FsQCA-based agricultural product electronic word-of-mouth configuration path identification method

Abstract

The invention discloses an FsQCA-based agricultural product electronic word-of-mouth configuration path identification method, which comprises the following steps: collecting data related to agricultural product electronic word-of-mouth, and extracting related factors according to the obtained data; quantizing the set variable data, and calibrating the membership degree of the condition variable and the result variable by using a fuzzy set membership degree method; based on an FsQCA method, analyzing a causal relationship between a condition variable and a result variable; and constructing a causal path and configuration configurations, obtaining a plurality of configuration configuration selection paths of the agricultural product electronic word-of-mouth influence factors according to a causal path analysis result, and screening and identifying effective configuration paths. The method can systematically identify, classify and explain the diversified influence of different factor combinations on the successful propagation of the electronic praise of the agricultural product and improve the scientificity of condition configuration adaptation in combination with the characteristics of the agricultural product, and can provide decision support for the brand party of the electronic praise of the agricultural product.

Description

Technical Field

[0001] The invention belongs to the technical field of data analysis and agricultural product e-commerce, and specifically relates to a method for identifying a configuration path of electronic word-of-mouth of agricultural products based on FsQCA. Background Art

[0002] With the rapid development of Internet technology and the widespread popularity of e-commerce platforms, online sales of agricultural products have gradually become an important engine for agricultural economic growth and a key link in the transformation of modern agriculture. In this context, e-word-of-mouth, as an important source of consumer information acquisition and decision-making support, not only significantly affects consumers' brand awareness and trust in agricultural products, but also plays an irreplaceable role in shaping brand image, driving purchase decisions and optimizing marketing strategies.

[0003] However, the current methods for analyzing the influencing factors of agricultural product electronic word-of-mouth communication are mainly based on traditional statistical methods, which have the following two limitations:

[0004] First, the communication effect of agricultural product electronic word-of-mouth itself is affected by many complex factors, and agricultural product electronic word-of-mouth has its heterogeneity. Existing methods often select some factors for analysis and lack a systematic and scientific factor screening mechanism;

[0005] Second, the effect of electronic word-of-mouth communication of agricultural products is not determined by a single factor, but is a comprehensive effect produced by the interaction and combination of multiple factors. The existing methods do not pay attention to the combination path and adaptability between influencing factors, and cannot effectively reveal the diverse mechanisms for the successful electronic word-of-mouth communication of agricultural products.

[0006] Based on the above limitations, there is an urgent need for a scientific method that can comprehensively consider the interaction of multiple factors and reveal the causal path. Compared with the traditional quantitative linear relationship research, the research based on the FsQCA (fuzzy set qualitative comparative analysis) method has significant innovation and advantages. Revealing nonlinearity and adaptability: The traditional linear regression method assumes that there is a fixed linear relationship between factors. However, the communication effect of agricultural product electronic word-of-mouth is often a nonlinear result of the interaction of multiple factors. The FsQCA method can not only identify the combination patterns of different factors, but also reveal how these combination patterns match the adaptability of the result variable, thereby more truly reflecting the complexity of electronic word-of-mouth communication. At the same time, when faced with complex causal relationships, traditional statistical methods tend to ignore the interactions between certain factors, while the FsQCA method can flexibly handle different types of factors (such as categorical, ordinal, and interval variables), and can still ensure efficient and accurate analysis in situations where the sample size is small and the relationship between variables is complex.

[0007] Therefore, as an innovative method integrating the advantages of qualitative and quantitative analysis, the FsQCA method can effectively identify the combination pattern of influencing factors and their adaptation paths with the result variables, and has unique advantages in analyzing the multidimensional influencing factors and their interaction mechanisms of agricultural product electronic reputation. Based on the FsQCA method, the present invention systematically identifies the key influencing factors and configurations that affect agricultural product electronic reputation, reveals its multivariate mechanism, and provides theoretical guidance and practical support for the optimization of agricultural product e-commerce strategies. Summary of the invention

[0008] This paper proposes a configuration path identification method for agricultural product electronic word-of-mouth based on FsQCA, aiming to systematically explore the multiple influencing factors and their complex interactive relationships of the communication effect of agricultural product electronic word-of-mouth. This method constructs a scientific analysis framework and combines the fuzzy set membership method to quantitatively analyze the configuration path between influencing factors and communication effects, providing decision support for agricultural product e-commerce platforms.

