Product promotion method and system based on multi-source data

Abstract

The invention provides a product promotion method and system based on multi-source data, and relates to the technical field of e-commerce promotion, and the method comprises the steps: obtaining multi-source data in an e-commerce platform, extracting a product use scene and a function demand point through employing a natural language processing technology, and carrying out the promotion of the multi-source data; and the scene coverage rate and the pain point degree value are calculated based on the scene-function incidence matrix, and the priority of the function demand is evaluated. And further performing mapping matching on the function parameter information of the target electronic product and the function demand point, performing importance sorting on the function characteristics according to the scene coverage rate and the pain point degree value, and generating a hierarchical product core selling point sequence. Based on the sequence, a progressive display mode is adopted to generate promotion content, and a scenarized promotion scheme is formed in combination with a scene case, so that the problems of insufficient utilization of multi-source data, unclear promotion content hierarchy and insufficient accuracy in the prior art can be solved, and the promotion effect and the user experience are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of e-commerce promotion technology, and more specifically, to a product promotion method and system based on multi-source data. Background Technology

[0002] With the rapid development of e-commerce, consumers are increasingly relying on online channels to purchase goods, and the promotion methods for electronic products, as an important part of the consumer market, are constantly evolving. From traditional advertising and marketing to targeted advertising based on big data analysis, technological advancements have greatly improved promotion efficiency. However, current product promotion methods largely rely on user behavior data and consumption records, neglecting the deeper needs revealed by user feedback. For example, user reviews, after-sales problem descriptions, and inquiries on e-commerce platforms contain rich information about usage scenarios and functional requirements, but this unstructured data is often overlooked or underutilized in existing promotion methods. Furthermore, traditional promotional content generation methods lack in-depth analysis of users' contextual needs, making it difficult to accurately match users' potential pain points and functional requirements, thus weakening the targeting and effectiveness of promotions.

[0003] Existing technologies for product promotion using multi-source data, including review texts, after-sales issues, and user consultation data, have the following shortcomings: First, existing methods have limited processing capabilities for unstructured data, making it difficult to extract high-quality scenarios and functional requirements. Second, the generation of promotional content lacks hierarchical and scenario-based design, failing to highlight key functional features and their application value in typical scenarios. Finally, the failure to comprehensively consider scenario coverage, user pain points, and functional synergy results in an unscientific and unreasonable prioritization of promotional content. These shortcomings significantly limit the effectiveness of existing methods in improving user experience and optimizing promotional results. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention is proposed. This invention provides a product promotion method and system based on multi-source data, which can, to some extent, solve the problems of congestion caused by uniformly played evacuation advertisements due to the inability to accurately determine the distribution of people in different areas, and the inability to receive evacuation instructions in some areas due to poor indoor positioning signals.

[0005] According to one aspect of the present invention, a product promotion method based on multi-source data is provided, which includes: acquiring review data, after-sales issues and user inquiries of electronic products from e-commerce platforms, and extracting product usage scenarios and functional requirements through natural language processing technology;

[0006] Based on the product usage scenarios and functional requirements, a scenario-function association matrix is ​​established, and based on the feature distribution of the scenario-function association matrix, the scenario coverage rate and pain point intensity value of each functional requirement are calculated.

[0007] Obtain the functional parameter information of the target electronic product, map and match the functional parameter information with the functional requirements, sort the functional characteristics of the target electronic product by importance according to the scenario coverage and the pain point degree value, and generate a sequence of core selling points of the product.

[0008] Based on the product's core selling points, promotional content is generated using a progressive display approach, prioritizing the display of features that address high pain points, and combining these with usage scenario examples to form a scenario-based promotional plan.

[0009] Furthermore, the scenario-function association matrix establishes a correspondence between scenario elements and functional requirements, and the matrix weights are calculated by co-occurrence frequency and adjusted by sentiment analysis to adjust the association strength.

[0010] Furthermore, the pain point severity value is calculated based on the scenario-function correlation matrix, and is expressed as:

[0011]

[0012] in, Indication Function The degree of pain point, Representing a scene Next function The intensity of negative emotions, Indicates the degree of impact of the problem;

[0013] The intensity of negative emotion is calculated as the ratio of the number of negative evaluations to the total number of evaluations, combined with the function. In the scene The associated value Weighted adjustment relative to the total correlation value across all scenarios;

[0014] The impact of the problem is comprehensively considered. In the scene The ratio of task completion time to baseline completion time, and related values. The positive correlation with the degree of influence was obtained through calculation.

