Mail classification method and device based on user requirements, equipment and storage medium

Abstract

The invention relates to the field of classification models, and provides a mail classification method based on user demands, and the method comprises the steps: obtaining the mail classification preference of a target user, and determining a target classification model according with the classification habit of the target user based on the mail classification preference; extracting mail features of the to-be-classified mails based on a preset feature extraction rule; and inputting the mail features of the to-be-classified mails into the target classification model, and outputting a classification result of the to-be-classified mails. According to the method and the device, the mails are classified through the target classification model, the frequency of manually setting and adjusting the classification rule by the user is reduced, then the classification model conforming to the mail classification habit of each user is further determined based on the mail classification preference of each user, and the mail classification efficiency is improved through the classification model conforming to the mail classification habit of the user. According to the method and the device, the mails are classified, mistaken classification or missing classification of the mails is avoided, the accuracy of the mail classification result is improved, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the field of classification models and financial technology, and in particular to a method, device, computer equipment and computer-readable storage medium for email classification based on user needs. Background Art

[0002] With the popularization and rapid development of Internet technology, email has become an important means of communication for people's work and life due to its fast, convenient and low-cost features. The current email classification system mainly relies on manually setting classification rules to classify emails. However, the manually set classification rules are difficult to meet the complex classification needs of users, and the processing flexibility is poor, which leads to low accuracy of email classification results and reduces user experience.

[0003] Therefore, how to solve the low accuracy of current email classification results has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] The main purpose of the present invention is to provide a method, device, equipment and computer-readable storage medium for mail classification based on user needs, aiming to solve the technical problem of low accuracy of current mail classification results.

[0005] To achieve the above object, the present invention provides an email classification method based on user needs, and the text association method comprises the following steps:

[0006] Acquire the email classification preference of the target user, and determine a target classification model that meets the classification habits of the target user based on the email classification preference;

[0007] Based on the preset feature extraction rules, feature extraction is performed on the mails to be classified to obtain mail features of the mails to be classified;

[0008] The mail features of the mail to be classified are input into the target classification model, and the classification result of the mail to be classified is output.

[0009] In addition, to achieve the above purpose, the present invention also provides a mail classification device based on user needs, and the mail classification device based on user needs includes:

[0010] A classification model determination module is used to obtain the email classification preference of the target user and determine a target classification model that meets the classification habits of the target user based on the email classification preference;

[0011] An email feature extraction module is used to extract features of the emails to be classified based on preset feature extraction rules to obtain email features of the emails to be classified;

[0012] The classification result output module is used to input the mail features of the mail to be classified into the target classification model and output the classification result of the mail to be classified.

[0013] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer device, which includes a processor, a memory, and an email classification program based on user needs stored in the memory and executable by the processor, wherein when the email classification program based on user needs is executed by the processor, the steps of the email classification method based on user needs as described above are implemented.

[0014] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which is stored an email classification program based on user needs, wherein when the email classification program based on user needs is executed by a processor, the steps of the email classification method based on user needs as described above are implemented.

[0015] The present application provides a method for email classification based on user needs, the method obtains the email classification preferences of the target user, and determines a target classification model that meets the classification habits of the target user based on the email classification preferences; based on the preset feature extraction rules, feature extraction is performed on the emails to be classified to obtain the email features of the emails to be classified; the email features of the emails to be classified are input into the target classification model, and the classification results of the emails to be classified are output. In the above manner, the present application first classifies each email through the target classification model, reducing the frequency of users manually setting and adjusting classification rules, and then further determines the classification model that meets the email classification habits of each user based on the email classification preferences of each user, and classifies the emails through the classification model that meets the email classification habits of the user, avoiding misclassification or omission of emails, improving the accuracy of the email classification results, and enhancing the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of a first embodiment of a method for classifying emails based on user needs according to the present invention;

[0017] Figure 2 This is a flow chart of a second embodiment of the email classification method based on user needs of the present invention;

[0018] Figure 3 This is a flow chart of a third embodiment of the email classification method based on user needs of the present invention;

[0019] Figure 4 It is a schematic diagram of a mail classification device based on user needs of the present invention;

[0020] Figure 5The figure is a schematic diagram of the hardware structure of a computer device involved in an embodiment of the present invention.

