Text classification model training method and device for Internet social media
By constructing polar web page text datasets and optimizing and updating them using ERNIE models, the problems of high manual annotation cost and poor model applicability are solved, and efficient and accurate text classification model training is achieved.
Patent Information
- Application Number
- CN202311596487.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-11-27
- Publication Date
- 2025-05-27
AI Technical Summary
In the existing text classification technology, manual adding real tags is time-consuming, laborious, costly, and subjective and consistency challenges, resulting in the model not being applicable to other fields.
The training data set is constructed by obtaining text data obtained from web pages of different polarity, using the ERNIE model for data preprocessing and classification, and the loss between the predicted results and the real classification label is constructed to optimize the update model.
It reduces the cost of manual labeling and improves the efficiency and accuracy of the model, making the model not only suitable for the current field, but also can analyze public opinion changes and track netizens' views in real time.
Smart Images
Figure CN120045969A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a text classification model training method and device for Internet social media. Background Art
[0002] Natural Language Processing (NLP) is an important direction in the fields of computer science and artificial intelligence. It studies various theories and methods that can achieve effective communication between people and computers using natural language. Natural language processing is a science that integrates linguistics, computer science, and mathematics. It is mainly used in machine translation, public opinion monitoring, automatic summarization, opinion extraction, text classification, question answering, text semantic comparison, speech recognition, Chinese OCR, etc.
[0003] Internet social media has become an important channel for people to obtain information and communicate, and it is of great significance to individuals, governments, enterprises and other organizations and groups. The application of public opinion monitoring and text classification technology in natural language processing in Internet social media is also of great significance, such as being able to timely understand public sentiment and attitudes, discover potential problems and crises, and strengthen security monitoring and early warning.
[0004] For public opinion monitoring and text classification technologies in natural language processing, supervised learning algorithms are currently mostly used. For supervised learning algorithms, whether it is traditional machine learning or the current mainstream deep learning, they are all driven by training data. The training data set is usually in the form of {(I 1 , O 1 ), (I 2 , O 2 ),...,(I n , O n )},I n represents the nth sample in the training data set, O n Represents sample I in the training data set n Corresponding true labels. Only after collecting enough training data and adding corresponding true labels can various downstream natural language processing classification tasks be carried out. At present, the true labels of samples are mainly added manually. Therefore, when faced with a large amount of training data, the method of manually adding true labels is time-consuming, labor-intensive, and costly, and there may be challenges in subjectivity and consistency. At the same time, the corresponding model obtained by training after manually adding true labels is only applicable to the field to which the current training data and its true labels belong. When processing other fields, the corresponding true labels need to be added again. Summary of the invention
[0005] In view of this, the embodiments of the present invention provide a text classification model training method and device for Internet social media to eliminate or improve one or more defects existing in the prior art, and solve the problems in the prior art of manually adding real labels, resulting in high time and labor costs, challenges of subjectivity and consistency in manually adding real labels, and the problem that the model trained based on manually adding real labels is not applicable to other fields.
[0006] In one aspect, the present invention provides a text classification model training method for Internet social media, the method comprising the following steps:
[0007] Acquire a first training data set, wherein the first training data set includes a plurality of samples, each sample includes text data obtained from web pages of different polarities; the polarities include positive sentiment tendency polarities and negative sentiment tendency polarities;
[0008] Obtaining an initial model, wherein the initial model adopts an ERNIE model, and the initial model includes a data preprocessing module and a classification module; inputting each sample into the data preprocessing module in batches for data cleaning and preprocessing operations, and inputting the processed samples into the classification module for classification to obtain prediction results of each sample;
[0009] The initial model is trained using the first training data set, and the loss between the prediction result and the true classification label is constructed. The initial model is optimized and updated with the goal of minimizing the loss to obtain a final text classification model; the true classification label includes positive sentiment tendency and negative sentiment tendency.
[0010] In some embodiments of the present invention, each sample includes text data obtained from web pages of different polarities, and further includes:
[0011] Based on natural language processing technology and a pre-trained web page polarity classification model, web pages are classified into polarity categories to obtain web pages with positive sentiment tendency polarity and web pages with negative sentiment tendency polarity; the web page polarity classification model is obtained based on machine learning or deep learning training.
[0012] In some embodiments of the present invention, each sample includes text data obtained from web pages of different polarities, and further includes:
[0013] A preset crawler program is used to crawl the required text data from the polarity web pages; wherein the true classification of the text data crawled from the positive sentiment tendency polarity web pages is positive sentiment tendency, and the true classification of the text data crawled from the negative sentiment tendency polarity web pages is negative sentiment tendency.