[0009] A method for identifying the configuration path of electronic reputation of agricultural products based on FsQCA, the method comprising the following steps:

[0010] S1: Collect data related to e-WOM of agricultural products, including product presentation content, consumer interaction and social media data, supplementary data, etc.;

[0011] S2: Extract relevant factors based on the obtained data, build a framework model of influencing factors of agricultural product e-reputation, and set conditional variables and outcome variables;

[0012] S3: Quantify the variable data and use the fuzzy set membership method to calibrate the membership of the conditional variable and the outcome variable;

[0013] S4: Based on the FsQCA method, the causal relationship between the conditional variables and the outcome variables was analyzed;

[0014] S5: Construct causal paths and configurations. According to the results of causal path analysis, obtain multiple configuration selection paths for factors affecting the electronic reputation of agricultural products, and screen and identify effective configuration paths.

[0015] Furthermore, the specific contents of S1 include:

[0016] The product presentation content is subdivided into basic content information (such as origin, brand, quality certification, sales nature, etc.), content theme (such as production process, cultural introduction, health function, etc.), content standard, content emotion, content publisher influence and publisher professionalism;

[0017] The consumer interactions are subdivided into like interactions, comment interactions, sharing interactions and user-generated content;

[0018] The social media data is subdivided into communication index and marketing data (such as sales volume, sales amount, conversion rate, etc.); the supplementary data may also include data on external factors such as product advertising, promotional activities, and logistics service quality as needed, so as to further analyze the impact of these factors on electronic word-of-mouth communication.

[0019] Furthermore, the specific contents of S2 include:

[0020] Data collation: Clean and screen the collected electronic word-of-mouth data of agricultural products to remove invalid or duplicate data;

[0021] Factor extraction: extracting key factors that may affect e-WOM communication from the data, such as content quality, consumer interaction, product characteristics, etc.;

[0022] Variable setting: Determine the conditional variables (influencing factors) and outcome variables (electronic word-of-mouth communication effects).

[0023] Conditional variables may include content quality, consumer behavior, communication channels, etc. Specifically, they include content standards (the number of professional terms used in agricultural product introduction videos) and content sentiment (the positive and negative sentiment and intensity are obtained through sentiment analysis tools); like interaction (number of likes / number of views), comment interaction (number of comments / number of views) and sharing interaction (number of shares / number of views); publisher influence (logarithm of the number of fans) and publisher professionalism (whether there is an official certification mark, with a value of 0 or 1);

[0024] The result variable is the effect of electronic word-of-mouth communication. The result variable is measured by the communication index (communication index data obtained by the Massive Arithmetic Platform) and sales revenue. The normalized weight is calculated by the entropy method to ensure objective weight distribution.

[0025] S1: The diffusion index is x1, the sales revenue is x2, and the normalization formula (assuming there are m samples):

[0026]

[0027] Among them, x ij is the original value of the jth indicator of the ith sample, min(x j ) and max(x j ) is the minimum and maximum value of index j, x′ ij It is the normalized value, which represents the standardized value of the jth indicator of the i-th sample.

[0028] S2: The entropy method determines the weight of each indicator based on its distribution information. The steps are as follows: Calculate the normalized value ratio p of each indicator ij :

[0029]

[0030] Among them, x′ ij is the normalized value of the jth indicator of the i-th sample. Calculate the entropy value e j , the entropy value e of each indicator j It indicates the uncertainty of its information. The smaller the entropy value, the more important the indicator is.

[0031]

[0032] Among them, p ij is the ratio of the jth indicator of the ith sample, and ln(m) is the natural logarithm of the number of samples. Calculate the weight w j , the larger the weight, the more important the indicator is.

[0033]

[0034] Among them, e j is the entropy value of the jth indicator, and n is the number of indicators (here 2, namely, the communication index and sales revenue). Calculate the comprehensive result variable, and the final communication effect evaluation Y is obtained by weighted average of the normalized values ​​of the communication index and sales revenue:

[0035] Y=w1·x'1+w2·x'2

[0036] Among them, x′1 is the normalized value of the communication index, x′2 is the normalized value of sales revenue, w1 and w2 are the weights of the communication index and sales revenue, respectively calculated by the entropy method.

