Heat calculation method and device, electronic equipment and storage medium

Abstract

The invention provides a popularity calculation method and device, electronic equipment and a storage medium, and the method comprises the steps: determining core parameters for a target vehicle type based on behavior data of a user; the core parameters comprise at least one of user asset data, public opinion data and content data; performing fitting processing on the core parameters through a regression fitting model to obtain an initial fitting model of the hot indexes of the target vehicle type; under the condition that evaluation of the initial fitting model is not passed, parameters to be adjusted of the core parameters in the initial fitting model are adjusted, and a target popularity model is obtained; and determining a popularity index of the target vehicle type according to the target popularity model.

Description

Technical Field

[0001] The present disclosure relates to the technical field of computers, and in particular, to a heat calculation method, device, electronic device and storage medium. Background Art

[0002] With the rapid development of computer technology, the application of automobile service platform has emerged. Automobile service platform can provide users with real and professional automobile content and convenient and reliable purchase services. At the same time, automobile service platform can also show users the popularity ranking of various car series, so as to provide users with reference when they have purchase needs.

[0003] However, with the continuous increase in the number of new energy brands and penetration rates in the industry, changes in the behavioral characteristics of young consumers, the maturity of the data system and the changes in the ecological system within the automobile service platform, the diversification of the operating methods of brand car series on the automobile service platform and other objective facts, the current calculation method of the heat parameter is relatively simple, and it is difficult to intuitively reflect the brand's long-term asset accumulation, and thus it is impossible to accurately reflect the actual needs of users for each car series, which in turn leads to a poor user experience. Summary of the invention

[0004] The embodiments of the present disclosure at least provide a heat calculation method, device, electronic device and storage medium.

[0005] In a first aspect, an embodiment of the present disclosure provides a heat calculation method, including:

[0006] Determine core parameters for the target vehicle type based on the user's behavior data; the core parameters include at least one of the following: user asset data, public opinion data, and content data;

[0007] Fitting the core parameters through a regression fitting model to obtain an initial fitting model of the popularity index of the target vehicle type;

[0008] When the evaluation of the initial fitting model fails, adjusting the parameters to be adjusted of the core parameters in the initial fitting model to obtain a target heat model;

[0009] A popularity index of the target vehicle type is determined according to the target popularity model.

[0010] In an optional implementation manner, determining the core parameters for the target vehicle type based on the user's behavior data includes:

[0011] Acquire the user's behavior data for the target vehicle type, and calculate the behavior correlation between the user and the target vehicle type based on the behavior data;

[0012] The intention degree of each user for the target vehicle type is determined based on the behavior correlation, and users with different intention degrees are counted as user asset data of different levels for the target vehicle type.

[0013] In an optional implementation manner, determining the core parameters for the target vehicle type based on the user's behavior data includes:

[0014] Determining basic behavior data from the user's behavior data for the target vehicle type; wherein the basic behavior data includes public opinion data and / or content data;

[0015] A preset number of basic behavior data whose correlation with the popularity index of the target vehicle type meets the requirement is determined, and public opinion data and / or content data in the core parameters are obtained.

[0016] In an optional implementation manner, the core parameters are fitted by a regression fitting model to obtain an initial fitting model of the target vehicle type, including:

[0017] The core parameters are fitted by the regression fitting model according to preset constraints to obtain exponential parameters and linear parameters of each core parameter, wherein the preset constraints include a first condition and / or a second condition, the first condition being that each exponential item in the initial fitting model changes in the same direction as the hot index, and the second condition being that the marginal utility of each exponential item in the initial fitting model decreases.

[0018] In an optional implementation, when the evaluation of the initial fitting model fails, adjusting the parameters to be adjusted of the core parameters in the initial fitting model to obtain the target heat model includes:

[0019] Performing model evaluation on the initial fitting model to obtain an evaluation result; wherein the evaluation result includes a rationality evaluation result and / or a stress resistance evaluation result;

[0020] When it is determined based on the evaluation result that the evaluation of the initial fitting model fails, the parameters to be adjusted of the core parameters in the initial fitting model are adjusted to obtain a target heat model.

[0021] In an optional implementation manner, the performing model evaluation on the initial fitting model to obtain an evaluation result includes:

[0022] Based on the ranking of the hot index determined by the initial fitting model and / or the change in the ranking of the hot index, the rationality of the initial fitting model is evaluated to obtain the rationality evaluation result;

[0023] Based on the degree of influence of the network delivery data for the target vehicle type on the popularity index determined based on the initial fitting model, the initial fitting model is evaluated for stress resistance to obtain the stress resistance evaluation result.