[0014] In some embodiments of the present invention, after obtaining the first training data set, the method further includes:
[0015] The first training data set is filtered using a preset screening method, and a small batch data set is constructed to train the initial model to obtain a preliminary text classification model; the preset screening method includes keyword matching and model recognition.
[0016] In some embodiments of the present invention, the method further comprises:
[0017] The sample is subjected to data cleaning and preprocessing operations using text processing techniques; the text processing techniques include removing stop words, performing stemming, lemmatization, and correcting spelling errors.
[0018] In some embodiments of the present invention, the method further comprises:
[0019] Construct the cross entropy loss between the prediction result and the true classification label. The cross entropy loss calculation formula is:
[0020] L=-∑ i p(x i )log(q(x i ));
[0021] Wherein, L represents the cross entropy loss; p(x i ) represents the sample x i The true classification label of q(x i ) represents the sample x i prediction results.
[0022] In some embodiments of the present invention, when no relevant polarity web page can be found to obtain text data, the method further includes:
[0023] Obtain a second training data set, where the second training data set includes multiple samples, each sample includes a piece of text data; randomly obtain samples from the second training data set, construct a small batch training set, and add a true classification label to each sample in the small batch training set;
[0024] Acquire the initial model, which includes a data preprocessing module and a classification module; input each sample in the small batch training set into the data preprocessing module in batches for data cleaning and preprocessing operations, input the processed samples into the classification module for classification, and obtain the prediction results of each sample;
[0025] The initial model is trained using the small batch training set, and the loss between the prediction result and the true classification label is constructed to obtain a preliminary text classification model; the preliminary text classification model is then trained using the second training data set, and the preliminary text classification model is optimized and updated to obtain a final text classification model.
[0026] In some embodiments of the present invention, the method further comprises:
[0027] Input each sample in the small batch training set into the preliminary text classification model in batches, obtain a prediction result for each sample, and determine the correctness of the prediction result;
[0028] The samples with wrong prediction results are screened and analyzed by using preset optimization means; the preset optimization means include keyword matching and manual verification.
[0029] In another aspect, the present invention provides a text classification model training device for Internet social media, the device comprising:
[0030] A data collection module is used to collect text data from web pages of different polarities as samples to construct a training data set;
[0031] A result prediction module, used to input each sample in the training data set into the model to be trained to obtain a prediction result;
[0032] The model training module is used to train the model according to the training data set, and construct the loss between the prediction result and the corresponding true classification label, and optimize and update the model with the goal of minimizing the loss.
[0033] On the other hand, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any one of the methods mentioned above.
[0034] The beneficial effects of the present invention are at least:
[0035] The present invention provides a text classification model training method and device for Internet social media, including: obtaining a first training data set, which includes multiple samples, each sample includes text data obtained from web pages of different polarities; obtaining an initial model, the initial model adopts an ERNIE model, and includes a data preprocessing module and a classification module; inputting each sample into the data preprocessing module in batches for data cleaning and preprocessing operations, inputting the processed samples into the classification module for classification, and obtaining the prediction results of each sample; using the first training data set to train the initial model, constructing the loss between the prediction result and the true classification label, and optimizing and updating the initial model with the goal of minimizing the loss to obtain the final text classification model. The model training method provided by the present invention can make full use of Internet public data information, and does not completely rely on manual annotation, thereby reducing labor costs and saving development time; the trained text classification model has high efficiency and high precision; at the same time, according to the classification results, it can analyze the changes in public opinion in real time and track the changes in netizens' opinions.
[0036] Additional advantages, purposes, and features of the present invention will be described in part in the following description, and will become apparent to those skilled in the art after studying the following, or may be learned from the practice of the present invention. The purposes and other advantages of the present invention may be achieved and obtained by the structures specifically indicated in the specification and the accompanying drawings.
[0037] Those skilled in the art will appreciate that the objectives and advantages that can be achieved with the present invention are not limited to the above specific description, and the above and other objectives that can be achieved by the present invention will be more clearly understood from the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of the present application, and do not constitute a limitation of the present invention. In the drawings:
[0039] Figure 1 The figure is a schematic diagram of the steps of a text classification model training method for Internet social media in one embodiment of the present invention.
[0040] Figure 2 The present invention is a flowchart of a text classification model training method for Internet social media in one embodiment of the present invention. DETAILED DESCRIPTION
[0041] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.
[0042] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, only structures and / or processing steps closely related to the solutions according to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.