[0015] Furthermore, based on the functional parameter information of the target electronic product and the functional requirements, a mapping and matching process is performed to examine the correspondence between each specific functional parameter and the user's requirements, and the basic mapping strength is calculated.

[0016] Furthermore, based on the aforementioned basic mapping strength, and taking into account the comparison between the product's functional parameters and the average value of similar products in the market, a functional satisfaction index is constructed. On this basis, combined with scenario coverage and pain point severity values, a comprehensive importance index is calculated.

[0017] Furthermore, based on the aforementioned functional satisfaction index, a functional synergy index is introduced to measure the synergistic effect between functions, and the group influence of functions is calculated based on the synergy index.

[0018] Furthermore, the functional synergy index is expressed by the formula:

[0019]

[0020] in, Indication Function With function The synergy index between them ranges from [0,1]. and They represent the products. Medium function and functions Satisfaction index and They represent Scene coverage This indicates the total number of reference products in the set.

[0021] Furthermore, based on the importance of the group influence assessment function in the product system, a trend gain factor is introduced, and combined with the expected market growth rate and forecast period of the function, a higher future weight is given to the high group influence function.

[0022] By taking into account the fundamental importance, development trends, and group influence, the overall importance score of the function as a core selling point is calculated, and standardized weight coefficients are generated to construct a sequence of core selling points.

[0023] Furthermore, based on the core selling point sequence, the highest priority function is selected as the primary focus of promotion among the main core selling points. The usage scenario feature vector of each function is extracted, and the most representative typical scenarios are selected by calculating Euclidean distance for product promotion.

[0024] According to another aspect of the present invention, a product promotion system based on multi-source data is provided, comprising:

[0025] The data collection module acquires review data, after-sales issues, and user inquiries for electronic products from e-commerce platforms, and extracts product usage scenarios and functional requirements through natural language processing technology.

[0026] The association module establishes a scenario-function association matrix based on the product usage scenarios and functional requirements, and calculates the scenario coverage rate and pain point severity value of each functional requirement based on the feature distribution of the scenario-function association matrix.

[0027] The matching module acquires the functional parameter information of the target electronic product, maps and matches the functional parameter information with the functional requirements, sorts the functional characteristics of the target electronic product by importance according to the scenario coverage and the pain point degree value, and generates a sequence of core selling points of the product.

[0028] The promotion module generates promotional content based on the product's core selling points sequence, using a progressive display method. It prioritizes showcasing features that address high pain points and combines them with usage scenario examples to form a scenario-based promotional plan.

[0029] Compared to existing technologies, this invention has significant advantages in utilizing multi-source data and precisely designing promotional content. By deeply mining user reviews, after-sales issues, and consultation data, and combining this with natural language processing technology, it accurately extracts usage scenarios and functional requirements from user feedback, significantly improving the depth and accuracy of data processing. Based on an evaluation method that considers scenario coverage and pain point intensity, it scientifically determines the priority of functional requirements, overcoming the shortcomings of existing technologies in terms of unclear promotional content hierarchy and insufficient targeting. Furthermore, by generating a hierarchical sequence of core selling points and combining it with a scenario-based display strategy, this invention not only improves the matching degree between promotional content and user needs but also significantly enhances user experience and promotional effectiveness. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0031] Figure 1 This is a flowchart of a product promotion method based on multi-source data according to an embodiment of the present invention. Detailed Implementation

[0032] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0033] As mentioned in the background section, existing technologies suffer from three main problems: First, they fail to adequately utilize unstructured data such as user reviews, after-sales issues, and consultation data from e-commerce platforms, making it impossible to deeply mine usage scenarios and functional needs in user feedback, thus failing to fully release the value of the data. Second, the hierarchical and scenario-based design of promotional content lacks scientific rigor, failing to effectively highlight key functional characteristics and their application value in typical scenarios, thereby weakening the accuracy of promotion and user experience. Third, they fail to comprehensively consider scenario coverage, the degree of user pain points, and functional synergy, resulting in an unreasonable prioritization of promotional content and difficulty in meeting diverse user needs.

[0034] Our invention addresses these technical pain points by proposing a product promotion method and system based on multi-source data.