[0021] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0023] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0024] It should be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0025] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0026] The email classification method based on user needs involved in the embodiment of the present invention is mainly applied to computer equipment. The email classification generating device based on user needs can be a device with display and processing functions such as a PC, a portable computer, and a mobile terminal.

[0027] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0028] Reference Figure 1 , Figure 1 This is a flow chart of a first embodiment of the email classification method based on user needs of the present invention.

[0029] like Figure 1 As shown, an embodiment of the present invention provides an email classification method based on user needs, and the email classification method based on user needs includes steps S10 to S30.

[0030] Step S10: obtaining the target user's email classification preference, and determining a target classification model that meets the target user's classification habits based on the email classification preference;

[0031] The current email classification system mainly relies on manually setting classification rules to classify emails. However, users need to constantly update and adjust email classification rules to adapt to new email content and senders. For example:

[0032] 1. Manual rule setting is cumbersome: Users need to constantly update and adjust classification rules to adapt to new email content and senders. In addition, classification rules are complex and difficult to maintain. Manually setting and maintaining a large number of classification rules is time-consuming and error-prone, especially for users with a large amount of emails.

[0033] 2. Low keyword matching accuracy: Simple keyword matching is prone to misjudgment, resulting in important emails being misclassified or spam entering the inbox. The processing flexibility is poor and simple keyword matching cannot adapt to the diverse content of emails, especially when facing emails in multiple languages ​​and fields.

[0034] 3. Lack of adaptive capabilities: The target classification system is difficult to automatically learn the user's classification preferences and cannot make adjustments when the user's habits change, resulting in a poor user experience. In addition, due to the existence of misclassification and missed classification, users often need to manually check and adjust the classification results, which increases time costs and usage burdens.

[0035] Therefore, manually set classification rules are difficult to meet the complex classification needs of users, and the processing flexibility is poor, which leads to low accuracy of current email classification results and reduces user experience.

[0036] In order to solve the above problem, this embodiment first classifies each email through a target classification model, reducing the frequency of users manually setting and adjusting classification rules, and then further determines a classification model that meets the email classification habits of each user based on the email classification preferences of each user, and classifies emails through the classification model that meets the user's email classification habits, thereby avoiding misclassification or omission of emails, improving the accuracy of email classification results, and enhancing user experience.

[0037] Specifically, classification models include support vector algorithms, random forest algorithms, and deep learning algorithms. Support vector machine algorithms (SVM) are supervised learning algorithms that are often used for classification and regression problems. Random forest algorithms (Random forest) are supervised machine learning algorithms based on classification trees. They improve the prediction accuracy of classification models by summarizing a large number of classification trees. They are new models that replace traditional machine learning methods such as neural networks. Deep learning refers to the classification or regression of unknown data by training multi-layer network structures. Deep learning algorithms include supervised learning algorithms (such as deep feedforward networks, convolutional neural networks, recurrent neural networks, etc.) and unsupervised learning algorithms (deep belief networks, deep Boltzmann machines, and deep autoencoders, etc.).

[0038] Among them, the email classification preference is the user's habit of classifying historical emails, and the email classification includes spam and normal emails. The closeness relationship between the target user and any user in the social network can be calculated, and the accuracy of the email classification algorithm can be improved by using the characteristic that the higher the closeness, the lower the probability of sending spam between users, which is to improve the accuracy of the email classification algorithm.

[0039] The user's classification habits for historical normal emails may be: classify emails according to customer groups, customer sources, customer status, customer types, customer star ratings, customer activity, etc.; or classify emails according to categories such as inquiries, quotations, PI, CI, PL, water orders, etc.; or classify emails according to different role categories (such as customers, colleagues, or roles in the address book, etc.), or according to sender attributes (such as the sender's geographical location, gender, etc.), or according to the activity to which the email subject belongs or the business type to which the email belongs; emails can also be classified according to the email language type.

[0040] Different email classification preferences may correspond to different classification models. During training, a classification model with the highest classification accuracy may be determined among various classification models as the classification model corresponding to each email classification preference, and the corresponding relationship may be stored.