[0043] It should be emphasized that the term “include / comprises” when used herein refers to the presence of features, elements, steps or components, but does not exclude the presence or addition of one or more other features, elements, steps or components.
[0044] It should also be noted that, unless otherwise specified, the term “connection” herein may refer not only to a direct connection but also to an indirect connection involving an intermediate.
[0045] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals represent the same or similar components, or the same or similar steps.
[0046] It should be emphasized here that the step marks mentioned below are not intended to limit the order of the steps, but it should be understood that the steps can be executed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be executed simultaneously.
[0047] In order to solve the problems in the existing text classification technology that manual addition of real labels leads to high time and labor costs, there are challenges of subjectivity and consistency in manual addition of real labels, and the model trained based on manual addition of real labels is not applicable to other fields, the present invention proposes a text classification model training method for Internet social media, such as Figure 1 As shown, the method includes the following steps S101 to S103:
[0048] Step S101: Obtain a first training data set, wherein the first training data set includes a plurality of samples, each sample includes text data obtained from web pages of different polarities; the polarities include positive sentiment tendency polarities and negative sentiment tendency polarities.
[0049] Step S102: Obtain an initial model, wherein the initial model adopts the ERNIE model, and the initial model includes a data preprocessing module and a classification module. Each sample is input into the data preprocessing module in batches for data cleaning and preprocessing operations, and the processed samples are input into the classification module for classification to obtain the prediction results of each sample.
[0050] Step S103: The initial model is trained using the first training data set, and the loss between the prediction result and the true classification label is constructed. The initial model is optimized and updated with the goal of minimizing the loss to obtain the final text classification model. The true classification label includes positive sentiment tendency and negative sentiment tendency.
[0051] like Figure 2 As shown, it is a flowchart of the text classification model training method for Internet social media. It takes whether the web page polarity data with distinct emotional opinions can be collected as a judgment condition, and is divided into two training situations. First, the situation of being able to collect web page polarity data with distinct emotional opinions is further explained.
[0052] In step S101, a first training data set is obtained (constructed) for subsequent model training.
[0053] First, web pages of different polarities are identified. In some embodiments, web pages can be classified based on natural language processing technology and pre-trained web page polarity classification models to obtain web pages with positive sentiment polarity and web pages with negative sentiment polarity. The web page polarity classification model can be obtained based on machine learning or deep learning training, such as support vector machine, naive Bayes classifier, random forest, recurrent neural network, Transformer, etc., or by using pre-trained language models, such as BERT, GPT, etc., or by extracting text features, such as bag-of-words model, TF-IDF, word embedding, etc., to convert text into numerical features that can be understood by machine learning algorithms, and feature selection methods are combined for feature screening.
[0054] Text data is captured from web pages with different polarities as samples to construct a training data set. For example, the actual classification of text data captured from web pages with positive sentiment polarity is positive sentiment, and the actual classification of text data captured from web pages with negative sentiment polarity is negative sentiment.
[0055] In some embodiments, a preset crawler program is used to crawl required text data from polarity web pages according to set rules. Technical means such as HTML parsing, XPath or CSS selectors can also be used to crawl text data.
[0056] In some embodiments, the captured text data may have a high noise content. A preset screening method, such as keyword matching, model recognition, etc., is used to filter the first training data set, build a small batch data set to train the initial model, obtain a preliminary text classification model, and then use the first training data set to train the preliminary text classification model. Among them, in text data, noise generally refers to irrelevant or erroneous information, which is introduced for various reasons, such as low text quality, spelling errors, grammatical errors, non-standard expressions, etc.; subjective bias, the author has personal or organizational subjective bias, which leads to misleading, exaggeration, devaluation of things or opinions, etc.; and spam, etc.
[0057] In some embodiments, the accuracy of the preliminary text classification model obtained by training based on the small batch data set needs to reach a preset value, such as 70%. If it fails to reach the preset value, the training steps need to be repeated and manual verification is used to assist.
[0058] In step S102, an initial model is obtained. Exemplarily, the initial model adopts the ERNIE model. The ERNIE (Enhanced Representation through Knowledge Integration) model is a pre-trained language model launched by Baidu. Its main feature is to integrate the rich semantic knowledge in the knowledge base into the pre-training process, thereby improving the performance of the model. The initial model includes a data preprocessing module and a classification module. Each sample is input into the data preprocessing module in batches for data cleaning and preprocessing operations, and then the processed samples are input into the classification module for classification to obtain the prediction results of each sample.
[0059] In some embodiments, text processing technology is used to perform data cleaning and preprocessing operations on the samples, wherein the text processing technology includes removing stop words, performing stemming, word form restoration, correcting spelling errors, and the like.