[0035] Figure 1 This is a system block diagram of a product promotion method based on multi-source data according to an embodiment of the present invention. Figure 1 As shown, the product promotion method based on multi-source data includes:

[0036] S1: Obtain review text data, after-sales problem data, and user consultation data of electronic products from e-commerce platforms, and based on the review text data, after-sales problem data, and user consultation data, use natural language processing technology to extract the product usage scenarios and corresponding functional requirements of user feedback;

[0037] Acquire text data of product reviews, after-sales issues, and user inquiries from e-commerce platforms;

[0038] Comment text data includes product reviews, follow-up reviews, and text information in accompanying images posted by users on e-commerce platforms;

[0039] The evaluation content includes text descriptions corresponding to the star rating, descriptions of the product's advantages and disadvantages, and descriptions of the user experience. Follow-up reviews include users' continuous evaluations of the product at different time periods after purchase. The text information in the accompanying images is extracted using optical character recognition technology.

[0040] After-sales issue data includes descriptions of product malfunctions submitted by users, abnormal usage situations, and reasons for return and exchange requests;

[0041] The product malfunction description includes the specific manifestation of the malfunction, the description of the usage environment, and the degree of impact of the malfunction. Abnormal usage situations include unexpected situations that occur during product use and user feedback on operating habits. Reasons for return and exchange requests include the specific manifestations of product functions not meeting expectations and a description of user expectations.

[0042] User consultation data includes pre-sales consultation records initiated by users on product pages and problem descriptions in customer service tickets;

[0043] Among them, pre-sales consultation records include user questions about product functions, descriptions of expected usage scenarios, and comparisons with other products; customer service work orders include user feedback on product usage issues, operational inquiries, and function suggestions.

[0044] Text preprocessing is performed on comment text data, after-sales problem data, and user consultation data, including removing special characters, standardizing punctuation, word segmentation, and removing stop words. At the same time, non-standard terms are standardized and synonyms are normalized. Named entity recognition technology is used to identify scene elements such as time, location, and environment related to the usage scenario from the preprocessed text.

[0045] The time element includes usage time period, seasonal characteristics and usage frequency; the location element includes indoor and outdoor environment, specific usage space and geographical features; the environmental element includes temperature and humidity conditions, lighting conditions and surrounding equipment environment. The semantic similarity of the identified scene elements is calculated through word vector model, and the hierarchical clustering algorithm is used to aggregate scene elements with similarity higher than a preset threshold to obtain a set of product usage scenarios.

[0046] Specifically, when calculating the semantic similarity between scene elements, the word vector representation of the scene elements is first constructed based on a large-scale pre-trained language model. When the scene element text is input, word segmentation technology is used to split the scene element into basic word units. For each word unit, the corresponding word vector representation is obtained through the word embedding layer of the pre-trained model. When the scene element contains multiple word units, the word vectors of these word units are weighted and averaged. The weight value is determined according to the importance of the word unit in the corpus, thus obtaining the vector representation of the scene element. If the scene element contains time information, such as "morning" or "late at night", the word vectors of these time word units are given higher weights. When it contains location information, such as "indoor" or "outdoor", the word vectors of these location word units are also given higher weights. For general modifiers, lower weights are given.

[0047] After obtaining the vector representations of scene elements, for any two scene elements, the cosine similarity between their vector representations is calculated. When the cosine similarity value of two scene elements is greater than a preset threshold of 0.8, the two scene elements are considered to have high semantic similarity. If a scene element has high semantic similarity with multiple other scene elements, these scene elements are classified into the same category. When cross-category semantic similarity occurs, the scene element is classified into the category with the highest average similarity by calculating the intra-category average similarity. For time-related scene elements, such as "used in the morning" and "used in the early morning", when their semantic similarity is high, they are preferentially classified into the same category. For location-related scene elements, such as "used in indoor office" and "used in the office", when they express similar meanings, they are also preferentially classified into the same category.

[0048] After completing the initial clustering, further semantic similarity analysis is performed on the scene elements within each category. When scene elements with significant semantic differences are found within a category, they are separated from the current category to form a new category or assigned to another more suitable category. Finally, based on the semantic similarity calculation results, a hierarchical scene element clustering structure is obtained, in which the scene elements in each category have high semantic relevance.

