_24:count}->{ip_prefix_24:count+1};

[0145] Type 4: {e.ip:count}->{e.ip:count+1}.

[0146] Figure 2 The flowchart of the ecological law algorithm for extracting corporate emails provided by the present invention is as follows: Figure 2 As shown, by reading all historical sample corporate emails within a period T, then performing data cleaning on the sample corporate emails, calculating the corporate email ecological laws, and updating the corporate email ecological law features, an updated corporate email ecological law database is obtained.

[0147] Based on the above embodiment, after obtaining the emails to be detected in the preset period, the method further includes:

[0148] The mail to be detected is subjected to data cleaning processing, wherein the data cleaning includes type exclusion processing, garbled code processing and URL field cleaning processing.

[0149] In the present invention, data cleaning is performed on each email to be detected, including:

[0150] Perform type exclusion processing on each email to be detected (optional step). If the phishing email detection method provided by the present invention works together with other email classification methods, if it is determined to be a spam or promotion type email based on the tag field, the email will be deleted;

[0151] Perform garbled processing on each email to be tested, check each character in all fields, and delete the character if the Unicode code is not in the visible or meaningful character range (including common languages ​​such as Chinese and English, and other language code ranges can also be added according to the company's business);

[0152] The URL field of each email to be detected is cleaned. For each URL in the urls field, the image link cleaning, unsubscribe link cleaning and rendering file link cleaning are performed respectively; among them, the image link cleaning includes removing the regular matching image file suffix and the regular matching base64 represented image; the unsubscribe link cleaning includes removing the keyword matching results; the rendering file link cleaning includes removing the regular matching css and js file suffixes.

[0153] By performing data cleaning processing such as type exclusion, garbled code processing and URL field cleaning on the emails to be detected, some errors in the email log fields can be allowed, the requirements for the email logs are relatively loose, and it is beneficial to improve the accuracy of email detection.

[0154] On the basis of the above embodiment, obtaining the score of the email to be detected according to the feature value, sorting the score, and obtaining the phishing email detection result of the email to be detected includes:

[0155] S1. Obtain each feature value of the first mail to be detected in the target ecological law group, and each feature value of each second mail to be detected in the target ecological law group; the first mail to be detected is any mail other than the scored mails in the mails to be detected in the current scoring round; the second mail to be detected is any mail other than the first mail to be detected in the mails to be detected in the current scoring round;

[0156] S2, comparing each feature value of the first mail to be detected with each feature value of the second mail to be detected in sequence, and after the first mail to be detected is compared with all the second mails to be detected, obtaining a score of the first mail to be detected according to the number of feature value items in the first mail to be detected that meet a preset comparison condition;

[0157] S3, repeating steps S1 to S3 to obtain the score of each email in the email to be detected;

[0158] S4. Sort the score of each email in the emails to be detected, and obtain the phishing email detection result of the emails to be detected according to the sorting result.

[0159] In the present invention, by obtaining each characteristic value of the current mail to be detected in the target ecological law group, and each characteristic value of each other mail in this batch of real-time mails to be detected in the target ecological law group, each characteristic value of the current mail to be detected is then compared with each characteristic value of each other mail in this batch of real-time mails to be detected in turn, that is, every two mails to be detected are compared. If the current mail to be detected meets the number of characteristic value items of the preset comparison conditions, the current mail to be detected is given extra points to obtain the score of the current mail to be detected. After all the mails to be detected are compared with each other, the score of each mail in the mail to be detected is obtained, and then the score of each mail is sorted to obtain the phishing mail detection result of the mail to be detected.

[0160] Optionally, the preset comparison condition may be a dimension that is greater than 70% of the corresponding target email ecological rule group.

[0161] Optionally, the score of each email is sorted, and can be arranged in order from high to low, with the higher the score, the higher the ranking.

[0162] Optionally, in practical applications, the sorting algorithm can be simplified by converting double precision to quad precision (Convert Double to Quad, CDQ for short) and the algorithm complexity is O(nlog^3n).

