Consumer behavior motivation classification model training and consumer behavior motivation classification method

Abstract

The invention relates to the technical field of data analysis, and discloses a consumer behavior motivation classification model training and consumer behavior motivation classification method, which comprises the following steps: collecting sample cigarette consumer consumption data, and carrying out motivation priority division on the sample cigarette consumer consumption data; performing feature extraction on the consumption data of the different behavior motivations, and calculating to obtain behavior motivation features corresponding to the consumption data; training each level of vector machine model in the initial cigarette consumer behavior motivation classification model by utilizing the behavior motivation characteristics; comparing output results to obtain the behavior motivation type of the highest probability of the cigarette consumer and the relative importance corresponding to the behavior motivation type; and the vector machine models of all levels are adjusted and optimized. According to the method, the consumer behavior motivation is effectively classified through the cigarette consumer consumption information, the classification capability of the vector machine model is fully utilized, and the characteristics of the cigarette consumer consumption data are combined, so that the cigarette consumer behaviors are efficiently and accurately classified.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and particularly to the training of a consumer behavior motivation classification model and a consumer behavior motivation classification method. Background Art

[0002] With the rapid development of information technology, especially the progress of big data and machine learning technologies, enterprises can now collect and analyze a large amount of data on cigarette consumer behavior. As an important part of the fast-moving consumer goods market, the understanding and classification of cigarette consumer behavior are crucial for formulating effective marketing strategies. However, traditional methods for analyzing cigarette consumer behavior often rely on simple statistical analysis or expert experience, and these methods have limitations in dealing with large-scale and multi-dimensional data, making it difficult to capture the complexity and dynamic changes of cigarette consumer behavior.

[0003] To improve the accuracy and efficiency of classification, more and more research has started to explore the use of machine learning algorithms to analyze cigarette consumer behavior data. Support Vector Machine (SVM), as a powerful classifier, has been proven to have good performance in many fields such as pattern recognition, image processing, and text classification. SVM distinguishes different categories by finding the optimal decision boundary, i.e., the hyperplane, in the feature space. Its core advantage is that it can effectively process data even in high-dimensional spaces and find the best classification boundary.

[0004] In related technologies, although SVM has shown great potential in many fields, its application in classifying cigarette consumer behavior motivations is relatively less. In addition, existing SVM-based classification models often lack in-depth analysis of the complexity of cigarette consumer behavior and may require a large number of parameter adjustments and optimizations in practical applications. Summary of the Invention

[0005] In view of this, the present invention provides a method for training a consumer behavior motivation classification model and a consumer behavior motivation classification method to solve the problem of the lack of existing methods for analyzing cigarette consumer behavior.

[0006] In a first aspect, the present invention provides a method for training a consumer behavior motivation classification model, the method comprising:

[0007] Collecting sample cigarette consumer consumption data and dividing the sample cigarette consumer consumption data into motivation priorities;

[0008] Extracting features from the consumption data of different behavior motivations and calculating the behavior motivation features corresponding to the consumption data;

[0009] Using behavioral motivation characteristics, train the vector machine models at all levels in the initial cigarette consumer behavioral motivation classification model to obtain the output results of the vector machine models at all levels in the initial cigarette consumer behavioral motivation classification model;

[0010] Compare the output results to obtain the behavioral motivation type with the highest probability for cigarette consumers and the relative importance corresponding to the behavioral motivation type;

[0011] Based on the behavioral motivation type and relative importance, optimize the vector machine models at all levels to obtain the target cigarette consumer behavioral motivation classification model.

[0012] In the present invention, the problem of the existing lack of a method for analyzing cigarette consumer behavior is solved. It realizes the effective classification of consumer behavioral motivations through cigarette consumer consumption information, can make full use of the classification ability of the vector machine model, and combines the characteristics of cigarette consumer consumption data to achieve efficient and accurate classification of cigarette consumer behavior. Through the present invention, it can provide deeper market insights for cigarette manufacturers and retailers, help them better understand cigarette consumer needs, optimize product promotion strategies, and improve market competitiveness.

[0013] In an alternative embodiment, using behavioral motivation characteristics, training the vector machine models at all levels in the initial cigarette consumer behavioral motivation classification model to obtain the output results of the vector machine models at all levels in the initial cigarette consumer behavioral motivation classification model includes:

[0014] Input the behavioral motivation characteristics into vector machine models at different levels;

[0015] Based on the distribution of sample cigarette consumer consumption data, determine the kernel function corresponding to the vector machine model training;

[0016] Using the kernel function corresponding to the vector machine model training, control the influence range of a single behavioral motivation characteristic, and train the vector machine models at all levels to obtain the output results of the vector machine models at all levels in the initial cigarette consumer behavioral motivation classification model.

[0017] In this method, by passing in different input features to train the SVM vector machine models at all levels, the fitting degree of the model to the training data and the generalization ability are optimized, the multi-classification problem is solved, and a balance is achieved between model performance and training efficiency.

[0018] In an alternative embodiment, comparing the output results to obtain the behavioral motivation type with the highest probability for cigarette consumers and the relative importance corresponding to the behavioral motivation type includes:

[0019] Take the output result of the current-level vector machine model as the score, and the score is used to represent the relative importance of the behavioral motivation of the current-level vector machine model;

[0020] Taking the score as the conditional probability of the next-layer vector machine model and combining it with the output result of the next-layer vector machine model, the behavior motivation type of the next-layer vector machine model and the relative importance of the next-layer vector machine model are obtained, until the behavior motivation types and relative importance of the vector machine models at all levels in the initial cigarette consumer behavior motivation classification model are obtained.