[0049] The topic modeling technique based on the attention mechanism is used to extract topics from the preprocessed text, identify the product functional attributes that users pay attention to in different scenarios, and calculate the importance weight of each functional attribute word based on the word frequency-inverse document frequency algorithm. The functional attribute words with the top 30% weight values ​​are selected as functional requirements. At the same time, the distribution consistency of functional attribute words in different data sources is considered to ensure that the selected functional requirements are universal.

[0050] S2: Based on the product usage scenarios and functional requirements, establish a scenario-function association matrix, and calculate the scenario coverage rate and pain point severity value of each functional requirement based on the feature distribution of the scenario-function association matrix.

[0051] By establishing a correspondence between scenario elements and functional requirements in the product usage scenario set, a scenario-function correspondence matrix is ​​constructed. The weight values ​​in the matrix are calculated by the co-occurrence frequency of scenario elements and functional requirements in the same text segment, thereby forming the product usage scenarios and corresponding functional requirements based on user feedback.

[0052] It is important to note that when establishing the correspondence between scene elements and functional requirements, a two-dimensional matrix structure is first constructed. The rows of the matrix represent the set of scene elements, and the columns represent the set of functional requirements. Deep text analysis technology is used to extract the correlation information between scene elements and functional requirements from user feedback. Semantic structure analysis is performed on each piece of user feedback text to identify the logical relationship between the scene and the function expressed in the text. Combined with contextual information, the degree of semantic correlation between scene elements and functional requirements is evaluated.

[0053] When calculating co-occurrence frequencies, the emotional characteristics of user feedback are fully considered. Sentiment analysis techniques are used to identify users' evaluation tendencies towards specific scenario-function combinations, and emotional intensity is incorporated as a moderating factor into the calculation of association strength. For scenario-function combinations with significantly positive evaluations, their association strength will be correspondingly increased; while negative evaluations will decrease the association strength.

[0054] For each scene element, analyze its association distribution characteristics with different functional requirements. Based on the statistical characteristics of association strength, identify the core functions with statistical significance in that scene. When a scene element has a strong association with multiple functional requirements, comprehensively consider the numerical characteristics of association strength and the distribution of user evaluations to establish a ranking criterion for the importance of functional requirements.

[0055] The basic scenario-function correspondence matrix is ​​represented as follows:

[0056]

[0057] in, Indicates the basic association value, Representing a scene With function Direct co-occurrence frequency, The maximum value of all co-occurrence frequencies. This indicates the number of times the scenario-feature is mentioned in user feedback.

[0058] The scenario coverage rate for each functional requirement point is calculated based on the scenario-function correspondence matrix and expressed as follows:

[0059]

[0060] in, Indication Function Scene coverage Indicates the total number of scenes. Representing a scene The weight reflects the importance of the scenario.

[0061] The higher the weight of important scenarios, the better. The more significant the contribution.

[0062] Less important scenarios contribute less to the denominator, reducing the dilution effect and preventing low-weight scenarios from dragging down the overall scenario coverage.

[0063] Furthermore, when calculating the pain point severity value, if users frequently report deficiencies or problems with a certain function in a certain scenario, it indicates that the function has obvious user pain points in that scenario; if user feedback contains strong negative emotional expressions, it indicates that the function problem has a significant impact on user experience. Therefore, when calculating the pain point severity value, it is necessary to comprehensively consider the emotional tendency of user feedback, the frequency of problems, and the degree of impact.

[0064] The formula for calculating the pain point severity value is:

[0065]

[0066] in, This represents the pain level value of function j. Representing a scene Next function The intensity of negative emotions, Indicates the degree of impact of the problem.

[0067] Furthermore, the intensity of negative emotions Represented as:

[0068]

[0069] in, Indicates the number of negative reviews. Indicates the total number of evaluations. This represents the total number of all scenarios. The index variable representing the scene, from 1 to... .

[0070] Impact of the problem Represented as:

[0071]

[0072] in, Indication Function In the scene The average task completion time in the data. The index term is the baseline completion time. Ensure that the degree of impact is positively correlated with the strength of the association.

[0073] After calculating the scene coverage and pain point severity values, a two-dimensional feature distribution of functional requirement points is obtained.

[0074] S3: Obtain the functional parameter information of the target electronic product, map and match the functional parameter information with the functional requirements, sort the functional characteristics of the target electronic product according to the scene coverage and the pain point degree value, and generate a sequence of core selling points of the product.