[0041] Furthermore, before step S10, the method further includes:

[0042] Based on the email classification preference, determining a target classification algorithm that meets the classification habits of the target user;

[0043] Acquire historical email classification data of the target user as a training set, and train the target classification algorithm based on the training set to obtain the target classification model;

[0044] Based on a preset validation set, the target classification model is validated to obtain the classification accuracy of the target classification algorithm;

[0045] Based on the updated training set, the target classification model is iteratively trained until the classification accuracy of the target classification model exceeds a preset classification accuracy threshold.

[0046] In this embodiment, since different classification algorithms are suitable for classifying different data, the classification accuracy of each classification algorithm is statistically analyzed in advance based on each email classification preference and the historical classification results of each classification algorithm, and the classification algorithm with the highest accuracy is selected as the corresponding target classification algorithm.

[0047] Based on the target user's email classification preference, after determining a target classification algorithm that meets the target user's classification habits, further obtain the target user's historical email classification data as a training set for the target classification algorithm, and based on the training set, further perform classification training on the target classification algorithm to fit and adjust the model parameters of the target classification algorithm.

[0048] According to the classification habits of the target user, the emails and their corresponding classification results are simulated, and the simulated emails and their corresponding classification results are used as the validation set. After the training is completed, the emails in the validation set are trained with the trained target classification algorithm, and based on the classification results corresponding to each email in the validation set, the accuracy corresponding to the current target classification algorithm is calculated.

[0049] When the accuracy corresponding to the current target classification algorithm is less than the preset classification accuracy threshold, the training set is updated again until the classification accuracy of the target classification algorithm exceeds the classification accuracy threshold, then the training of the target classification algorithm is stopped and the target classification algorithm is used as the target classification model.

[0050] Specifically, each classification algorithm can be trained according to a preset training set, and the classification accuracy of each trained classification algorithm can be calculated based on the verification set. Among the classification algorithms whose classification accuracy exceeds the classification accuracy threshold, the classification algorithm with the highest classification accuracy is selected as the target classification model.

[0051] In a specific embodiment, the hyperparameters of the trained target classification algorithm can be further debugged through the validation set, so as to evaluate the model performance and generalization ability of each target classification algorithm through the test set, and select the algorithm with the highest classification accuracy from the debugged target classification algorithms as the target classification model.

[0052] Step S20: extracting features of the mails to be classified based on preset feature extraction rules to obtain mail features of the mails to be classified;

[0053] In this embodiment, the feature extraction rules include one or more of keyword extraction (i.e., using algorithms such as TF-IDF to extract important keywords in emails), semantic feature extraction (i.e., using NLP technology to perform semantic analysis and extract semantic features of emails), and metadata feature extraction (i.e., extracting metadata such as senders and timestamps).

[0054] According to the preset feature extraction rule, feature information corresponding to the mail to be classified is extracted as the mail feature of the mail to be classified.

[0055] Step S30: inputting the mail features of the mail to be classified into the target classification model, and outputting the classification result of the mail to be classified.

[0056] In this embodiment, the mail features of the mail to be classified are input into the target classification model, so as to classify the mail to be classified based on the target classification model, and output the classification result of the mail to be classified.

[0057] Furthermore, after step S30, the method further includes:

[0058] According to the classification result of the mail to be classified, the mail to be classified is classified into a corresponding folder, or a corresponding mail classification label is added to the mail to be classified.

[0059] In this embodiment, after obtaining the classification result of the mail to be classified, the classification result can be used as a label, and the label corresponding to the classification result can be added to the mail to be classified; or the folder corresponding to the classification result can be searched in the folder storing each classified mail, and the mail to be classified can be stored in the folder corresponding to the classification result; or the label corresponding to the classification result can be added to the mail to be classified, and the mail to be classified can be stored in the folder corresponding to the classification result, so as to complete the classification of the mail to be classified.