[0021] In this method, since consumers' purchase decisions are often not driven by a single motivation, but by multiple motivations intertwined. By introducing the concept of the priority of consumer behavior motivations, it is allowed to rank the consumer behavior motivations according to their importance in the decision-making process. The probability of each layer is the relative importance relative to other behavior motivations after excluding the influence of the upper-layer behavior motivations.

[0022] In a second aspect, the present invention provides a method for classifying consumer behavior motivations, the method comprising:

[0023] Obtaining consumption data of cigarette consumers to be classified;

[0024] Inputting the consumption data of cigarette consumers to be classified into the target cigarette consumer behavior motivation classification model to obtain the behavior motivation type corresponding to the consumption data of cigarette consumers to be classified, wherein the target cigarette consumer behavior motivation classification model is trained by using the consumer behavior motivation classification model training method of any item in the first aspect.

[0025] In the present invention, using the cigarette consumer behavior motivation classification model to classify the consumption data of the cigarette consumers to be classified can not only provide the classification result of the consumer behavior motivation, but also provide more in-depth insights for the marketing strategy, helping the enterprise to better understand the needs and preferences of consumers, so as to formulate a more effective market strategy.

[0026] In an optional implementation manner, the method further comprises, when adding a new consumer behavior motivation type, inserting the vector machine model corresponding to the new consumer behavior motivation type into the target cigarette consumer behavior motivation classification model.

[0027] In this way, if a new consumer behavior motivation type needs to be added in the future, the vector machine model corresponding to the new consumer behavior motivation type can be directly inserted into the target cigarette consumer behavior motivation classification model without retraining the entire model. The addition of the new layer will naturally integrate into the existing conditional probability framework, enabling the model to flexibly adapt to market changes and the evolution of consumer behavior.

[0028] In a third aspect, the present invention provides a device for training a consumer behavior motivation classification model, the device comprising:

[0029] A sample data acquisition module, configured to collect consumption data of sample cigarette consumers and classify the priorities of the motivations for the consumption data of sample cigarette consumers;

[0030] A feature extraction module, configured to extract features from the consumption data of different behavioral motivations and calculate the behavioral motivation features corresponding to the consumption data;

[0031] A model training module, configured to use the behavioral motivation features to train each level of vector machine model in the initial cigarette consumer behavioral motivation classification model and obtain the output results of each level of vector machine model in the cigarette consumer behavioral motivation classification model;

[0032] An output comparison module, configured to compare the output results to obtain the behavioral motivation type with the highest probability for cigarette consumers and the relative importance corresponding to the behavioral motivation type;

[0033] A model tuning module, configured to tune each level of vector machine model based on the behavioral motivation type and the relative importance to obtain a target cigarette consumer behavioral motivation classification model.

[0034] Fourthly, the present invention provides a device for classifying consumer behavioral motivations, and the device includes:

[0035] A data acquisition module, configured to acquire consumption data of cigarette consumers to be classified;

[0036] A behavioral motivation classification module, configured to input the consumption data of cigarette consumers to be classified into the target cigarette consumer behavioral motivation classification model to obtain the behavioral motivation type corresponding to the consumption data of cigarette consumers to be classified, where the target cigarette consumer behavioral motivation classification model is trained by using the consumer behavioral motivation classification model training device in the third aspect.

[0037] Fifthly, the present invention provides a computer device, including: a memory and a processor, which are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to execute the consumer behavioral motivation classification model training method in the first aspect or any corresponding embodiment thereof or execute the consumer behavioral motivation classification method in the second aspect or any corresponding embodiment thereof.

[0038] Sixthly, the present invention provides a computer-readable storage medium, on which computer instructions are stored, and the computer instructions are used to cause a computer to execute the consumer behavioral motivation classification model training method in the first aspect or any corresponding embodiment thereof or execute the consumer behavioral motivation classification method in the second aspect or any corresponding embodiment thereof.

[0039] Seventh aspect, the present invention provides a computer program product, including computer instructions for causing a computer to execute the consumer behavior motivation classification model training method according to the first aspect or any corresponding embodiment thereof, or execute the consumer behavior motivation classification method according to the second aspect or any corresponding embodiment thereof. Description of the Drawings

[0040] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0041] Figure 1 It is a flowchart of the consumer behavior motivation classification model training method according to an embodiment of the present invention.

[0042] Figure 2 It is a flowchart of a cigarette consumer behavior motivation classification method based on an SVM classification model according to an embodiment of the present invention.

[0043] Figure 3 It is a schematic diagram of the user distribution of a first-level gift-type SVM user classifier according to an embodiment of the present invention.

[0044] Figure 4 It is a schematic diagram of the user distribution of a second-level curiosity-seeking new-type SVM user classifier according to an embodiment of the present invention.

[0045] Figure 5 It is a schematic diagram of the user distribution of a third-level novelty-seeking, herd-imitation, and practical SVM user classifier according to an embodiment of the present invention.

[0046] Figure 6 It is a flowchart of another consumer behavior motivation classification model training method according to an embodiment of the present invention.

[0047] Figure 7 It is a flowchart of a consumer behavior motivation classification method according to an embodiment of the present invention.