[0075] Comprehensive collection and systematic organization of functional parameter information for the target electronic product are necessary. This includes hardware specifications (such as processor performance, memory capacity, and display parameters), software features (such as operating system features and application functions), and unique innovative functions. During information collection, the completeness and accuracy of the functional parameters must be ensured. If necessary, cross-verification can be performed through multiple channels such as product manuals, official technical documents, and actual testing. Functional parameter information for similar competing products should also be collected accordingly.

[0076] After obtaining complete functional parameter information, it is necessary to establish a mapping relationship between these parameters and the previously identified functional requirements. The mapping process needs to consider the following aspects:

[0077] Direct mapping relationship, that is, the functional parameter directly corresponds to a certain functional requirement point;

[0078] Combination mapping relationship, that is, multiple functional parameters jointly support a certain functional requirement;

[0079] Indirect mapping relationship, that is, functional parameters indirectly satisfy functional requirements in some way.

[0080] First, we examine the correspondence between each specific functional parameter and the user's requirement point, and calculate the basic mapping strength, expressed as:

[0081]

[0082] in, Indicates the first The first functional feature is related to the first The basic mapping strength of individual user needs Representing the The first functional feature The actual values ​​of each functional parameter This represents the maximum value among parameters of the same type. Indicates the first The first functional parameter and the first The correlation coefficient between individual user needs This is the attenuation coefficient.

[0083] After determining the basic mapping strength, a market-level comparison needs to be introduced. Therefore, a functional satisfaction index is constructed, calculated using the following formula:

[0084]

[0085] in, Indicates product Medium function Satisfaction index This represents the average value of this functional parameter for similar products on the market. The standard deviation of similar products in the market reflects the degree of market fragmentation. This is a weighting coefficient used to adjust the influence of market leadership on satisfaction scores.

[0086] Based on the functional satisfaction index, combined with the aforementioned scenario coverage and pain point severity values, a comprehensive importance index is calculated. This establishes an organic link between the three dimensions of technical parameters, market performance, and user needs. The formula for the comprehensive importance index is as follows:

[0087]

[0088] in, This is the balance coefficient, and its value range is [0 ≤ 1]. ≤2] is used to adjust the relative importance of scene coverage and pain point severity values.

[0089] when When the threshold is 0, the pain point is completely ignored, and the overall importance is determined solely by the scenario coverage rate. and market satisfaction This decision applies to situations where users are less sensitive to pain points (such as in mature markets where users focus on the overall product rather than specific pain points).

[0090] when When the value is 2, the weight of pain point on overall importance increases significantly, which is suitable for situations where pain points are prominent (such as emerging markets where pain points dominate user feedback).

[0091] when When the value is 1, the weights of pain point and scenario coverage are balanced, which is suitable for ordinary scenarios. Users care about both the comprehensiveness of the product and their specific pain points.

[0092] at the same time, The value is adjusted based on market feedback; if the market feedback shows a strong negative sentiment... High, the degree of impact of the problem Serious, high settings The value range is 1.5≤ ≤2.

[0093] If market feedback indicates that users are not very sensitive to pain points, set a lower threshold. The value range is 0≤ ≤0.5.

[0094] Furthermore, to comprehensively assess the interaction between functional characteristics, a functional synergy index is introduced, expressed by the formula:

[0095]

[0096] in, Indication Function With function The synergy index between them ranges from [0,1]. and They represent the products. Medium function and functions Satisfaction index and They represent Scene coverage This indicates the total number of reference products in the set.

[0097] Based on the synergy index, the group influence of each function is calculated and expressed as:

[0098]

[0099] in, Indicate target function Group influence This indicates the total number of all product features. Indicates the first A comprehensive importance score for each function.

[0100] It is important to note that when analyzing the formula for calculating group influence, three characteristics need to be considered: reflexivity, symmetry, and accumulation.

[0101] When encountering during the calculation process = In this case, it is actually examining the relationship between the function and itself, reflecting the independent integrity and internal consistency of the function, and providing a benchmark reference point for the overall evaluation;

[0102] Synergy Index Symmetry features (i.e.) = This reflects the reciprocal nature of interactions between functions, indicating that if functions... Function If there is a facilitating or disruptive effect, then the counter-effect will also exist to the same extent. However, it is worth noting that although the synergy index is symmetrical, the final group influence... and They may not be equal because different importance indices must be considered during the calculation. And the unique positioning and function of each feature within the overall system;

[0103] The summation operation ensures that the interaction between the target function and all other functions in the system (including itself) can be fully considered, including both direct functional synergy and the formation of a complete functional interaction network.