[0060] This embodiment provides a method for email classification based on user needs, the method obtains the email classification preferences of the target user, and determines a target classification model that meets the classification habits of the target user based on the email classification preferences; based on the preset feature extraction rules, feature extraction is performed on the email to be classified to obtain the email features of the email to be classified; the email features of the email to be classified are input into the target classification model, and the classification results of the email to be classified are output. In the above manner, the present application first classifies each email through the target classification model, reducing the frequency of users manually setting and adjusting classification rules, and then further determines the classification model that meets the email classification habits of each user based on the email classification preferences of each user, and classifies the emails through the classification model that meets the email classification habits of the user, avoiding misclassification or omission of emails, improving the accuracy of the email classification results, and enhancing the user experience.

[0061] Reference Figure 2 , Figure 2 This is a flow chart of a second embodiment of the email classification method based on user needs of the present invention.

[0062] Based on the above Figure 1 In the embodiment shown, in this embodiment, the step S20 specifically includes:

[0063] Step S21, performing semantic analysis on the email to be classified, and extracting features from the analyzed semantics to obtain semantic features of the email to be classified;

[0064] Step S22, extracting keywords of the email to be classified based on keyword feature extraction rules to obtain keyword features of the email to be classified;

[0065] Step S23, based on the preset metadata type, extracting metadata features of the email to be classified to obtain metadata features of the email to be classified;

[0066] Step S24: generating email features of the email to be classified based on the semantic features, keyword features and metadata features of the email to be classified.

[0067] In this embodiment, firstly, a natural language processing (NLP) model is used to perform semantic analysis on the emails to be classified, and semantic features of the emails to be classified are extracted according to the semantic analysis results.

[0068] Then, the TF-IDF algorithm is used to perform keyword detection on the mails to be classified, and the keywords in the mails are extracted as keyword features of the mails to be classified.

[0069] Finally, based on the extracted metadata features such as sender, timestamp, etc., combined with semantic features and keyword features, the email features of the email to be classified are generated.

[0070] It is understandable that email metadata can be extracted in advance based on the target user's email classification habits and stored in a metadata database. Specifically, a graph database can be used to manage the extracted email metadata to improve the efficiency of the classification algorithm.

[0071] Furthermore, the classification model can be optimized and trained based on the characteristic that the reply degree of spam is zero to reduce the misjudgment rate of emails; the classification model can also be optimized based on the user's behavior of deleting spam emails mistakenly judged as normal emails and restoring normal emails mistakenly judged as spam emails.

[0072] Furthermore, before step S20, the method further includes:

[0073] The text of the mail to be classified is cleaned up to clean up the hypertext markup language HTML tags, special characters, redundant symbols and repeated words in the mail to be classified.

[0074] Furthermore, before extracting features of the mails to be classified based on the preset feature extraction rules and obtaining the mail features of the mails to be classified in step S20, the method further includes:

[0075] The cleaned emails to be classified are subjected to word segmentation processing to convert the texts corresponding to the emails to be classified into a word list.

[0076] In this embodiment, the email information of the email to be classified is preprocessed in advance, wherein the preprocessing operation specifically includes text cleaning, such as removing HTML tags, special characters and redundant spaces, etc., and the email information can be further segmented after the text cleaning, that is, the email content is segmented, and the email text after segmentation is converted into a corresponding vocabulary list, so as to facilitate subsequent feature extraction of the vocabulary list to obtain the email features of the email to be classified.

[0077] Through the above-mentioned method, this embodiment provides a method for email classification based on user needs. The method classifies emails through an email classification model that conforms to the user's classification habits, reduces misclassification and missed classification, and improves the accuracy and efficiency of email classification; and reduces the frequency of users manually setting and adjusting classification rules, thereby improving the user experience. In addition, this embodiment can perform semantic analysis and text translation on email information through the NLP language model, thereby meeting the needs of email classification in multiple languages ​​and multiple fields to adapt to diversified needs; finally, through the feedback mechanism, the classification model is self-optimized to maintain an efficient classification effect. In this way, accurate email classification is achieved, spam interference is reduced, spam is effectively filtered, and user experience is improved.