[0048] Figure 8 It is a block diagram of the structure of a consumer behavior motivation classification model training device according to an embodiment of the present invention.

[0049] Figure 9 It is a block diagram of the structure of a consumer behavior motivation classification device according to an embodiment of the present invention.

[0050] Figure 10It is a schematic diagram of the hardware structure of the computer device according to an embodiment of the present invention. Detailed implementation manners

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0052] In related technologies, although SVM has shown great potential in many fields, its application in the classification of cigarette consumer behavior motivations is relatively few. In addition, existing SVM-based classification models often lack in-depth analysis of the complexity of cigarette consumer behavior, and may require a large amount of parameter adjustment and optimization in practical applications.

[0053] To solve the above problems, an embodiment of the present invention provides a method for training a consumer behavior motivation classification model for use in a computer device. It should be noted that the execution subject thereof may be a consumer behavior motivation classification model training device, and this device can be implemented as part or all of the computer device through software, hardware, or a combination of software and hardware. Among them, the computer device may be a terminal, a client, or a server. The server may be a single server or a server cluster composed of multiple servers. The terminal in the embodiments of the present application may be other intelligent hardware devices such as a smart phone, a personal computer, or a tablet computer. In the following method embodiments, the execution subject being a computer device is taken as an example for description.

[0054] The computer device in this embodiment is applicable to the usage scenario of classifying the behavior motivations of cigarette consumers by using cigarette consumer consumption information. By providing the method for training a consumer behavior motivation classification model according to the present invention, the problem of the existing lack of a method for analyzing cigarette consumer behavior is solved, and the effective classification of consumer behavior motivations through cigarette consumer consumption information is realized. It can make full use of the classification ability of the vector machine model and combine the characteristics of cigarette consumer consumption data to achieve efficient and accurate classification of cigarette consumer behavior. Through the present invention, deeper market insights can be provided for cigarette manufacturers and retailers, helping them better understand cigarette consumer needs, optimize product promotion strategies, and improve market competitiveness.

[0055] According to an embodiment of the present invention, an embodiment of a method for training a consumer behavior motivation classification model is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0056] In this embodiment, a method for training a consumer behavior motivation classification model is provided, which can be used in the above-mentioned computer device. Figure 1 is a flowchart of a method for training a consumer behavior motivation classification model according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0057] Step S101, collecting consumption data of sample cigarette consumers, and dividing the consumption data of sample cigarette consumers into motivation priorities.

[0058] In one example, the behavioral motivations of cigarette consumers are prioritized: the consumption information of cigarette consumers is collected.

[0059] For example, the priority of cigarette consumer behavior motivation is divided, specifically including in-depth analysis of consumer consumption data through statistical methods, analysis of consumer group portraits corresponding to each behavior motivation label, observation of their objective consumption patterns, and division of consumer behavior motivation priorities through different consumption characteristics of the target research population. Among them, cigarette consumer consumption information includes but is not limited to: membership code, sales date, cigarette name, cigarette barcode, recommended retail price, selling price, quantity, unit, payment amount, marketing department, retail store license number, and behavior motivation.

[0060] Step S102, extracting features from consumption data of different behavioral motivations, and calculating behavioral motivation features corresponding to the consumption data.

[0061] In one example, different features are extracted for different priorities of cigarette consumer behavior motivations, and the features selected for cigarette consumer behavior motivations include but are not limited to conformity and divergence. Conformity is used to describe the degree to which their behavior is positively influenced by social behavior, that is, the degree to which cigarette consumers respond positively to social consistency behavior; divergence is used to describe the degree to which their behavior is negatively influenced by social behavior, that is, the degree to which cigarette consumers respond negatively to social consistency behavior.

[0062] For example, the conformity degree is the total degree to which all the purchasing behaviors of the cigarette consumer are positively influenced, which specifically includes: assuming that member i buys cigarette specification k at a certain store j, the sales volume of the specification in the store is recorded as S jk , the store's maximum cigarette sales volume is recorded as The popularity of this cigarette purchase is The member's purchase amount this time is recorded as B ijk , the degree of impact on the member's consumption behavior this time is The member's conformity includes:

[0063] The degree of difference is the total degree to which all purchasing behaviors of the cigarette consumer are affected by the reverse influence, which specifically includes: Assuming that member i buys cigarette specification k at a certain store j, the sales volume of the specification in the store is recorded as S jk , the store’s minimum cigarette sales volume is recorded as The niche degree of this cigarette purchase is The member's purchase amount this time is recorded as B ijk , the degree of impact on the member's consumption behavior this time is Then the member's different degree is:

[0064]

[0065] Step S103, using the behavior motivation features, training the vector machine models at all levels in the initial cigarette consumer behavior motivation classification model to obtain output results of the vector machine models at all levels in the initial cigarette consumer behavior motivation classification model.

[0066] In one example, various levels of SVM models are trained, and SVM models are trained respectively for each single or multiple behavioral motivations calculated above.

[0067] Step S104, comparing the output results to obtain the highest probability behavior motivation type of the cigarette consumer and the relative importance corresponding to the behavior motivation type.

[0068] In one example, by comparing the output results of the SVM model at each level, we can obtain the most likely behavioral motivation type of cigarette consumers and its relative importance in each level of classification. This will prevent the judgment of the behavioral motivation at the next level from being affected by the excessively high probability of the behavioral motivation classified at the previous level, and will allow us to capture consumer behavioral motivations in a more detailed manner.