[0037] Finally, all the above variables are normalized.

[0038] Furthermore, the specific contents of S3 include:

[0039] The influencing factors of agricultural product electronic word-of-mouth are taken as conditional variables, and the communication effect of agricultural product electronic word-of-mouth is taken as the result variable. The calibration anchor points of the conditional variables are calculated by direct calibration method. For fixed-distance and fixed-ratio variables, three calibration anchor points are set by quantile method:

[0040] 95% quantile: as the anchor point of complete membership (value 1), indicating that the variable fully meets the set characteristics; 50% quantile: as the anchor point of the intersection (value 0.5), indicating that the variable partially meets the set characteristics; 5% quantile as the anchor point of complete non-membership (value 0), indicating that the variable does not meet the set characteristics at all. Based on the above calibration anchor points, the conditional variable is calibrated to the interval [0,1] to obtain the membership score of each sample. The calibrated membership score is used to indicate the degree to which the conditional variable belongs to a specific set (such as the "high content standard" set).

[0041] Furthermore, the specific contents of S4 include:

[0042] Based on the calibration anchor point results of the variables, the consistency and coverage between the single conditional variables and the outcome variables are calculated, and the steps of determining the necessary conditions for the spread of electronic word-of-mouth of agricultural products include:

[0043] The consistency between the conditional variable and the outcome variable is calculated by the following formula:

[0044]

[0045] Consistency is used to measure the degree of causal matching between the condition variable X and the result variable Y. It reflects the strength of causal matching by calculating the proportion of the intersection of the condition variable membership and the result variable membership in the condition variable membership. It means the proportion of the result variable Y that is satisfied when the condition variable X is satisfied. The consistency value range is [0,1]. The closer the result is to 1, the stronger the causal matching between X and Y.

[0046] X i represents the membership value of the conditional variable in the i-th sample, Y i Indicates the membership value of the outcome variable in the i-th sample, min(X i ,Y i ) indicates that the conditional variable and the outcome variable membership take the smaller value of the two, ∑min(X i ,Y i ) is the sum of X and Y in all samples, ∑X i is the sum of the memberships of the conditional variables, and the ratio indicates whether the membership of the conditional variables is sufficient to cover the outcome variable.

[0047] The coverage between the condition variable and the outcome variable is calculated by the following formula:

[0048]

[0049] The coverage response indicates how much of the realization of the outcome variable Y can be explained by the condition variable X. By calculating the proportion of the intersection of the condition variable and the outcome variable to the total amount of the outcome variable, the coverage value range is [0,1]. The closer the result is to 1, the greater the range of Y that X can explain. ∑Y i It is the sum of the membership degrees of the outcome variable, and the ratio reflects the proportion of the outcome variable Y that can be explained by the conditional variable X.

[0050] Consistency focuses on whether the conditional variable reliably predicts the outcome variable, while coverage focuses on the degree to which the conditional variable explains the outcome variable. In causal analysis, high consistency indicates a strong causal relationship, but if coverage is low, it means that the influence of the condition is limited. Conditional variables with consistency values ​​greater than the consistency threshold are regarded as necessary conditions for the outcome variable.

[0051] Furthermore, the specific contents of S5 include:

[0052] The original consistency threshold is set, the consistency and coverage between the condition configuration and the result variable are calculated, and the simple solution, intermediate solution and complex solution are obtained through configuration sufficiency analysis. The core conditions and edge conditions are determined according to the simple solution, intermediate solution and complex solution, and the configuration configuration of successful electronic word-of-mouth communication of agricultural products under the preset original consistency threshold is obtained. Multiple guided paths are obtained by screening and identifying and classifying various configuration features.

[0053] Furthermore, it also includes: Robustness test: Raise the consistency threshold (e.g., from 0.8 to 0.85) to verify the stability of the configuration results. If the experiment shows that the path results are consistent under different consistency settings, the robustness of the analysis is verified.