[0024] In an optional implementation, after determining the popularity index of the target vehicle type according to the target popularity model, the method further includes:

[0025] Determine the historical key events of the target vehicle type and the change data of the popularity index at the historical moment corresponding to the historical key events;

[0026] Training a neural network model based on the historical key events and the change data;

[0027] According to the trained neural network model, the change information of the hot index corresponding to the key events to be occurred is predicted.

[0028] In a second aspect, an embodiment of the present disclosure provides a heat calculation device, including:

[0029] A first determination unit is used to determine core parameters for a target vehicle type based on user behavior data; the core parameters include at least one of the following: user asset data, public opinion data, and content data;

[0030] A construction unit, configured to perform fitting processing on the core parameters through a regression fitting model to obtain an initial fitting model of the target vehicle type;

[0031] An adjusting unit, configured to adjust the to-be-adjusted parameters of the core parameters in the initial fitting model to obtain a target heat model when the initial fitting model fails to pass the evaluation;

[0032] The second determining unit is used to determine the popularity index of the target vehicle type according to the target popularity model.

[0033] In a third aspect, an embodiment of the present disclosure further provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the above-mentioned first aspect, or any possible implementation of the first aspect are performed.

[0034] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned first aspect or any possible implementation of the first aspect are executed.

[0035] The disclosed embodiment provides a heat calculation method, device, electronic device and storage medium. In the disclosed embodiment, first, the core parameters for the target vehicle type are determined based on the user's behavior data; wherein the core parameters include at least one of the following: user asset data, public opinion data, content data; then, the core parameters are fitted through a regression fitting model to obtain an initial fitting model of the target vehicle type; if the initial fitting model fails to pass the evaluation, the parameters to be adjusted of the core parameters in the initial fitting model are adjusted to obtain a target heat model; finally, the popularity index of the target vehicle type is determined according to the target heat model.

[0036] From the above description, it can be seen that by determining the core parameters including at least user asset data, public opinion data and content data, the in-depth user behavior of users related to vehicle brands on the automobile service platform can be reflected, thereby reflecting the user's intentions and preferences for each vehicle brand, providing a reasonable calculation basis for the calculation of the hot index, and being able to more reasonably reflect the vehicle brand assets and popularity. By fitting the core parameters and adjusting the parameters of the initial fitting model that failed the evaluation, the model can improve its reflection of the popularity of the vehicle brand, and further improve the accuracy and stability of the hot index determined by the target heat model.

[0037] In order to make the above-mentioned objectives, features and advantages of the present disclosure more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following is a brief introduction to the drawings required for use in the embodiments. The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and are used together with the specification to illustrate the technical solutions of the present disclosure. It should be understood that the following drawings only illustrate certain embodiments of the present disclosure and should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can also be obtained based on these drawings without creative work.

[0039] Figure 1 A flow chart of a heat calculation method provided by an embodiment of the present disclosure is shown;

[0040] Figure 2 A schematic diagram showing the influence of the improvement of core parameters on the popularity index under different quantiles of the popularity index provided by an embodiment of the present disclosure is shown;

[0041] Figure 3 A schematic diagram of a heat calculation device provided by an embodiment of the present disclosure is shown;

[0042] Figure 4 A schematic diagram of an electronic device provided by an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical scheme and advantages of the embodiments of the present disclosure clearer, the technical scheme in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all of the embodiments. The components of the embodiments of the present disclosure generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present disclosure provided in the drawings is not intended to limit the scope of the present disclosure for protection, but merely represents the selected embodiments of the present disclosure. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present disclosure.

[0044] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0045] The term "and / or" herein only describes an association relationship, indicating that three relationships may exist. For example, A and / or B may represent the following three situations: A exists alone, A and B exist at the same time, and B exists alone. In addition, the term "at least one" herein represents any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C may represent including any one or more elements selected from the set consisting of A, B, and C.