[0069] Exemplarily, the output of each layer of the model can be regarded as a score, which reflects the relative importance of the behavioral motivation of the current layer when the behavioral motivation of the previous layer of consumers is known. For example, if the first-layer model identifies the probability of a certain consumer behavioral motivation A, then in the second-layer model, the output corresponding to this non-behavioral motivation A will be used as a conditional probability and will not affect the judgment of other behavioral motivations. In this way, even if the priority of the first layer is higher, it will not cause the judgment of the second layer to lose balance, ensuring the diversity and accuracy of the model output.

[0070] Step S105: Optimize each level of vector machine models based on the behavior motivation type and relative importance to obtain the target cigarette consumer behavior motivation classification model.

[0071] In one example, the accuracy of each level of models is improved through continuous model optimization to obtain an available multi-level cigarette consumer behavior motivation classification model.

[0072] Specifically, the optimization of the model may include:

[0073] S5-1: Try to use different kernel functions, such as linear kernel, polynomial kernel, and radial basis function (RBF) kernel, etc.

[0074] S5-2: Hyperparameter optimization: Optimize the hyperparameters in the SVM model, including the regularization parameter C and kernel function parameters (such as the gamma value in the RBF kernel), etc.

[0075] S5-3: Use the test set to evaluate the performance of the trained SVM model and optimize the model according to the evaluation results.

[0076] In an implementation scenario, Figure 2 is a schematic flowchart of a method for classifying cigarette consumer behavior motivation based on an SVM hierarchical model according to an embodiment of the present invention. As Figure 2 shown, the method for classifying cigarette consumer behavior motivation based on an SVM hierarchical model includes:

[0077] S1: Divide the priority of cigarette consumer behavior motivation: Collect cigarette consumer consumption information.

[0078] Specifically, S1 includes: For mid- to high-end cigarette consumers, the gift-giving type and novelty-seeking type motivations occupy a relatively high priority. Gift-giving type consumers value more the social value and personal image enhancement brought by the brand when purchasing, which is in line with their relatively high economic strength. They tend to choose brands that can reflect their social status and taste. Novelty-seeking type consumers are more sensitive to novel and fashionable products and are willing to pay a higher price for unique consumption experiences.

[0079] While the behavior motivation priorities of the individualistic type, conformist type, and practical type consumers are set relatively low. This is because individualistic type consumers pursue individuality, conformist type consumers prefer popular products, and practical type consumers pay more attention to cost performance. These motivations deviate to a certain extent from the core needs of mid- to high-end cigarette consumers, who attach more importance to brand value and consumption experience.

[0080] Exemplarily, Figure 3 is a schematic diagram of the user distribution of a first-level gift-giving type SVM user classifier according to an embodiment of the present invention. Figure 4It is a schematic diagram of the user distribution of a second-level novelty-seeking SVM user classifier according to an embodiment of the present invention. Figure 5 It is a schematic diagram of the user distribution of a third-level dissimilarity-seeking, conformist-imitating, and practical SVM user classifier according to an embodiment of the present invention.

[0081] As Figure 3 , 4 , shown in 5, in summary, the priority of consumer behavior motivation: gift type > novelty-seeking type > dissimilarity-seeking type, conformist type, practical type.

[0082] Regarding the features selected for different priorities, the key lies in specifically what available consumption data there is. Through statistical methods, in-depth analysis of consumers' consumption data is carried out, the portraits of consumer groups corresponding to each behavior motivation label are analyzed, their objective consumption laws are observed, and model features are scientifically selected based on this to ensure that the diversity and complexity of consumers' behaviors can be accurately captured. Based on the basis of data analysis, for gift-type users, the influence of their social needs is identified by comparing their purchase behaviors during holidays and on weekdays; for novelty-seeking and dissimilarity-seeking users, the reason for this feature selection is mainly based on consumers' behaviors and preferences. And the three types of cigarettes, brand A (specification A), brand B (specification B), and brand C (specification C), entering a certain market as out-of-town cigarettes have novelty and uniqueness and can attract the attention of this part of consumers; for dissimilarity-seeking, conformist, and practical users, we measure their responses to social behaviors through the degree of conformity and the degree of dissimilarity.

[0083] Collecting consumers' consumption information includes but is not limited to the unique ID of cigarette consumers, the unique ID of the store, the sales time, the code of the purchased cigarettes, the name of the cigarettes, the recommended retail price of the cigarettes, the quantity, the unit, the amount of this consumption, the area where the store belongs, the store level, the store format, the store address, the membership behavior motivation, the number of boxes purchased during holidays, the number of boxes purchased on weekdays, the average number of boxes per day during holidays, the average number of boxes per day on weekdays, the total number of boxes consumed, the sales volume of the purchased cigarette specifications in the purchased store, the maximum sales volume of the purchased cigarette specifications in the purchased store, the minimum sales volume of the purchased cigarette specifications in the purchased store, and the total number of boxes consumed of specific specifications purchased by cigarette consumers, such as brand A (specification A), brand B (specification B), and brand C (specification C). Such specific specifications are newly launched cigarette specifications.