[0054] The present invention can combine the characteristics of agricultural products, systematically identify, classify and explain the diverse impacts of different combinations of factors on the success of agricultural product electronic word-of-mouth communication and improve the scientific nature of condition configuration adaptation, and can provide decision support for agricultural product electronic word-of-mouth brands. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0056] Figure 1 The present invention shows a flow chart of a method for identifying a configuration path of an electronic reputation of agricultural products based on FsQCA. DETAILED DESCRIPTION

[0057] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further described below in conjunction with the accompanying drawings and preferred embodiments.

[0058] In order to systematically explore and analyze the adaptation path of the influencing factors and configuration of the electronic word-of-mouth of agricultural products, and systematically explore and analyze the multiple influencing factors and their complex interactions in the successful dissemination of the electronic word-of-mouth of agricultural products, the present invention proposes a configuration path identification method for the electronic word-of-mouth of agricultural products based on FsQCA. This method is based on the construction of a framework model to reasonably and systematically screen the influencing factors of the electronic word-of-mouth of agricultural products, and continue to adapt to the configuration path between the factors, aiming to provide decision support for the successful dissemination of the electronic word-of-mouth of agricultural products.

[0059] Figure 1The figure shows a flow chart of a method for identifying a configuration path of an electronic reputation of agricultural products based on FsQCA according to an embodiment of the present invention.

[0060] like Figure 1 As shown, a method for identifying the configuration path of the electronic word-of-mouth of agricultural products based on FsQCA is provided. The method starts with step S1, and obtains the relevant data of the influencing factors according to the influencing factor framework of the electronic word-of-mouth of agricultural products.

[0061] The data related to the electronic word-of-mouth of agricultural products in Pu'er, Yunnan Province on the Douyin platform were selected to study the influencing factors and configuration of the electronic word-of-mouth of agricultural products. The sample data of the head videos were screened out through the platform, and the sample data with incomplete information was deleted, and finally the index of the influencing factors of the electronic word-of-mouth of agricultural products with a certain number of samples was obtained.

[0062] The main data collected include content themes, content IP, video content, publisher characteristics, interaction volume, etc. The effect of online agricultural product electronic word-of-mouth communication is taken as the result variable. Since the impact of electronic word-of-mouth on the sales volume of corporate products is the focus of research in the field of electronic word-of-mouth communication, the dependent variables used in research are generally indicators that can reflect product sales (such as sales revenue). This patent takes into account that the effectiveness of a single indicator is relatively low, so it comprehensively evaluates word-of-mouth communication and marketing effects to fully reflect the breadth of electronic word-of-mouth communication of agricultural products on social platforms and its effect of converting into actual purchasing behavior. Word-of-mouth communication is measured by the communication index provided by the huge arithmetic platform, and the marketing effect is measured by the sales revenue of the product. After the two are normalized, the entropy method is used to confirm the rights, and the above data are normalized.

[0063] S2, construct a framework model of factors influencing the electronic word-of-mouth of agricultural products.

[0064] Based on the Media Richness Theory, Information Processing Theory and Opinion Leader Theory, the framework of factors influencing the electronic word-of-mouth of agricultural products includes information conditions, interaction conditions and opinion conditions; the information conditions include the quality of information content and the level of appeal, which are subdivided into content standards and content emotions; the interaction conditions include audience interaction, which are subdivided into like interaction, comment interaction and sharing interaction; the opinion conditions include publisher characteristics, which are subdivided into publisher influence and publisher professionalism. The antecedent variable design covers three main dimensions: content quality, audience interaction and publisher characteristics, as shown in Table 1.

[0065] Table 1 Antecedent variables

[0066]

[0067]

[0068] S3, after quantifying the data collected on the factors affecting the e-WOM of agricultural products, variable calibration was performed, and the membership of the conditional variables and the outcome variables was calibrated using the fuzzy set membership method.

[0069] In content dissemination, information richness and emotional expression are key factors. Based on the Media Richness Theory, it is believed that the detail and professionalism of information help to improve the understanding and attractiveness of information. In the introduction video of agricultural products, the quality of content is mainly measured by the use of professional terms, such as whether it contains scientific names or industry standards related to agriculture. This professional information helps to establish the authority and credibility of the content. Content sentiment is measured based on the intensity of emotional expression (positive or negative) of the content. The video text content is analyzed to generate a sentiment score. The range of sentiment scores is -1 to 1, indicating the intensity of sentiment from negative to positive.