[0163] Optionally, by sorting the emails, a batch of emails with the highest threat level scores from a batch of real-time emails to be detected are screened out as warning emails, and combined with the preset warning threshold k, the first k emails with the highest threat level scores are obtained, and the output warning emails are sent to the administrator for manual review. For warning emails sent from non-corporate mailboxes, the administrator can manually review them to confirm that they are phishing emails, and then issue a warning to the recipient; for warning emails sent from corporate mailboxes, the administrator can manually review them to confirm whether their corporate mailbox accounts have been stolen and take corresponding measures.

[0164] Based on the above embodiment, the target email ecological rule group includes four categories: A, B, C, and D, and the corresponding database values ​​are:

[0165] A: Lateral penetration phishing emails, including Type 1, Type 2, Type 3, Type 4, Type 9, and Type 10;

[0166] B: Forged phishing emails, including Type 5, Type 6, Type 9, and Type 10;

[0167] C: Obfuscated phishing emails, including Type 7, Type 8, Type 9, and Type 10;

[0168] D: Unknown type phishing emails, including type 9, type 10, type 11, and type 12.

[0169] According to the categories of incoming and outgoing mails of corporate mails, two databases, RCPT_DATABASE and SEND_DATABASE, are obtained respectively. Among them, RCPT_DATABASE is the ecological characteristic database of incoming mails, including key-value pair types: type 1, type 2, type 3, type 4, type 5, type 6, type 7, type 8, type 9, type 10, type 11, type 12; SEND_DATABASE is the ecological characteristic database of outgoing mails, including key-value pair types: type 5, type 6, type 7, type 8, type 9, type 10, type 11, type 12.

[0170] In one embodiment, taking the distinction of type A lateral penetration phishing emails as an example, six key-value pairs of type 1, type 2, type 3, type 4, type 9, and type 10 are involved, with a total of six dimensions. Calculate a batch of real-time emails, assuming that there are three emails a, b, and c, and the initial scores are all 0 points. If a is compared with b, the ecological law feature library is queried for a and b, and each feature value in the ecological law group corresponding to a and each feature value in the ecological law group corresponding to b are obtained, and each feature value in a and b is compared in turn. If a has a smaller feature value count in more than 70% of the dimensions (6*70%=4.2, rounded down to 4, that is, more than 4 dimensions), then the score of email a is increased by one point; email a is continued to be compared with email c, and the ecological law feature library is queried respectively. If a has a smaller count value in more than 4 dimensions, then email a is continued to be increased by one point. Then email b is compared with email c, and so on, to obtain the score of each email, and then all email scores are sorted, and the higher the score, the higher the ranking.

[0171] Figure 3 A flow chart of the real-time email alert algorithm to be detected provided by the present invention is shown in FIG. Figure 3 As shown, within the specified cycle value t, an alarm threshold k is set, and at the end of each cycle, a real-time email alarm is issued, and the k emails with the highest threat level scores are output. Specifically, by reading all real-time emails to be detected within cycle t, data cleaning is performed, and the real-time email enterprise ecological law characteristics are queried from the enterprise email ecological law characteristic database, the threat level of all real-time emails is scored, and an alarm email is output, including the k emails with the highest threat level.

[0172] In one embodiment, a cycle value t (e.g., 1 hour) and an alarm threshold k (e.g., 10 emails) are specified. At the end of each cycle, for the set S of all emails sent and received by the corporate mailbox in this cycle, each feature value corresponding to the email to be detected in the target email ecological law group is obtained by querying the corporate ecological law database. Specifically, if the email to be detected is an email sent from the corporate mailbox, the sent email ecological law feature library is used for query; if the email to be detected is an email received by the corporate mailbox, the received email ecological law feature library is used for query. Based on the feature values ​​obtained by querying the ecological law database, the scores of the emails to be detected are obtained, and the scores of the emails to be detected are sorted. The top k emails are selected as alarm output emails. The specific algorithm code is as follows:

[0173] For each email e in S:

[0174] Calculate the corresponding four email ecological rule groups eA, eB, eC, eD / three email ecological rule groups eB, eC, eD (e.is_sender == True);

[0175] According to e.timestamp:

[0176] Calculate hour, which is the hour in the day corresponding to e.timestamp;

[0177] Calculate weekday, which is the day of the week corresponding to e.timestamp;