[0084] S2 Extract features: Extract different features for the behavior motivations of cigarette consumers with different priorities, specifically including:

[0085] Dense deviation degree. The ratio of the number of boxes purchased by cigarette consumers during holidays to the sum of the number of boxes purchased during holidays and on weekdays:

[0086]

[0087] Total weight. The ratio of the average number of cigarette packs per day consumed by cigarette consumers to the sum of the average number of cigarette packs per day during holidays and the average number of cigarette packs per day during normal days:

[0088]

[0089] Ratio of the number of packs of a particular brand of cigarettes consumed to the total number of packs consumed:

[0090]

[0091] Conformity is the total degree to which all purchasing behaviors of the cigarette consumer are positively influenced, which specifically includes: Assuming that member i buys cigarette specification k at a certain store j, the sales volume of the specification in the store is recorded as S jk , the store's maximum cigarette sales volume is recorded as The popularity of this cigarette purchase is The member's purchase amount this time is recorded as B ijk , the degree of impact on the member's consumption behavior this time is The member's conformity includes:

[0092] The degree of difference is the total degree to which all purchasing behaviors of the cigarette consumer are affected by the reverse influence, which specifically includes: Assuming that member i buys cigarette specification k at a certain store j, the sales volume of the specification in the store is recorded as S jk , the store’s minimum cigarette sales volume is recorded as The niche degree of this cigarette purchase is The member's purchase amount this time is recorded as B ijk , the degree of impact on the member's consumption behavior this time is Then the member's different degree is:

[0093]

[0094] The above features are used as input parameters for the prediction results of the consumer behavior motivation classification model. The first-level SVM model is used to solve the classification problem of whether the member is a gift type, and its input features are [intensive weight, total weight]; the second-level SVM model is used to solve the classification problem of whether the member is curious about seeking new types, and its input features are the proportion of the number of consumption boxes of a specific brand of cigarettes [brand A (specification A), brand B (specification B), brand C (specification C)] to the total number of consumption boxes; the third-level SVM model is used to solve the classification problem of distinguishing between members seeking differences, following the crowd and imitating, and seeking reality, and its input features are [following the crowd, seeking differences].

[0095] S3. Train SVM models at all levels: train SVM models for each single or multiple behavioral motivations separately.

[0096] S4. Compare the output results of the SVM models at each level, and obtain the type of behavioral motivation with the highest probability for cigarette consumers, as well as the relative importance in each layer of classification. The judgment of the behavioral motivation in the next layer will not be affected by the overly high probability of the behavioral motivation classified in the previous layer.

[0097] S5. Improve the accuracy of the models at each level through continuous model tuning, and obtain an available multi-level cigarette consumer behavioral motivation classification model. In the future, if new types of consumer behavioral motivations need to be added, they can be directly inserted into the model as a new level without retraining the entire model.

[0098] The training method for the consumer behavioral motivation classification model provided in this embodiment solves the problem of the existing lack of methods for analyzing cigarette consumer behavior, realizes the effective classification of consumer behavioral motivations through cigarette consumer consumption information, can make full use of the classification ability of the vector machine model, and combines the characteristics of cigarette consumer consumption data to achieve efficient and accurate classification of cigarette consumer behavior. Through the present invention, more in-depth market insights can be provided for cigarette manufacturers and retailers, helping them better understand cigarette consumer needs, optimize product promotion strategies, and improve market competitiveness.

[0099] In this embodiment, a training method for a consumer behavioral motivation classification model is provided, which can be used in the above-mentioned computer device. Figure 6 It is a flowchart of another training method for a consumer behavioral motivation classification model according to an embodiment of the present invention, as Figure 6 shown. The process includes the following steps:

[0100] Step S601, collect sample cigarette consumer consumption data, and divide the sample cigarette consumer consumption data according to the motivation priority. For details, please refer to Figure 1 step S101 of the embodiment shown, which will not be elaborated here.

[0101] Step S602, extract features from the consumption data of different behavioral motivations, and calculate the behavioral motivation features corresponding to the consumption data. For details, please refer to Figure 1 step S102 of the embodiment shown, which will not be elaborated here.

[0102] Step S603, use the behavioral motivation features to train the vector machine models at all levels in the initial cigarette consumer behavioral motivation classification model, and obtain the output results of the vector machine models at all levels in the initial cigarette consumer behavioral motivation classification model.

[0103] Specifically, the above step S603 includes:

[0104] Step S6031, input the behavioral motivation features into the vector machine models at different levels.

[0105] Step S6032: Determine the kernel function corresponding to the training of the support vector machine model based on the distribution of the consumption data of sample cigarette consumers.

[0106] Step S6033: Use the kernel function corresponding to the support vector machine model training to control the influence range of a single behavioral motivation feature, and train each level of the support vector machine model to obtain the output results of each level of the support vector machine model in the initial cigarette consumer behavioral motivation classification model.

[0107] In one example, each level of SVM model is trained with different input features as follows:

[0108] S3-1: Use the conformity degree and difference-seeking degree corresponding to each consumer as the input parameters for the training of the third-level SVM model; other levels of SVM models use other features as input parameters.

[0109] S3-2: Select a non-linear kernel or a linear kernel for training according to whether the distribution of users is linearly separable;

[0110] S3-3: Control the influence range of a single training sample in the RBF kernel to optimize the fitting degree and generalization ability of the model to the training data. (The core idea of the RBF kernel is that the similarity between samples decreases rapidly as the distance between them increases, and a Gaussian function is usually used to achieve this effect. Briefly, the RBF kernel maps the input space to a higher-dimensional feature space, making the originally linearly inseparable data linearly separable in the new space).