[0060] In step S103, the initial model is trained using the first training data set, the loss between the prediction result of each sample and its true classification label is constructed, and the initial model is updated and optimized with the goal of minimizing the loss.
[0061] In some embodiments, considering that the present invention is a binary classification task, a cross entropy loss between the prediction result and the true classification label is constructed, wherein the cross entropy loss calculation formula is shown in formula (1):
[0062]
[0063] Where L represents the cross entropy loss; p(x i ) represents the sample x i The true classification label of q(x i ) represents the sample x i prediction results.
[0064] In some embodiments, Figure 2 As shown, when web page polarity data with distinct emotional viewpoints cannot be collected, the present invention further provides a text classification model training method for Internet social media, comprising the following steps S201 to S203:
[0065] Step S201: Obtain a second training data set, wherein the second training data set includes multiple samples, each sample includes a piece of text data. Randomly obtain samples from the second training data set, construct a small batch training set, and add real classification labels to the small batch training set. Train a preliminary classification model based on a small portion of manually annotated text data.
[0066] Step S202: Obtain an initial model. For example, the initial model uses the Baidu ERNIE model. The initial model includes a data preprocessing module and a classification module. Each sample in the small batch training set is input into the data preprocessing module in batches for data cleaning and preprocessing operations, and then the processed samples are input into the classification module for classification to obtain the prediction results of each sample.
[0067] Step S203: First, the initial model is trained using a small batch training set with real labels added, and the loss between the prediction results and the real classification labels is constructed to obtain a preliminary text classification model; then the preliminary text classification model is trained using a second training data set, and the preliminary text classification model is optimized and updated to obtain a final text classification model.
[0068] In some embodiments, in step S203, a cross entropy loss between the prediction result and the true classification label is also constructed, and the calculation formula is shown in formula (1).
[0069] In some embodiments, each sample in the small batch training set is input into the preliminary text classification model in batches to obtain the prediction results of each sample, and the correctness of the prediction results is judged; preset optimization means are used to screen out samples with incorrect prediction results, and analysis is performed to optimize the performance of the preliminary text classification model. Exemplarily, the preset optimization means include keyword matching, manual verification, etc.
[0070] The present invention also provides a text classification model training device for Internet social media, the device comprising:
[0071] The data collection module is used to collect text data from web pages of different polarities as samples to construct a training data set.
[0072] The result prediction module is used to input each sample in the training data set into the initial model to be trained to obtain the prediction result.
[0073] The model training module is used to train the model according to the training data set, and construct the loss between the predicted results and the corresponding true classification labels, and optimize and update the model with the goal of minimizing the loss.
[0074] The present invention will be further described below in conjunction with a specific embodiment.
[0075] After determining the corresponding field, a preset crawler program is used to crawl about 100,000 related post bar articles and news media data as training data for positive sentiment tendencies, and about 100,000 related post bar articles and news media data are crawled as training data for negative sentiment tendencies. Exemplarily, 2,500 pairs of positive and negative training data are selected as test sets, and the rest are used as training sets. The initial ERNIE model is trained with the training set to obtain a text classification model, and the text classification model is tested with the test set, with an accuracy rate of up to 99%.
[0076] It should be noted that during actual training and testing, the training data of the training set and the test set will be screened and filtered, resulting in differences in the number of training sets and test sets.
[0077] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a text classification model training method for Internet social media.
[0078] Corresponding to the above method, the present invention also provides a device, which includes a computer device, the computer device includes a processor and a memory, the memory stores computer instructions, the processor is used to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the device implements the steps of the method described above.
[0079] The embodiment of the present invention also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of the aforementioned edge computing server deployment method are implemented. The computer-readable storage medium can be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium known in the technical field.
[0080] In summary, the present invention provides a text classification model training method and device for Internet social media, including: obtaining a first training data set, which includes multiple samples, each sample includes text data obtained from web pages of different polarities; obtaining an initial model, the initial model adopts an ERNIE model, and includes a data preprocessing module and a classification module; inputting each sample into the data preprocessing module in batches for data cleaning and preprocessing operations, and inputting the processed samples into the classification module for classification to obtain the prediction results of each sample; using the first training data set to train the initial model, constructing the loss between the prediction result and the true classification label, and optimizing and updating the initial model with the goal of minimizing the loss to obtain the final text classification model. The model training method provided by the present invention can make full use of Internet public data information, and does not completely rely on manual annotation, reducing labor costs and saving development time; the trained text classification model is efficient and accurate; at the same time, according to the classification results, it can analyze public opinion changes in real time and track changes in netizens' opinions.