[0070] In the process of information dissemination, the audience's interactive behavior reflects the attractiveness and popularity of the content. Based on Information Processing Theory, the audience's interactive behaviors, such as likes, comments, and shares, can further amplify the dissemination effect of information. Like interactions, comment interactions, and share interactions are measured by the ratio of the number of likes, comments, and shares to the number of views, respectively, indicating the importance of interactive behavior in content dissemination. The ratio of likes to views can be seen as a reflection of the audience's direct preference for the content, the number of comments to views reflects the audience's thinking and feedback on the content, and the number of shares to views indicates that the audience is willing to spread the content to more people.

[0071] The influence and professionalism of the publisher are directly related to their authority on the platform and the credibility of their content. The influence of the publisher is measured based on the number of their fans. According to the Opinion Leader Theory, publishers with a high number of fans are considered to have greater influence and can drive a wider range of user groups. The professionalism of the publisher is measured by whether it has obtained official certification from Douyin, which is represented by a binary variable (0 / 1). Officially certified publishers usually have a certain professional background or qualifications, which enhances the authority and credibility of the content. In the electronic word-of-mouth communication of agricultural products, the audience tends to trust publishers with higher professionalism, so professionalism plays an important role in the communication effect.

[0072] After the variables are measured, the fuzzy set data is calibrated. Taking content quality as an example, FsQCA only knows that the values ​​between them are different, but it is not clear to what extent the content quality of a case belongs to the high attention set. After data calibration, the purpose of clarifying the approximate position and membership of the content quality value can be achieved. Calibration is to regard the outcome variable and each conditional variable as a set, and convert each variable value set into a membership between 0 and 1. According to the characteristics of the original data, the direct calibration method is adopted for calibration. In the selection of three anchor points, the 95% quantile is used as the complete membership anchor point, the 5% quantile is used as the completely non-membership anchor point, and the 50% quantile is used as the intersection anchor point. The specific anchor point values ​​are shown in Table 2. Since 0.5 cannot determine whether it is more inclined to complete membership or more inclined to complete non-membership, after calibration, the variable value of 0.5 must be manually changed to 0.499 or 0.501. Here, it is changed to 0.501 to avoid the software automatically eliminating the case corresponding to 0.5 due to the law of fuzzy set intersection.

[0073] Table 2 Calibration anchor points of conditional variables

[0074]

[0075] S4, based on the FsQCA method, analyze the causal relationship between the conditional variable and the outcome variable

[0076] According to the causal logic of the set, if condition X always occurs when a specific result Y occurs, then X is a necessary condition for Y, and the necessity of condition X for result Y means whether Y can deduce X. In the study of FsQCA, consistency is generally used as the criterion for necessity, and its threshold is defined as 0.9. Each condition variable also corresponds to a complement (indicated by "~"), which indicates the opposite or missing state of the characteristic. Necessary Conditions in FsQCA is used to conduct a necessity analysis on the condition variables that affect the effect of agricultural product electronic word-of-mouth. The results are shown in Table 3. It can be seen from the table that the consistency of content quality is 0.637, indicating that it has a high correlation with the result; the coverage is 0.788, indicating that there is a large coverage range in the sample; the complement (~content quality) has a low correlation (0.363) and a significantly lower coverage (0.676), which further verifies that the positive impact of "content quality" is more significant. The consistency of all conditions is lower than 0.9, and any one of these conditions cannot induce the outcome variable alone. Therefore, it can be considered that the effect of agricultural product electronic word-of-mouth is affected by the combination of multiple conditional variables.

[0077] Table 3 Necessity analysis of single factors of the effect of electronic word-of-mouth of agricultural products

[0078]

[0079] Different from the necessity analysis, the sufficiency discusses whether condition X can derive a specific result Y, and the condition X here refers to the condition configuration rather than a single condition variable, so this part of the analysis is used to determine the sufficiency of different condition configurations for the generated results. When the truth table is generated, there are three thresholds that need to be set. Some scholars set 0.75 as the consistency critical value. There is no unified measurement standard for this threshold in the academic community. Combined with the sample situation, this description uses 0.8 as the consistency threshold; the frequency threshold is set according to the case size. The case belongs to a small and medium sample, so the frequency is set to 1; PRI consistency is used to further screen the truth table rows. The larger the value, the less likely it is that the same cause has different results. The setting of the PRI threshold refers to the threshold value of 0.75 proposed by scholar Du Yunzhou in his research. FsQCA can calculate complex solutions, intermediate solutions and simple solutions. In the configuration analysis, the intermediate solution among the three solutions is used as the basis, and the simple solution is used as a supplement, and the complex solution is not considered.