[0111] S3-4: Enable probability estimation and use the default termination criterion to solve multi-classification problems and help balance between model performance and training efficiency.

[0112] In this method, each level of SVM support vector machine model is trained by passing in different input features, optimizing the fitting degree and generalization ability of the model to the training data, solving multi-classification problems, and achieving a balance between model performance and training efficiency.

[0113] Step S604: Compare the output results to obtain the behavioral motivation type with the highest probability of cigarette consumers and the relative importance corresponding to the behavioral motivation type.

[0114] Specifically, the above-mentioned step S604 includes:

[0115] Step S6041: Use the output result of the current-level support vector machine model as the score, and the score is used to represent the relative importance of the behavioral motivation of the current-level support vector machine model.

[0116] Step S6042: Use the score as the conditional probability of the next-layer vector machine model, and combine it with the output result of the next-layer vector machine model to obtain the behavior motivation type of the next-layer vector machine model and the relative importance of the next-layer vector machine model, until the behavior motivation types and relative importance of the vector machine models at all levels in the initial cigarette consumer behavior motivation classification model are obtained.

[0117] In one example, the output of each level of SVM model not only includes probability estimation but also conditional probability. Specifically: The output of each layer of the model can be regarded as a score, which reflects the relative importance of the current-level behavior motivation given the previous-level consumer behavior motivation. For example, if the first-layer model identifies the probability of a certain consumer behavior motivation A, then in the second-layer model, the output corresponding to the non-behavior motivation A will be used as the conditional probability and will not affect the judgment of other behavior motivations. In this way, even if the priority of the first layer is relatively high, it will not cause the judgment of the second layer to lose balance, ensuring the diversity and accuracy of the model output.

[0118] In this method, since consumers' purchase decisions are often not driven by a single motivation but by multiple motivations intertwined. By introducing the concept of the priority of consumer behavior motivations, it is allowed to rank the consumer behavior motivations according to their importance in the decision-making process. The probability of each layer is the relative importance relative to other behavior motivations after excluding the influence of the upper-layer behavior motivations.

[0119] Step S605: Based on the behavior motivation type and relative importance, optimize each level of vector machine model to obtain the target cigarette consumer behavior motivation classification model. For details, please refer to Figure 1 Step S105 in the illustrated embodiment, which will not be elaborated here.

[0120] The consumer behavior motivation classification model training method provided in this embodiment trains each level of SVM vector machine model by inputting different input features, optimizes the fitting degree and generalization ability of the model to the training data, solves the multi-classification problem, and achieves a balance between model performance and training efficiency. Since consumers' purchase decisions are often not driven by a single motivation but by multiple motivations intertwined. By introducing the concept of the priority of consumer behavior motivations, it is allowed to rank the consumer behavior motivations according to their importance in the decision-making process. The probability of each layer is the relative importance relative to other behavior motivations after excluding the influence of the upper-layer behavior motivations.

[0121] In this embodiment, a consumer behavior motivation classification method is provided, which can be used in the above computer device. Figure 7 It is a flowchart of a consumer behavior motivation classification method according to an embodiment of the present invention. As Figure 7 shown, the process includes the following steps:

[0122] Step S701: Obtain the consumption data of cigarette consumers to be classified.

[0123] Step S702: Input the consumption data of cigarette consumers to be classified into the target cigarette consumer behavior motivation classification model to obtain the behavior motivation type corresponding to the consumption data of cigarette consumers to be classified.

[0124] In the embodiment of the present invention, the target cigarette consumer behavior motivation classification model is trained by using the above-mentioned consumer behavior motivation classification model training method.

[0125] Step S703: When adding a new consumer behavior motivation type, insert the vector machine model corresponding to the new consumer behavior motivation type into the target cigarette consumer behavior motivation classification model.

[0126] In one example, if a new consumer behavior motivation type needs to be added in the future, it can be directly inserted as a new layer into the model without retraining the entire model. The addition of the new layer will naturally integrate into the existing conditional probability framework, enabling the model to flexibly adapt to market changes and the evolution of consumer behavior.

[0127] The consumer behavior motivation classification method provided in this embodiment uses the cigarette consumer behavior motivation classification model to classify the consumption data of cigarette consumers to be classified, which can not only provide the classification result of consumer behavior motivation, but also provide more in-depth insights for marketing strategies, helping enterprises better understand the needs and preferences of consumers, so as to formulate more effective marketing strategies.

[0128] In this embodiment, a consumer behavior motivation classification model training device is also provided. This device is used to implement the above-mentioned embodiment and preferred implementation manners, and those that have been described will not be repeated here. As used below, the term "module" can be a combination of software and / or hardware that can achieve a predetermined function. Although the device described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0129] This embodiment provides a consumer behavior motivation classification model training device, as Figure 8 shown, including:

[0130] The sample data acquisition module 801 is used to collect the consumption data of sample cigarette consumers and divide the motivation priorities of the consumption data of sample cigarette consumers. For details, please refer to Figure 1 Step S101 of the embodiment shown here, which will not be repeated here.

[0131] The feature extraction module 802 is used to extract features from consumption data with different behavioral motivations and calculate the behavioral motivation features corresponding to the consumption data. For details, please refer to Figure 1 Step S102 of the embodiment shown in the figure, which will not be elaborated here.