[0078] Reference Figure 3 , Figure 3 This is a flow chart of the third embodiment of the email classification method based on user needs of the present invention. Figure 1 In the embodiment shown, in this embodiment, after step S30, the following steps are further included:

[0079] Step S40, collecting evaluation parameters of the target user on the classification result of the email to be classified based on the plug-in interface;

[0080] Step S50: Based on the classification result of the mail to be classified and the evaluation parameter, feedback adjustment is performed on the model parameters of the target classification model to achieve feedback optimization of the target classification model.

[0081] In this embodiment, the target user's satisfaction evaluation on the classification results of the classified emails is collected through the plug-in interface, such as the user's rating of the current classification results, and the model parameters are adjusted according to the user feedback to complete the optimization and update of the classification rules in the model. In this way, the classification model is continuously learned and optimized, and the classification accuracy and user satisfaction of the classification model are gradually improved.

[0082] Specifically, when the user's score for the classification result of an email to be classified is lower than a preset score threshold, it means that the target user is not satisfied with the classification result of the email to be classified. The target user's accurate classification result of the email to be classified can be further collected, and the target classification model can be optimized and trained based on the email to be classified and the accurate classification result.

[0083] When the user's score for the classification result of an email to be classified is not lower than the preset score threshold, it means that the target user is satisfied with the classification result of the email to be classified. The email to be classified and the classification result can be used as a training set, a validation set or a test set to subsequently optimize the target classification model.

[0084] Through the above-mentioned method, this embodiment provides a method for email classification based on user needs. The method uses machine learning and NLP language to accurately classify emails, reduce misclassification and missed classification, and improve classification accuracy and efficiency; through the training and feedback mechanism of the classification model, emails are automatically classified, which reduces the frequency of manual setting and adjustment of rules by users, and improves user experience; and can meet the needs of email classification in multiple languages ​​and multiple fields, and adapt to the personalized email classification needs of different users, so as to better adapt to the diverse needs of users; finally, through the feedback mechanism, the classification model is updated and optimized according to user feedback and usage habits, so as to dynamically adapt to user classification habits, so as to maintain efficient classification effects. In this way, accurate classification of emails is achieved, spam interference is reduced, spam is effectively filtered, and user experience is improved.

[0085] See also Figure 4 , Figure 4 This is a functional module diagram of a mail classification device based on user needs provided in an embodiment of the present application.

[0086] like Figure 4 As shown, the mail classification device 400 based on user needs includes:

[0087] A classification model determination module 10 is used to obtain the email classification preference of the target user and determine a target classification model that meets the classification habits of the target user based on the email classification preference;

[0088] The mail feature extraction module 20 is used to extract features of the mails to be classified based on preset feature extraction rules to obtain mail features of the mails to be classified;

[0089] The classification result output module 30 is used to input the mail features of the mail to be classified into the target classification model and output the classification result of the mail to be classified.

[0090] Furthermore, the email feature extraction module 20 specifically includes:

[0091] A semantic feature extraction unit, used for performing semantic analysis on the email to be classified, and extracting features from the analyzed semantics to obtain semantic features of the email to be classified;

[0092] A keyword feature extraction unit, configured to extract keywords of the email to be classified based on a keyword feature extraction rule to obtain keyword features of the email to be classified;

[0093] A metadata feature extraction unit, configured to extract metadata features of the email to be classified based on a preset metadata type, to obtain metadata features of the email to be classified;

[0094] The mail feature generating unit is used to generate mail features of the mail to be classified based on the semantic features, keyword features and metadata features of the mail to be classified.

[0095] Furthermore, the mail classification device 400 based on user needs also includes:

[0096] An evaluation parameter collection module, used for collecting the evaluation parameters of the target user's classification result of the email to be classified based on the plug-in interface;

[0097] The model feedback optimization module is used to perform feedback adjustment on the model parameters of the target classification model based on the classification result of the mail to be classified and the evaluation parameter, so as to achieve feedback optimization of the target classification model.

[0098] Furthermore, the mail classification device 400 based on user needs also includes:

[0099] The mail text cleaning module is used to clean the text of the mail to be classified, so as to clean up the hypertext markup language HTML tags, special characters, redundant symbols and repeated words in the mail to be classified.