[0081] It should be understood by those skilled in the art that the exemplary components, systems and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software or a combination of the two. Whether it is performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link via a data signal carried in a carrier.
[0082] It should be clear that the present invention is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present invention.
[0083] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with features of other embodiments or replace features of other embodiments.
[0084] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the embodiments of the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A text classification model training method for Internet social media, It is characterized in that The method comprises the following steps: Acquire a first training data set, wherein the first training data set includes a plurality of samples, each sample includes text data obtained from web pages of different polarities; the polarities include positive sentiment tendency polarities and negative sentiment tendency polarities; Obtaining an initial model, wherein the initial model adopts an ERNIE model, and the initial model includes a data preprocessing module and a classification module; inputting each sample into the data preprocessing module in batches for data cleaning and preprocessing operations, and inputting the processed samples into the classification module for classification to obtain prediction results of each sample; The initial model is trained using the first training data set, and the loss between the prediction result and the true classification label is constructed. The initial model is optimized and updated with the goal of minimizing the loss to obtain a final text classification model; the true classification label includes positive sentiment tendency and negative sentiment tendency.
2. According to the text classification model training method for Internet social media according to claim 1, It is characterized in that Each sample contains text data obtained from web pages of different polarities, and also includes: Based on natural language processing technology and a pre-trained web page polarity classification model, web pages are classified into polarity categories to obtain web pages with positive sentiment tendency polarity and web pages with negative sentiment tendency polarity; the web page polarity classification model is obtained based on machine learning or deep learning training.
3. The text classification model training method for Internet social media according to claim 2, It is characterized in that Each sample contains text data obtained from web pages of different polarities, and also includes: A preset crawler program is used to crawl the required text data from the polarity web pages; wherein the true classification of the text data crawled from the positive sentiment tendency polarity web pages is positive sentiment tendency, and the true classification of the text data crawled from the negative sentiment tendency polarity web pages is negative sentiment tendency.
4. The text classification model training method for Internet social media according to claim 1, It is characterized in that After acquiring the first training data set, the method further includes: The first training data set is filtered using a preset screening method, and a small batch data set is constructed to train the initial model to obtain a preliminary text classification model; the preset screening method includes keyword matching and model recognition.
5. The text classification model training method for Internet social media according to claim 1, It is characterized in that The method further comprises: The sample is subjected to data cleaning and preprocessing operations using text processing techniques; the text processing techniques include removing stop words, performing stemming, lemmatization, and correcting spelling errors.
6. The text classification model training method for Internet social media according to claim 1, It is characterized in that The method further comprises: Construct the cross entropy loss between the prediction result and the true classification label. The cross entropy loss calculation formula is: L=-∑ i p(x i )log(q(x i )); Wherein, L represents the cross entropy loss; p(x i ) represents the sample x i The true classification label of q(x i ) represents the sample x i prediction results.
7. The text classification model training method for Internet social media according to claim 1, It is characterized in that When no relevant polarity web page can be found to obtain text data, the method further includes: Obtain a second training data set, where the second training data set includes multiple samples, each sample includes a piece of text data; randomly obtain samples from the second training data set, construct a small batch training set, and add a true classification label to each sample in the small batch training set; Acquire the initial model, which includes a data preprocessing module and a classification module; input each sample in the small batch training set into the data preprocessing module in batches for data cleaning and preprocessing operations, input the processed samples into the classification module for classification, and obtain the prediction results of each sample; The initial model is trained using the small batch training set, and the loss between the prediction result and the true classification label is constructed to obtain a preliminary text classification model; the preliminary text classification model is then trained using the second training data set, and the preliminary text classification model is optimized and updated to obtain a final text classification model.
8. According to the text classification model training method for Internet social media according to claim 7, It is characterized in that The method further comprises: Input each sample in the small batch training set into the preliminary text classification model in batches, obtain a prediction result for each sample, and determine the correctness of the prediction result; The samples with wrong prediction results are screened and analyzed by using preset optimization means; the preset optimization means include keyword matching and manual verification.
9. A text classification model training device for Internet social media, It is characterized in that The device comprises: A data collection module is used to collect text data from web pages of different polarities as samples to construct a training data set; A result prediction module, used to input each sample in the training data set into the model to be trained to obtain a prediction result; The model training module is used to train the model according to the training data set, and construct the loss between the prediction result and the corresponding true classification label, and optimize and update the model with the goal of minimizing the loss.
10. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.