[0178] For each url in e.urls:

[0179] Calculate fqdn, which is the FQDN of the url;

[0180] According to e.ip:

[0181] Calculate ip_prefix_16, which is the first 16 bits of e.ip;

[0182] Calculate ip_prefix_24, which is the first 24 bits of e.ip;

[0183] Let the currently calculated ecological law group be Z;

[0184] For features type1, type2, type3, type4, type5, type6, type7, type8, type9, type10, type11, type12:

[0185] If Z contains this feature:

[0186] if e.is_sender == False:

[0187] use RCPT_DATABASE;

[0188] if e.is_sender == True:

[0189] use SEND_DATABASE;

[0190] Query the value count of the corresponding key in the database and add it to the ecological rule group Z;

[0191] For four email ecological rule groups eA, eB, eC, eD / three email ecological rule groups eB, eC, eD (e.is_sender == True):

[0192] Sort all mails;

[0193] Output the first k emails as alerts.

[0194] The phishing email detection method provided by the present invention can effectively help enterprises deal with various phishing email attacks, and mainly has the following advantages: First, based on the ecological laws of enterprise emails, real-time emails are detected, which can more effectively discover sender forgery, obfuscation, lateral penetration, and even unknown attack methods; second, the detection algorithm has loose requirements on email logs. On the one hand, the complete email body and attachment information are not required, which has a certain degree of protection for the privacy of enterprise users. On the other hand, the detection algorithm includes a data cleaning part, which allows some errors in the email log field; third, the detection algorithm runs completely offline, based on the ecological laws of enterprise emails, and does not rely on any third-party detection data, which is conducive to protecting enterprise data security.

[0195] Figure 4 A schematic diagram of the structure of the phishing email detection system provided by the present invention is shown in FIG. Figure 4 As shown, the present invention provides a phishing email detection system, including an email acquisition module 401 for detecting, an ecological law characteristic value acquisition module 402 and a scoring and sorting module 403, wherein the email acquisition module 401 for detecting is used to obtain emails to be detected in a preset period; the ecological law characteristic value acquisition module 402 is used to obtain each characteristic value of the email to be detected in the target email ecological law group according to the ecological law characteristic database of corporate emails and preset phishing email types, and the ecological law characteristic database includes an ecological characteristic database of received emails and an ecological characteristic database of sent emails; the scoring and sorting module 403 is used to obtain the score of the email to be detected according to the characteristic value, and sort the score to obtain the phishing email detection result of the email to be detected.

[0196] The phishing email detection system provided by the present invention determines the target email ecological law group by pre-setting the phishing email type, obtains each feature value of the email to be detected in the target email ecological law group according to the ecological law feature database of the corporate email, and then obtains the score of the email to be detected according to the feature value, and performs score sorting to obtain the detection result of the phishing email, thereby realizing automatic detection of phishing emails that may exist in the corporate email system, reducing the workload of email security administrators, and improving the accuracy of the company's detection of phishing emails.

[0197] Based on the above embodiment, the system further includes:

[0198] Sample enterprise email acquisition module, used to acquire sample enterprise emails in the target period;

[0199] The ecological law characteristic database construction module is used to collect statistics on the sample enterprise emails based on the email logs of the sample enterprise emails, obtain ecological law key-value pairs, and construct an ecological law characteristic database based on the ecological law key-value pairs.

[0200] In the present invention, an extensible email log unit can be set in the sample enterprise email acquisition module to record the enterprise's historical email logs. On the one hand, the email log does not record information such as email text and attachments to protect the privacy of enterprise users; on the other hand, the phishing email detection system provided by the present invention supports collaboration with other email classification algorithms to provide better detection results, specifically, by adding an extended classification field output by other classification algorithms to the email log.

[0201] Furthermore, the detection algorithm unit can be set in the ecological law feature database construction module, and the detection algorithm unit can run offline. On the one hand, the historical data already in the email log unit is used to obtain the corporate email ecological law; on the other hand, for the real-time input emails, the threat level of the emails is scored and sorted, and an alarm is generated according to the threshold pre-set by the administrator, and sent directly to the user or to the security administrator for review. The security administrator receives the alarm from the detection algorithm unit, conducts manual review, and then issues an alarm to the corporate user.