[0100] Furthermore, the mail classification device 400 based on user needs also includes:

[0101] The email text segmentation module is used to perform a segmentation process on the cleaned emails to be classified, so as to convert the text corresponding to the emails to be classified into a word list.

[0102] Furthermore, the mail classification device 400 based on user needs also includes:

[0103] A classification algorithm determination module, used to determine a target classification algorithm that meets the classification habits of the target user based on the email classification preference;

[0104] A classification model generation module is used to obtain the historical email classification data of the target user as a training set, and train the target classification algorithm based on the training set to obtain the target classification model;

[0105] A classification model verification module is used to verify the target classification model based on a preset verification set to obtain the classification accuracy of the target classification algorithm;

[0106] The classification model training module is used to iteratively train the target classification model based on the updated training set until the classification accuracy of the target classification model exceeds a preset classification accuracy threshold.

[0107] Furthermore, the mail classification device 400 based on user needs also includes:

[0108] The classified mail storage module is used to classify the mails to be classified into corresponding folders or add corresponding mail classification labels to the mails to be classified according to the classification results of the mails to be classified.

[0109] It should be noted that those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described device and each module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0110] The above-mentioned device can be implemented in the form of a computer program. Figure 5 Runs on the computer device shown.

[0111] See also Figure 5 , Figure 5 1 is a schematic block diagram of a computer device provided in an embodiment of the present application. The computer device may be a server.

[0112] See also Figure 5 The computer device includes a processor, a memory and a network interface connected through a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.

[0113] The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute any email classification method based on user requirements.

[0114] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

[0115] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor, the processor can execute any email classification method based on user needs.

[0116] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0117] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0118] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps:

[0119] Acquire the email classification preference of the target user, and determine a target classification model that meets the classification habits of the target user based on the email classification preference;

[0120] Based on the preset feature extraction rules, feature extraction is performed on the mails to be classified to obtain mail features of the mails to be classified;

[0121] The mail features of the mail to be classified are input into the target classification model, and the classification result of the mail to be classified is output.

[0122] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps:

[0123] Performing semantic analysis on the email to be classified, and extracting features from the analyzed semantics to obtain semantic features of the email to be classified;

[0124] Based on the keyword feature extraction rule, keywords of the email to be classified are extracted to obtain keyword features of the email to be classified;

[0125] Based on the preset metadata type, metadata features are extracted for the email to be classified to obtain metadata features of the email to be classified;

[0126] Based on the semantic features, keyword features and metadata features of the mail to be classified, the mail features of the mail to be classified are generated.

[0127] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps:

[0128] Collecting evaluation parameters of the target user on the classification result of the email to be classified based on the plug-in interface;

[0129] Based on the classification result of the mail to be classified and the evaluation parameter, feedback adjustment is performed on the model parameters of the target classification model to achieve feedback optimization of the target classification model.

[0130] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps:

[0131] The text of the mail to be classified is cleaned up to clean up the hypertext markup language HTML tags, special characters, redundant symbols and repeated words in the mail to be classified.

[0132] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps:

[0133] The cleaned emails to be classified are subjected to word segmentation processing to convert the texts corresponding to the emails to be classified into a word list.

[0134] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps:

[0135] Based on the email classification preference, determining a target classification algorithm that meets the classification habits of the target user;

[0136] Acquire historical email classification data of the target user as a training set, and train the target classification algorithm based on the training set to obtain the target classification model;

[0137] Based on a preset validation set, the target classification model is validated to obtain the classification accuracy of the target classification algorithm;

[0138] Based on the updated training set, the target classification model is iteratively trained until the classification accuracy of the target classification model exceeds a preset classification accuracy threshold.

[0139] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps:

[0140] According to the classification result of the mail to be classified, the mail to be classified is classified into a corresponding folder, or a corresponding mail classification label is added to the mail to be classified.

[0141] A computer-readable storage medium is also provided in an embodiment of the present application. The computer-readable storage medium stores a computer program. The computer program includes program instructions. The processor executes the program instructions to implement any one of the email classification methods based on user needs provided in the embodiments of the present application.