[0080] The configuration analysis is carried out by expanding the two solutions. The existence of the conditions is recorded as a solid circle, and the lack of conditions is recorded as a cross circle. The large solid circle represents the core condition, and the small solid circle represents the edge condition. After determining the representation method, the configuration obtained by the intermediate solution is first drawn with a small solid circle, and then combined with the simple solution. If it exists in both solutions, the solid circle is adjusted to a large size. If it is missing in both solutions, the cross circle is adjusted to a large size. The overall results are shown in Table 4. It can be seen from the configuration table that there are 7 configurations that bring high agricultural product electronic word-of-mouth effects. The consistency of the overall and each configuration meets the requirement of being greater than the threshold of 0.8, and the overall coverage is also good, proving that the 7 configurations have strong interpretability.

[0081] Table 4 Conditional configuration table of high agricultural product electronic reputation effect

[0082]

[0083] Observing each configuration vertically, the core conditions of configuration a are content quality and publisher professionalism, and the existence of publisher influence, content sentiment, and like interaction plays an auxiliary role. In configuration b, the existence of content quality and publisher professionalism is the core condition, and the existence of comment interaction content sentiment and like interaction is an auxiliary condition. In configuration c, the core condition is the existence of comment interaction, and the auxiliary conditions are content sentiment, like interaction, and publisher professionalism, while publisher influence is absent. In configuration d, the existence of comment interaction is the core condition, and the existence of content quality, content sentiment, and sharing interaction are auxiliary conditions.

[0084] S5: Construct causal paths and configurations. According to the results of causal path analysis, obtain multiple configuration selection paths for factors affecting the electronic reputation of agricultural products, and screen and identify effective configuration paths.

[0085] In the configuration analysis of the effect of electronic word-of-mouth communication of agricultural products, the four configurations can be divided into two types according to the core conditions: communication-oriented and interaction-oriented. Each type shows a different influence path and condition combination in the electronic word-of-mouth communication of agricultural products.

[0086] Communication-oriented (configuration a and configuration b): The communication-oriented configuration focuses on the authority and professionalism of the content. It takes content quality and publisher professionalism as core conditions, supplemented by publisher influence and user likes and other interactions, to form a wide spread of content. This type of configuration highlights the strategy of improving credibility and expanding brand awareness through professional content on social platforms. Communication-oriented is based on information richness theory and opinion leader theory. It has unique regional culture and industrial characteristics, so content quality is particularly important in the dissemination of such content. Users' trust in products often depends on whether the publisher has a high professional background and the authority of the information. In configuration a, the existence of publisher influence and content emotion further strengthens the credibility and attractiveness of the information. According to the dual processing theory, users are more inclined to trust authoritative content sources. In configuration b, like interaction, comment interaction and content emotion play an auxiliary role, reflecting the boosting effect of interactive behavior on the dissemination effect. Users' likes and comments not only provide social recognition externally, but also enhance the visibility of information, making the brand image more popular.

[0087] Interaction-oriented (configuration c and configuration d): The interaction-oriented configuration emphasizes the interactivity of users in the content, especially the role of comment interaction. This type of configuration creates an interactive atmosphere through user comments, likes and sharing, which enhances the brand influence on the social platform. This model pays more attention to the interactive behavior between users on the social platform, and expands the awareness and consumption willingness of the product through discussion and sharing between users. The interaction-oriented type is closely related to social interaction theory and information processing theory in theory. As a commodity with local characteristics, its unique regional attributes often stimulate users' interest and desire for discussion during the information dissemination process. In configuration c, comment interaction is the core condition, emphasizing the key role of user participation in content dissemination, while the professionalism of the publisher is an auxiliary condition to increase the credibility of the information. Comment interaction can stimulate discussion between users and enhance the attractiveness and credibility of content, which is in line with the mechanism of deep user participation in information processing theory. In configuration d, comment interaction is still the core condition, while content quality, content emotion and sharing interaction are auxiliary conditions to build a more interactive and sharing communication environment. Social capital theory further explains that commenting and sharing behaviors not only expand the scope of information dissemination, but also strengthen the social connections between users, making the content more influential and trustworthy among users. Through the two types of configuration analysis, we can see the different paths of communication-oriented and interaction-oriented in the electronic word-of-mouth communication of agricultural products, as well as their potential impact on product awareness and purchase intention. This result helps to further optimize the marketing strategy of social platforms and achieve more efficient dissemination and conversion.