[0202] The system provided by the present invention is used to execute the above-mentioned method embodiments. Please refer to the above-mentioned embodiments for the specific processes and detailed contents, which will not be repeated here.

[0203] Figure 5 A schematic diagram of the structure of an electronic device provided by the present invention, such as Figure 5 As shown, the electronic device may include: a processor 501, a communication interface 502, a memory 503 and a communication bus 504, wherein the processor 501, the communication interface 502 and the memory 503 complete mutual communication through the communication bus 504. The processor 501 may call the logic instructions in the memory 503 to execute the phishing email detection method, which includes: obtaining the emails to be detected in a preset period; obtaining each feature value of the email to be detected in the target email ecological law group according to the ecological law feature database of the enterprise email and the preset phishing email type, wherein the ecological law feature database includes the ecological feature database of the received email and the ecological feature database of the sent email, and the target email ecological law group is determined according to the preset phishing email type; obtaining the score of the email to be detected according to the feature value, and sorting the score to obtain the phishing email detection result of the email to be detected.

[0204] In addition, the logic instructions in the above-mentioned memory 503 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0205] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the phishing email detection method provided by the above-mentioned methods, and the method includes: obtaining emails to be detected in a preset period; obtaining each feature value of the emails to be detected in a target email ecological law group based on an ecological law feature database of corporate emails and a preset phishing email type, the ecological law feature database including an ecological feature database of received emails and an ecological feature database of sent emails, and the target email ecological law group is determined based on the preset phishing email type; obtaining the score of the email to be detected based on the feature value, and sorting the score to obtain the phishing email detection result of the email to be detected.

[0206] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the above-mentioned phishing email detection methods, the methods comprising: obtaining emails to be detected in a preset period; obtaining each feature value of the emails to be detected in a target email ecological law group based on an ecological law feature database of corporate emails and a preset phishing email type, the ecological law feature database comprising an ecological feature database of received emails and an ecological feature database of sent emails, and the target email ecological law group is determined based on the preset phishing email type; obtaining scores of the emails to be detected based on the feature values, and sorting the scores to obtain phishing email detection results of the emails to be detected.

[0207] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0208] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0209] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting phishing emails, characterized in that: include: Get the emails to be tested in the preset period; According to the ecological law feature database of corporate emails and the preset phishing email type, each feature value of the email to be detected in the target email ecological law group is obtained, the ecological law feature database includes a received email ecological feature database and a sent email ecological feature database, and the target email ecological law group is determined according to the preset phishing email type; According to the characteristic value, obtaining the score of the email to be detected, and sorting the score to obtain the phishing email detection result of the email to be detected; Before obtaining each feature value of the to-be-detected email in the target email ecological law group according to the corporate email ecological law feature database and the preset phishing email type, the method further includes: Get sample corporate emails for the target period; Based on the email logs of sample enterprise emails, various features of the sample enterprise emails are counted and statistically analyzed to obtain ecological law key-value pairs, and an ecological law feature database is constructed based on the ecological law key-value pairs; the ecological law key-value pairs are expressed in the form of {key:value}; The step of obtaining the score of the email to be detected according to the characteristic value, and sorting the score to obtain the phishing email detection result of the email to be detected includes: S1. Obtain each feature value of a first mail to be detected in a target mail ecological law group, and each feature value of each second mail to be detected in a target mail ecological law group; the first mail to be detected is any mail in the mails to be detected in the current scoring round, except for the scored mails; the second mail to be detected is any mail in the mails to be detected in the current scoring round, except for the first mail to be detected; S2, comparing each feature value of the first mail to be detected with each feature value of the second mail to be detected in sequence, and after the first mail to be detected is compared with all the second mails to be detected, obtaining a score of the first mail to be detected according to the number of feature value items in the first mail to be detected that meet a preset comparison condition; S3, repeating steps S1 to S3 to obtain the score of each email in the email to be detected; S4. Sort the score of each email in the emails to be detected, and obtain the phishing email detection result of the emails to be detected according to the sorting result.