[0132] The model training module 803 is used to train each level of the vector machine model in the initial cigarette consumer behavioral motivation classification model by using the behavioral motivation features, and obtain the output results of each level of the vector machine model in the cigarette consumer behavioral motivation classification model. For details, please refer to Figure 1 Step S103 of the embodiment shown in the figure, which will not be elaborated here.

[0133] The output comparison module 804 is used to compare the output results to obtain the behavioral motivation type with the highest probability for cigarette consumers and the relative importance corresponding to the behavioral motivation type. For details, please refer to Figure 1 Step S104 of the embodiment shown in the figure, which will not be elaborated here.

[0134] The model tuning module 805 is used to tune each level of the vector machine model based on the behavioral motivation type and relative importance to obtain the target cigarette consumer behavioral motivation classification model. For details, please refer to Figure 1 Step S105 of the embodiment shown in the figure, which will not be elaborated here.

[0135] In some optional embodiments, the model training module 803 includes:

[0136] The feature input unit is used to input the behavioral motivation features into different levels of the vector machine model.

[0137] The kernel function determination unit is used to determine the kernel function corresponding to the training of the vector machine model based on the distribution of the consumption data of sample cigarette consumers.

[0138] The model training unit is used to use the kernel function corresponding to the training of the vector machine model to control the influence range of a single behavioral motivation feature and train each level of the vector machine model to obtain the output results of each level of the vector machine model in the initial cigarette consumer behavioral motivation classification model.

[0139] In some optional embodiments, the output comparison module 804 includes:

[0140] The score determination unit is used to use the output result of the current level of the vector machine model as the score, and the score is used to represent the relative importance of the behavioral motivation of the current level of the vector machine model.

[0141] A conditional probability determination unit, which is used to use the score as the conditional probability of the next-layer vector machine model, and combine the output result of the next-layer vector machine model to obtain the behavior motivation type of the next-layer vector machine model and the relative importance of the next-layer vector machine model, until the behavior motivation type and relative importance of each-level vector machine model in the initial cigarette consumer behavior motivation classification model are obtained.

[0142] The further function descriptions of the above-mentioned various modules and units are the same as those in the corresponding embodiments above, and will not be elaborated here.

[0143] The consumer behavior motivation classification model training device in this embodiment is presented in the form of functional units. Here, the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and a memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0144] In this embodiment, a consumer behavior motivation classification device is also provided. This device is used to implement the above-mentioned embodiments and preferred implementation manners, and those that have been described will not be elaborated again. As used below, the term "module" can be a combination of software and / or hardware that can achieve a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0145] This embodiment provides a consumer behavior motivation classification device, as Figure 9 shown, including:

[0146] A data acquisition module 901, which is used to acquire consumption data of cigarette consumers to be classified. For details, please refer to Figure 7 step S701 of the embodiment shown, which will not be elaborated here.

[0147] A behavior motivation classification module 902, which is used to input the consumption data of cigarette consumers to be classified into the target cigarette consumer behavior motivation classification model to obtain the behavior motivation type corresponding to the consumption data of cigarette consumers to be classified, where the target cigarette consumer behavior motivation classification model is trained by using the above-mentioned consumer behavior motivation classification model training device. For details, please refer to Figure 7 step S702 of the embodiment shown, which will not be elaborated here.

[0148] In some alternative implementation manners, the consumer behavior motivation classification device includes:

[0149] A model addition unit, which is used to insert the vector machine model corresponding to the new consumer behavior motivation type into the target cigarette consumer behavior motivation classification model when a new consumer behavior motivation type is added.

[0150] The further function descriptions of the above-mentioned various modules and units are the same as those in the corresponding above-mentioned embodiments, and will not be elaborated here.

[0151] The consumer behavior motivation classification device in this embodiment is presented in the form of functional units. Here, the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and a memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0152] The embodiment of the present invention further provides a computer device having the above-mentioned Figure 8 shown consumer behavior motivation classification model training device and Figure 9 shown consumer behavior motivation classification device.

[0153] Please refer to Figure 10 , Figure 10 which is a schematic structural diagram of a computer device provided by an optional embodiment of the present invention. As Figure 10 shown, the computer device includes: one or more processors 10, a memory 20, and an interface for connecting each component, including a high-speed interface and a low-speed interface. Each component communicates with each other using different buses and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed within the computer device, including instructions stored in the memory or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (such as a server array, a set of blade servers, or a multi-processor system). Figure 10 In

[0154]

[0155] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above-mentioned embodiments.

[0156] The memory 20 may include a program storage area and a data storage area. Among them, the program storage area can store an operating system and application programs required for at least one function; the data storage area can store data created according to the use of the computer device, etc. In addition, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some alternative embodiments, the memory 20 may optionally include a memory remotely disposed relative to the processor 10, and these remote memories can be connected to the computer device through a network. Examples of the above-mentioned network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0157] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, a hard disk, or a solid-state drive; the memory 20 may further include a combination of the above types of memory.

[0158] The computer device further includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30, and the output device 40 can be connected through a bus or other means. Figure 10 Taking connection through a bus as an example.

[0159] The input device 30 can receive input digital or character information and generate key signal inputs related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a trackpad, a touchpad, a pointing stick, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (such as an LED), and a haptic feedback device (such as a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light-emitting diode, a display, and a plasma display. In some alternative embodiments, the display device may be a touch screen.