[0142] The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (Smart Medi a Card, SMC), a secure digital (Secure Digital, SD) card, a flash memory card (Flash Card), etc., equipped on the computer device.

[0143] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A mail classification method based on user needs, characterized in that: The text association method comprises the following steps: Acquire the email classification preference of the target user, and determine a target classification model that meets the classification habits of the target user based on the email classification preference; Based on the preset feature extraction rules, feature extraction is performed on the mails to be classified to obtain mail features of the mails to be classified; The mail features of the mail to be classified are input into the target classification model, and the classification result of the mail to be classified is output.

2. The method for classifying mails based on user needs as claimed in claim 1, characterized in that: The extracting features of the mails to be classified based on the preset feature extraction rules to obtain the mail features of the mails to be classified includes: Performing semantic analysis on the email to be classified, and extracting features from the analyzed semantics to obtain semantic features of the email to be classified; Based on the keyword feature extraction rule, keywords of the email to be classified are extracted to obtain keyword features of the email to be classified; Based on the preset metadata type, metadata features are extracted for the email to be classified to obtain metadata features of the email to be classified; Based on the semantic features, keyword features and metadata features of the mail to be classified, the mail features of the mail to be classified are generated.

3. The method for classifying mails based on user needs as claimed in claim 1, characterized in that: After inputting the mail features of the mail to be classified into the target classification model and outputting the classification result of the mail to be classified, the method further includes: Collecting evaluation parameters of the target user on the classification result of the email to be classified based on the plug-in interface; Based on the classification result of the mail to be classified and the evaluation parameter, feedback adjustment is performed on the model parameters of the target classification model to achieve feedback optimization of the target classification model.

4. The method for classifying mails based on user needs as claimed in claim 1, characterized in that: Before extracting features of the mails to be classified based on the preset feature extraction rules to obtain the mail features of the mails to be classified, the method further includes: The text of the mail to be classified is cleaned up to clean up the hypertext markup language HTML tags, special characters, redundant symbols and repeated words in the mail to be classified.

5. The method for classifying mails based on user needs as claimed in claim 1, characterized in that: Before extracting features of the mails to be classified based on the preset feature extraction rules to obtain the mail features of the mails to be classified, the method further includes: The cleaned emails to be classified are subjected to word segmentation processing to convert the texts corresponding to the emails to be classified into a word list.

6. The method for classifying mails based on user needs as claimed in claim 1, characterized in that: Before obtaining the target user's email classification preference and determining a target classification model that meets the target user's classification habits based on the email classification preference, the method further includes: Based on the email classification preference, determining a target classification algorithm that meets the classification habits of the target user; Acquire historical email classification data of the target user as a training set, and train the target classification algorithm based on the training set to obtain the target classification model; Based on a preset validation set, the target classification model is validated to obtain the classification accuracy of the target classification algorithm; Based on the updated training set, the target classification model is iteratively trained until the classification accuracy of the target classification model exceeds a preset classification accuracy threshold.

7. The method for classifying mails based on user needs according to any one of claims 1 to 6, characterized in that: After inputting the mail features of the mail to be classified into the target classification model and outputting the classification result of the mail to be classified, the method further includes: According to the classification result of the mail to be classified, the mail to be classified is classified into a corresponding folder, or a corresponding mail classification label is added to the mail to be classified.

8. A mail classification device based on user needs, characterized in that: The mail classification device based on user needs includes: A classification model determination module is used to obtain the email classification preference of the target user and determine a target classification model that meets the classification habits of the target user based on the email classification preference; An email feature extraction module is used to extract features of the emails to be classified based on preset feature extraction rules to obtain email features of the emails to be classified; The classification result output module is used to input the mail features of the mail to be classified into the target classification model and output the classification result of the mail to be classified.

9. A computer device, characterized in that: The computer device includes a processor, a memory, and a mail classification program based on user needs stored in the memory and executable by the processor, wherein when the mail classification program based on user needs is executed by the processor, the steps of the mail classification method based on user needs as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a mail classification program based on user needs, wherein when the mail classification program based on user needs is executed by a processor, the steps of the mail classification method based on user needs as claimed in any one of claims 1 to 7 are implemented.

Images