[0046] Research has found that the popularity index is an asset indicator that is visible, predictable, and interventionable on the consumer side and the product side, and has industry credibility. With the continuous rise of new energy brands and penetration rates in the industry, changes in the behavioral characteristics of young consumers, the maturity of the data system and the changes in the ecological system on the automotive service platform, and the diversification of the operating methods of brand car series on the automotive service platform, the current calculation method of the popularity parameter is relatively simple, and it is difficult to intuitively reflect the brand's long-term asset accumulation, and thus it is impossible to accurately reflect the user's actual needs for each car series, which in turn leads to a poor user experience.

[0047] Based on the above research, the present disclosure provides a heat calculation method, device, electronic device and storage medium. In the embodiment of the present disclosure, first, the core parameters for the target vehicle type are determined based on the user's behavior data; wherein the core parameters include at least one of the following: user asset data, public opinion data, content data; then, the core parameters are fitted through a regression fitting model to obtain an initial fitting model of the target vehicle type; if the evaluation of the initial fitting model fails, the parameters to be adjusted of the core parameters in the initial fitting model are adjusted to obtain a target heat model; finally, the popularity index of the target vehicle type is determined according to the target heat model.

[0048] From the above description, it can be seen that by determining the core parameters including at least user asset data, public opinion data and content data, the in-depth user behavior of users related to vehicle brands on the automobile service platform can be reflected, thereby reflecting the user's intentions and preferences for each vehicle brand, providing a reasonable calculation basis for the calculation of the hot index, and being able to more reasonably reflect the vehicle brand assets and popularity. By fitting the core parameters and adjusting the parameters of the initial fitting model that failed the evaluation, the model can improve its reflection of the popularity of the vehicle brand, and further improve the accuracy and stability of the hot index determined by the target heat model.

[0049] To facilitate understanding of this embodiment, a heat calculation method disclosed in the embodiment of the present disclosure is first introduced in detail. The execution subject of the heat calculation method provided in the embodiment of the present disclosure is generally an electronic device with certain computing capabilities. In some possible implementations, the heat calculation method can be implemented by a processor calling a computer-readable instruction stored in a memory.

[0050] See also Figure 1 FIG. 1 is a flow chart of a heat calculation method provided by an embodiment of the present disclosure, wherein the method includes steps S101 to S104, wherein:

[0051] S101: Determine core parameters for a target vehicle type based on user behavior data; the core parameters include at least one of the following: user asset data, public opinion data, and content data.

[0052] In the disclosed embodiment, the target vehicle type may be a specified vehicle series, a specified vehicle model, or a specified vehicle brand; the user's behavior data may be the user's operational behavior data for the target vehicle type on the automobile service platform.

[0053] The automobile service platform in the disclosed technical solution is a comprehensive platform for content, product information and dealer information for car buyers and car enthusiasts. The ranking of automobile popularity in the automobile service platform should be based on the user's online behavior related to the target vehicle type, including but not limited to advertising browsing, article reading, video watching, commenting, collecting, forwarding, active search, quotation inquiry, product parameter inquiry, product comparison, click jump, inquiry, coupon collection and other information. The above user behavior data together with the completeness of the basic infrastructure of the car series content and other modules jointly determine the popularity ranking of the car series on the automobile service platform.

[0054] Therefore, the fitting of the popularity index in the disclosed technical solution aggregates multiple core parameters such as user assets, content data, public opinion data, etc. of the vehicle series, and constructs a popularity index for indicating the popularity of the target vehicle type based on the multiple core parameters. The popularity index can more intuitively measure the popularity of each target vehicle type, so that the customer can be provided with a corresponding basis for adjusting the operation strategy based on the popularity index.

[0055] Here, the user asset data in the core parameters is used to indicate the level information of the user in the terminal relative to a single target vehicle type, through which the user's intention to the target vehicle type can be determined, for example, the user's intention to a certain car series. The content data in the core parameters can include the user's stay time, key touchpoint behavior, frequency, the competitive landscape of the current target vehicle type's competitors, decision cycle and other data.

[0056] In the embodiment of the present disclosure, partial behavior data that meets the relevance requirement is pre-screened from a variety of behavior data based on the user's behavior data, and then the core parameters are determined according to the partial behavior data.

[0057] In the technical solution disclosed in the present invention, the fitting model of the hot index (also known as the car understanding index) needs to solve the long-term and short-term change trends and cumulative effects of the brand assets of the car series in the industry, especially in the car service platform. Here, multiple behavioral data with the strongest correlation can be prioritized in advance from dozens of behavioral data involving the user level as core parameters. Among them, the core parameters have a strong correlation with the model fitting, and the degree of correlation between each core parameter and the hot index is greater than or equal to the preset correlation threshold, so that it has factual business representativeness for the asset accumulation of the car series brand.