[0160] Embodiments of the present invention also provide a computer-readable storage medium. The method according to the embodiments of the present invention can be implemented in hardware, firmware, or be implemented as computer code that can be recorded on a storage medium, or be implemented as computer code that is originally stored in a remote storage medium or a non-transitory machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored as such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory, a random access memory, a flash memory, a hard disk, or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code, and when the software or computer code is accessed and executed by the computer, the processor, or the hardware, the method shown in the above embodiments is implemented.

[0161] A part of the present invention can be applied as a computer program product, for example, computer program instructions, when executed by a computer, through the operation of the computer, can call or provide the method and / or technical solution according to the present invention. Those skilled in the art should be able to understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Herein, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible by the computer.

[0162] Although the embodiments of the present invention are described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for training a consumer behavior motivation classification model, characterized in that: The method comprises: Collecting consumption data of sample cigarette consumers, and prioritizing the consumption data of the sample cigarette consumers according to motivations; Extracting features of consumption data with different behavioral motivations, and calculating the behavioral motivation features corresponding to the consumption data; Using the behavior motivation features, training vector machine models at all levels in the initial cigarette consumer behavior motivation classification model to obtain output results of the vector machine models at all levels in the initial cigarette consumer behavior motivation classification model; Comparing the output results, obtaining the highest probability behavior motivation type of cigarette consumers and the relative importance corresponding to the behavior motivation type; Based on the behavior motivation type and the relative importance, the vector machine models at each level are optimized to obtain a target cigarette consumer behavior motivation classification model.

2. The method according to claim 1, characterized in that The method of using the behavior motivation features to train vector machine models at various levels in the initial cigarette consumer behavior motivation classification model to obtain output results of the vector machine models at various levels in the initial cigarette consumer behavior motivation classification model includes: Inputting the behavior motivation features into the vector machine models at different levels; Based on the distribution of the sample cigarette consumer consumption data, determining the kernel function corresponding to the vector machine model training; The vector machine model is used to train the corresponding kernel function, control the influence range of a single behavior motivation feature, train the vector machine models at all levels, and obtain the output results of the vector machine models at all levels in the initial cigarette consumer behavior motivation classification model.

3. The method according to claim 1, characterized in that The comparing the output results to obtain the behavior motivation type of the cigarette consumer with the highest probability and the relative importance corresponding to the behavior motivation type includes: The output result of the current LVM model is used as a score, where the score is used to characterize the relative importance of the behavior motivation of the current LVM model; The score is used as the conditional probability of the next layer of vector machine model, and combined with the output result of the next layer of vector machine model, the behavioral motivation type of the next layer of vector machine model and the relative importance of the next layer of vector machine model are obtained, until the behavioral motivation type and relative importance of the vector machine models at each level in the initial cigarette consumer behavioral motivation classification model are obtained.

4. A method for classifying consumer behavior motivations, characterized in that: The method comprises: Obtaining consumer consumption data of cigarettes to be classified; The consumption data of the cigarette consumers to be classified are input into a target cigarette consumer behavior motivation classification model to obtain the behavior motivation type corresponding to the consumption data of the cigarette consumers to be classified, wherein the target cigarette consumer behavior motivation classification model is trained using the consumer behavior motivation classification model training method described in any one of claims 1 to 3.

5. The method according to claim 4, characterized in that The method further comprises, when adding a new consumer behavior motivation type, inserting a vector machine model corresponding to the new consumer behavior motivation type into the target cigarette consumer behavior motivation classification model.

6. A consumer behavior motivation classification model training device, characterized in that: The device comprises: A sample data acquisition module, used to collect sample cigarette consumer consumption data and to prioritize the sample cigarette consumer consumption data according to motivations; A feature extraction module is used to extract features from consumption data of different behavioral motivations and calculate the behavioral motivation features corresponding to the consumption data; A model training module, used to train the vector machine models at various levels in the initial cigarette consumer behavior motivation classification model using the behavior motivation features, and obtain output results of the vector machine models at various levels in the initial cigarette consumer behavior motivation classification model; An output comparison module, used to compare the output results, obtain the behavior motivation type with the highest probability of cigarette consumers and the relative importance corresponding to the behavior motivation type; The model tuning module is used to tune the vector machine models at all levels based on the behavior motivation type and the relative importance to obtain a target cigarette consumer behavior motivation classification model.

7. A consumer behavior motivation classification device, characterized in that: The device comprises: A data acquisition module, used to acquire consumption data of cigarette consumers to be classified; A behavior motivation classification module is used to input the consumption data of the cigarette consumers to be classified into a target cigarette consumer behavior motivation classification model to obtain the behavior motivation type corresponding to the consumption data of the cigarette consumers to be classified, wherein the target cigarette consumer behavior motivation classification model is trained using the consumer behavior motivation classification model training device described in claim 6.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the consumer behavior motivation classification model training method described in any one of claims 1 to 3 or the consumer behavior motivation classification method described in any one of claims 4 to 5 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, which are used to enable a computer to execute the consumer behavior motivation classification model training method described in any one of claims 1 to 3 or the consumer behavior motivation classification method described in any one of claims 4 to 5.

10. A computer program product, characterized in that It includes computer instructions, which are used to enable a computer to execute the consumer behavior motivation classification model training method described in any one of claims 1 to 3 or the consumer behavior motivation classification method described in any one of claims 4 to 5.

Images