[0088] Robustness testing refers to the verification of the results obtained. If the results of the empirical study remain the same after changing certain conditions, it indicates that they are stable and reliable. This paper verifies the results by improving consistency. The first verification only changes the consistency threshold to 0.85, and the second verification only changes the PRI consistency threshold to 0.8. The results of the two verifications are the same as the existing research results, proving that the empirical study is robust.

[0089] Those skilled in the art should understand that the present invention is not limited by the above-mentioned embodiments. The above-mentioned embodiments and the description only illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the technical solutions described in the above-mentioned examples can still be adjusted, or some technical details can be replaced with similar functions. These adjustments or replacements will not change the core ideas of the relevant technical solutions, nor will they exceed the basic principles and coverage of the technical solutions of the various examples of the present invention. These changes and improvements all fall within the scope of the present invention to be protected. The scope of protection claimed in the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for identifying the configuration path of agricultural product electronic reputation based on FsQCA, characterized by: The steps of this method are as follows: S1. Collect data related to e-WOM of agricultural products, including product presentation content, consumer interaction and social media data, and supplementary data; S2. Extract relevant factors based on the obtained data, build a framework model of influencing factors of agricultural product e-reputation, and set conditional variables and outcome variables; S3, quantify the variable data, and use the fuzzy set membership method to calibrate the membership of the conditional variable and the outcome variable; S4. Analyze the causal relationship between conditional variables and outcome variables based on the FsQCA method; S5. Construct causal paths and configurations. According to the results of causal path analysis, obtain multiple configuration selection paths for factors affecting the electronic reputation of agricultural products, and screen and identify effective configuration paths.

2. According to claim 1, a method for identifying paths of electronic reputation of agricultural products based on FsQCA configuration, characterized in that: The specific contents of S1 include: The product presents content, which is subdivided into basic content information, content theme, content standard, content sentiment, content publisher influence and publisher professionalism; The consumer interactions are subdivided into like interactions, comment interactions, sharing interactions and user-generated content; The social media data is subdivided into communication index and marketing data; The supplementary data may include data on external factors such as product advertising, promotional activities, and logistics service quality as needed, so as to further analyze the impact of these factors on electronic word-of-mouth communication.

3. According to claim 1, a method for identifying paths of electronic reputation of agricultural products based on FsQCA configuration, characterized in that: The specific contents of S2 include: Data collation: Clean and screen the collected electronic word-of-mouth data of agricultural products to remove invalid or duplicate data; Factor extraction: extracting key factors that may affect e-WOM communication from the data, such as content quality, consumer interaction, and product characteristics; Variable setting: determine the conditional variables and outcome variables; Conditional variables include content quality, consumer behavior, and communication channels; specifically, they include content standards and content sentiment, like interactions, comment interactions, sharing interactions, publisher influence, and publisher professionalism; The outcome variable is the effect of electronic word-of-mouth communication; the outcome variable is measured by the communication index and sales revenue. The weights are calculated by the entropy method after normalization to ensure that the weight distribution is objective. The specific steps are as follows: 1) The diffusion index is x1, the sales revenue is x2, and the normalized formula is assuming that there are m samples: Among them, x ij is the original value of the jth indicator of the ith sample, min(x j ) and max(x j ) is the minimum and maximum value of index j, x′ ij is the normalized value, which represents the standardized value of the jth indicator of the i-th sample; 2) The entropy method determines the weight of each indicator based on its distribution information. The steps are as follows: 21) Calculate the normalized value ratio p of each indicator ij : Among them, x′ ij is the normalized value of the jth indicator of the i-th sample; 22) Calculate the entropy value e j , the entropy value e of each indicator j It indicates the uncertainty of its information. The smaller the entropy value, the more important the indicator is. Among them, p ij is the ratio of the jth indicator of the ith sample, and ln(m) is the natural logarithm of the number of samples; 23) Calculate the weight w j , the larger the weight, the more important the indicator is; Among them, e j is the entropy value of the jth indicator, n is the number of indicators; 24) Calculate the comprehensive result variable. The final communication effect evaluation Y is obtained by the normalized value of the weighted average communication index and sales revenue: Y=w1·x′1+w2·x′2 Among them, x′1 is the normalized value of the communication index, x′2 is the normalized value of sales revenue, w1 and w2 are the weights of the communication index and sales revenue, respectively, calculated by the entropy method; 25) Finally, all the above variables are normalized.