2. The phishing email detection method according to claim 1, characterized in that: Before counting and calculating various features of the sample enterprise emails based on the email logs of the sample enterprise emails to obtain ecological law key-value pairs, the method further includes: According to the field classification record of each email, the email log of the sample corporate email is obtained. The field classification record includes the recipient's email address, the sender's email address, the sender's name, the corporate user's login IP, the region where the corporate user's IP address is located, the timestamp, the URL that appears in the email body, the sending and receiving flag field, and the email classification algorithm classification field.

3. The phishing email detection method according to claim 1, characterized in that: The email log based on the sample enterprise emails counts and statistics various features of the sample enterprise emails to obtain ecological law key-value pairs, and constructs an ecological law feature database based on the ecological law key-value pairs, including: Obtain sender and receiver information according to the email log of the sample enterprise email; The sample corporate email is judged according to the sender and receiver information. If it is judged that the sample corporate email is an email sent by a corporate mailbox, according to the preset rules corresponding to each type of key-value pair, the key-value pairs corresponding to the sender information of the sample corporate email, the key-value pairs corresponding to the URL appearing in the email body, and the key-value pairs corresponding to the number of emails sent at the target time are counted to obtain a first ecological law characteristic value, and a sent email ecological characteristic database is constructed according to the first ecological law characteristic value; If it is determined that the sample corporate email is an email received by a corporate mailbox, the key-value pairs corresponding to all field classification records are counted according to the preset rules corresponding to each type of key-value pair, a second ecological law characteristic value is obtained, and a received email ecological characteristic database is constructed according to the second ecological law characteristic value; An ecological regularity characteristic database is obtained according to the sent mail ecological characteristic database and the received mail ecological characteristic database.

4. The phishing email detection method according to claim 1, characterized in that: After obtaining the mails to be detected in the preset period, the method further includes: The mail to be detected is subjected to data cleaning processing, wherein the data cleaning includes type exclusion processing, garbled code processing and URL field cleaning processing.

5. A phishing email detection system, characterized in that: include: The module for acquiring mails to be detected is used to acquire mails to be detected in a preset period; An ecological law characteristic value acquisition module is used to acquire each characteristic value of the to-be-detected email in the target email ecological law group according to the ecological law characteristic database of the corporate email and the preset phishing email type, wherein the ecological law characteristic database includes a received email ecological characteristic database and a sent email ecological characteristic database, and the target email ecological law group is determined according to the preset phishing email type; A scoring and sorting module, used to obtain the scores of the emails to be detected according to the feature values, and sort the scores to obtain the phishing email detection results of the emails to be detected; The system further comprises: Sample enterprise email acquisition module, used to acquire sample enterprise emails in the target period; The ecological law characteristic database construction module is used to count and count various characteristics of the sample enterprise emails based on the email logs of the sample enterprise emails, obtain ecological law key-value pairs, and construct an ecological law characteristic database according to the ecological law key-value pairs; the ecological law key-value pairs are expressed in the form of {key:value}; The scoring and ranking module is specifically used for: S1. Obtain each feature value of a first mail to be detected in a target mail ecological law group, and each feature value of each second mail to be detected in a target mail ecological law group; the first mail to be detected is any mail in the mails to be detected in the current scoring round, except for the scored mails; the second mail to be detected is any mail in the mails to be detected in the current scoring round, except for the first mail to be detected; S2, comparing each feature value of the first mail to be detected with each feature value of the second mail to be detected in sequence, and after the first mail to be detected is compared with all the second mails to be detected, obtaining a score of the first mail to be detected according to the number of feature value items in the first mail to be detected that meet a preset comparison condition; S3, repeating steps S1 to S3 to obtain the score of each email in the email to be detected; S4. Sort the score of each email in the emails to be detected, and obtain the phishing email detection result of the emails to be detected according to the sorting result.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the phishing email detection method according to any one of claims 1 to 4 are implemented.

7. A non-transitory 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 phishing email detection method according to any one of claims 1 to 4 are implemented.

Citation Information

Patent Citations

  • Spear phishing mail detection method and device

    CN105072137A

  • Harpoon mail detection method and device, equipment and computer readable storage medium

    CN111404806A