[0104] Group influence comprehensively reflects the synergistic effect and importance of a function with other functions. When a function has a strong synergistic relationship with multiple important functions, its group influence will be high, indicating that the function occupies a more important position in the overall product function system. Therefore, considering the development trend of this product function, a trend gain factor is introduced to give functions with strong group influence a higher weight in future development. The trend gain factor formula is expressed as:

[0105]

[0106] in, This indicates the expected growth rate of the feature in the market. For the prediction period. Index term. This ensures that the moderating effect of group influence on trend gain is achieved. This is the adjustment coefficient.

[0107] Ultimately, the importance score of the electronic product's core selling points comprehensively considered three dimensions: the fundamental importance of the function, development trends, and group influence. The calculation formula is expressed as:

[0108]

[0109] in, Indicates the first The first product The overall importance score of each feature as a core selling point.

[0110] Based on the calculated score values, a functional importance sequence is constructed. The score of each functional unit is normalized by dividing it by the total score, yielding a standardized weight coefficient for each function. Then, according to the magnitude of these standardized weight coefficients, the functional sequence is divided into three levels: the top 2-3 functional units with the highest weight coefficients constitute the primary core selling point level; the next 3-4 functional units with relatively high weights constitute the secondary selling point level; and the remaining functional units constitute the extended selling point level. This establishes a hierarchical product core selling point sequence.

[0111] S4: Based on the product's core selling points sequence, promotional content is generated using a progressive information display method, prioritizing the display of functional features with higher pain point values, and incorporating corresponding usage scenario examples into the displayed content to form a scenario-based product promotion plan.

[0112] The pain point score of each functional feature Its importance score Multiply to obtain the overall priority index for information display. .because The composite calculation has taken into account the overall importance index of the function, the trend gain factor, and the normalized group influence, so it can simultaneously reflect the market value of the function and the degree of user pain points.

[0113] For the calculated The values ​​are normalized to obtain standardized display priority coefficients, and the display order of functional features is sorted accordingly. In the main core selling points level, [the following is selected]. The features with the highest perceived value are prioritized for dissemination. For each feature, usage scenario feature vectors are extracted based on user behavior data analysis and field research. ,in, Quantitative indicators representing the characteristics of a scenario (such as frequency of use, scope of impact, and efficiency of resolution). =1,2,...n. By calculating the Euclidean distance between the scene feature vectors, 2-3 of the most representative typical scenes are selected.

[0114] For the selected typical scenarios, construct a three-part promotional content: The first part, a description of the scenario's pain points, should include specific pain point metrics (e.g., functional features in this scenario). The core problem to be solved occurs most frequently. Once per month, affecting the number of users. (This causes an average delay of T hours).

[0115] The second section demonstrating functional value needs to be combined with... The various indicators in the calculations emphasize the overall importance of this function. and future development potential ;

[0116] The third section, presenting the user experience, should reflect the influence of the group. The actual application effect of the function is demonstrated through specific user feedback data.

[0117] Based on the hierarchical position of functional characteristics within the core selling points sequence, set differentiated information release rhythms: main core selling points hierarchy ( The top 2-3 features will be displayed every 3-5 days, the next 3-4 features will be displayed every 5-7 days, and the extended features will be displayed every 7-10 days.

[0118] In summary, a product promotion method and system based on multi-source data, as described in this invention, is presented. By deeply mining user reviews, after-sales issues, and consultation data, and combining this with natural language processing technology, it accurately extracts usage scenarios and functional requirements from user feedback, significantly improving the depth and accuracy of data processing. Based on an evaluation method using scenario coverage and pain point severity values, the priority of functional requirements is scientifically determined, overcoming the shortcomings of existing technologies in terms of unclear promotional content hierarchy and insufficient targeting. Furthermore, by generating a hierarchical sequence of core selling points and combining it with a scenario-based display strategy, this invention not only improves the matching degree between promotional content and user needs but also significantly enhances user experience and promotional effectiveness.