4. According to claim 1, a method for identifying paths based on FsQCA agricultural product electronic reputation configuration, characterized in that: The specific contents of S3 include: The influencing factors of agricultural product electronic word-of-mouth are taken as conditional variables, and the communication effect of agricultural product electronic word-of-mouth is taken as the outcome variable. The calibration anchor points of the conditional variables are calculated using the direct calibration method. For fixed-distance and fixed-ratio variables, the quantile method is used to set three calibration anchor points.

5. According to claim 1, a method for identifying paths of electronic reputation of agricultural products based on FsQCA configuration, characterized in that: The specific contents of S4 include: Based on the calibration anchor point results of the variables, the consistency and coverage between the single conditional variables and the outcome variables are calculated, and the steps of determining the necessary conditions for the spread of electronic word-of-mouth of agricultural products include: The consistency between the conditional variable and the outcome variable is calculated by the following formula: Consistency is used to measure the degree of causal matching between conditional variable X and outcome variable Y. It reflects the strength of causal matching by calculating the proportion of the intersection of the conditional variable membership and the outcome variable membership in the conditional variable membership. It indicates the proportion of outcome variable Y that is satisfied when conditional variable X is satisfied. The consistency value range is [0, 1]. The closer the result is to 1, the stronger the causal matching between X and Y. X i represents the membership value of the conditional variable in the i-th sample, Y i Indicates the membership value of the outcome variable in the i-th sample, min(X i ,Y i ) indicates that the conditional variable and the outcome variable membership take the smaller value of the two, ∑min(X i ,Y i ) is the sum of X and Y in all samples, ∑X i is the sum of the membership of the conditional variables, and the ratio indicates whether the membership of the conditional variables is sufficient to cover the outcome variable; The coverage between the condition variable and the outcome variable is calculated by the following formula: Coverage reflects how much of the realization of the outcome variable Y can be explained by the condition variable X; by calculating the proportion of the intersection of the condition variable and the outcome variable to the total amount of the outcome variable, the coverage value range is [0, 1]. The closer the result is to 1, the greater the range of Y that X can explain; ∑Y i It is the sum of the membership degrees of the outcome variable, and the ratio reflects the proportion of the outcome variable Y that can be explained by the conditional variable X; Consistency focuses on whether the conditional variable can reliably predict the outcome variable, and coverage focuses on the degree to which the conditional variable explains the outcome variable. In causal analysis, high consistency indicates a strong causal relationship, but if the coverage is low, it means that the influence of the condition is limited. Conditional variables with consistency values ​​greater than the consistency threshold are regarded as necessary conditions for the outcome variable.

6. According to claim 1, a method for identifying paths of electronic reputation of agricultural products based on FsQCA configuration, characterized in that: The specific contents of S5 include: An original consistency threshold is set, the consistency and coverage between the conditional configuration and the result variable are calculated, and simple solutions, intermediate solutions and complex solutions are obtained through configuration sufficiency analysis; the core conditions and marginal conditions are determined based on the simple solutions, intermediate solutions and complex solutions, and the configuration configuration for the successful electronic word-of-mouth communication of agricultural products under the preset original consistency threshold conditions is obtained, and multiple guided paths are obtained through screening, identification and classification of various configuration features.

7. The method for identifying the path of electronic reputation of agricultural products based on FsQCA configuration according to claim 6 is characterized in that: Also includes: Robustness test: Raise the consistency threshold to verify the stability of configuration results; If the experiment shows that the path results are consistent under different consistency settings, the robustness of the analysis is verified.

Images