Claims

1. A product promotion method based on multi-source data, characterized in that, include: Acquire review data, after-sales issues, and user inquiries for electronic products on e-commerce platforms, and extract product usage scenarios and functional requirements using natural language processing technology; Based on the product usage scenarios and functional requirements, a scenario-function association matrix is ​​established, and based on the feature distribution of the scenario-function association matrix, the scenario coverage rate and pain point intensity value of each functional requirement are calculated. Obtain the functional parameter information of the target electronic product, map and match the functional parameter information with the functional requirements, sort the functional characteristics of the target electronic product by importance according to the scenario coverage and the pain point degree value, and generate a sequence of core selling points of the product. Based on the product's core selling points, promotional content is generated using a progressive display approach, prioritizing the display of features that address high pain points, and combining these with usage scenario examples to form a scenario-based promotional plan.

2. The product promotion method based on multi-source data according to claim 1, characterized in that, The scenario-function association matrix establishes a correspondence between scenario elements and functional requirements. The matrix weights are calculated based on co-occurrence frequency and adjusted for association strength by combining sentiment analysis.

3. The product promotion method based on multi-source data according to claim 2, characterized in that, The pain point severity value is calculated based on the scenario-function correlation matrix and expressed as follows: in, Indication Function The degree of pain point, Representing a scene Next function The intensity of negative emotions, Indicates the degree of impact of the problem; The intensity of negative emotion is calculated as the ratio of the number of negative evaluations to the total number of evaluations, combined with the function. In the scene The associated value Weighted adjustment relative to the total correlation value across all scenarios; The impact of the problem is comprehensively considered. In the scene The ratio of task completion time to baseline completion time, and related values. The positive correlation with the degree of influence was obtained through calculation.

4. The product promotion method based on multi-source data according to claim 3, characterized in that, Based on the functional parameter information of the target electronic product and the functional requirements, a mapping and matching process is performed to examine the correspondence between each specific functional parameter and the user's requirements, and the basic mapping strength is calculated.

5. The product promotion method based on multi-source data according to claim 4, characterized in that, Based on the aforementioned basic mapping strength, and taking into account the comparison between the product's functional parameters and the average value of similar products in the market, a functional satisfaction index is constructed. On this basis, combined with scenario coverage and pain point severity values, a comprehensive importance index is calculated.

6. The product promotion method based on multi-source data according to claim 5, characterized in that, Based on the aforementioned functional satisfaction index, a functional synergy index is introduced to measure the synergistic effect between functions, and the group influence of functions is calculated based on the synergy index.

7. The product promotion method based on multi-source data according to claim 6, characterized in that, The functional synergy index is expressed by the formula: in, Indication Function With function The synergy index between them ranges from [0,1]. and They represent the products. Medium function and functions Satisfaction index and They represent Scene coverage This indicates the total number of reference products in the set.

8. The product promotion method based on multi-source data according to claim 7, characterized in that, Based on the importance of the group influence assessment function in the product system, and by introducing a trend gain factor, combined with the expected market growth rate and forecast period of the function, functions with high group influence are given higher future weight. By taking into account the fundamental importance, development trends, and group influence, the overall importance score of the function as a core selling point is calculated, and standardized weight coefficients are generated to construct a sequence of core selling points.

9. The product promotion method based on multi-source data according to claim 8, characterized in that, Based on the core selling point sequence, the highest priority function is selected as the primary focus of promotion among the main core selling points. The usage scenario feature vector of each function is extracted, and the most representative typical scenarios are selected by calculating Euclidean distance for product promotion.

10. A product promotion system based on multi-source data, using the product promotion method based on multi-source data according to any one of claims 1-9 to promote electronic products, characterized in that, include: The data collection module acquires review data, after-sales issues, and user inquiries for electronic products from e-commerce platforms, and extracts product usage scenarios and functional requirements through natural language processing technology. The association module establishes a scenario-function association matrix based on the product usage scenarios and functional requirements, and calculates the scenario coverage rate and pain point severity value of each functional requirement based on the feature distribution of the scenario-function association matrix. The matching module acquires the functional parameter information of the target electronic product, maps and matches the functional parameter information with the functional requirements, sorts the functional characteristics of the target electronic product by importance according to the scenario coverage and the pain point degree value, and generates a sequence of core selling points of the product. The promotion module generates promotional content based on the product's core selling points sequence, using a progressive display method. It prioritizes showcasing features that address high pain points and combines them with usage scenario examples to form a scenario-based promotional plan.

Images