[0058] S102: Fitting the core parameters through a regression fitting model to obtain an initial fitting model of the popularity index of the target vehicle type.

[0059] In the initial fitting model, each core parameter includes a corresponding model parameter. The core parameters are fitted by the regression fitting model to determine the specific parameter value of the model parameter of each core parameter. Here, the model parameter of each core parameter can be determined by the exponential fitting model and the linear regression model.

[0060] S103: When the evaluation of the initial fitting model fails, adjusting the parameters to be adjusted of the core parameters in the initial fitting model to obtain a target heat model.

[0061] After obtaining the initial fitting model, the initial fitting model can be evaluated for rationality and stress resistance to obtain an evaluation result. If it is determined according to the evaluation result that the evaluation of the initial fitting model has not passed, then the parameter to be adjusted of the core parameter in the initial fitting model can also be adjusted. The parameter to be adjusted is the model parameter of the core parameter in the initial fitting model. After adjusting the parameter to be adjusted, if it is determined that the evaluation of the initial fitting model after the adjustment of the parameter is passed, the initial fitting model after the adjustment of the parameter to be adjusted is determined as the target heat model.

[0062] S104: Determine a popularity index of the target vehicle type according to the target popularity model.

[0063] After determining the target popularity model in the above manner, the parameter values ​​of the core parameters within a preset time period can be counted and substituted into the target popularity model to obtain the popularity index (ie, the car knowledge index) of the target vehicle type.

[0064] In the disclosed embodiment, the car understanding index can be understood as an indicator for evaluating the performance of a car series. The car understanding index needs to comprehensively consider the performance of the car series on the car service platform, and then abstract the various performances into quantitative dimensions, and then adjust the influence weights of each quantitative dimension in the car understanding index, so as to aggregate into a car understanding index, and use this car understanding index as the ranking basis for the car series hot list.

[0065] The above steps will be described in detail below in conjunction with specific implementation methods.

[0066] In an optional implementation, step S101 determines the core parameters for the target vehicle type based on the user's behavior data, and specifically includes the following steps:

[0067] Step S11: obtaining the user's behavior data for the target vehicle type, and counting the behavior correlation between the user and the target vehicle type based on the behavior data;

[0068] Step S12: determining the intention degree of each user for the target vehicle type based on the behavior correlation, and counting users with different intention degrees as different levels of user asset data for the target vehicle type.

[0069] Since there are many types of behavior data, it is not conducive to fitting the car knowledge index. Therefore, the disclosed technical solution divides the user's behavior data into several categories, divides the users into different intention depths according to preset rules (for example, the frequency or type of user behavior), and counts users with different intention depths as user assets of different levels of car series. This user asset classification method can abstract various behavior data into user assets of limited dimensions.

[0070] In the disclosed embodiment, the user's behavior data can be statistically processed according to preset rules to obtain the behavior correlation between the user and the target vehicle type. For example, the following behavior correlations can be obtained: A1 aware, A2 appeal, A3 ask, A4 act, and A5 advocate.

[0071] In specific implementation, corresponding statistical indicators are set for each behavior correlation. Here, the behavior data can be statistically processed according to the statistical indicators of each behavior correlation, so as to determine the behavior correlation according to the statistical results. For example, if the statistical results are determined to meet the statistical indicators of any behavior correlation, then the behavior correlation is determined to be the user's intention level for the target vehicle type. If the statistical results meet the statistical indicators of multiple behavior correlations, then the user's intention level for the target vehicle type, that is, the user level, is determined according to the highest level of behavior correlation among the multiple behavior correlations.

[0072] Assume that the behavioral relevance includes: A1 aware, A2 appeal, A3 ask, A4 act, and A5 advocate. Then the statistical indicators of each behavioral relevance can be as follows:

[0073] A1:

[0074] Platform A (last 30 days)

[0075] Advertising: 1-3 exposures;

[0076] Content: 1 exposure;

[0077] Live broadcast: 1 entry / watching time <10 seconds / card display 1-3 times.

[0078] A2 appeal:

[0079] Platform A (last 30 days)

[0080] Advertising: 4-15 impressions / 1 click;

[0081] Content: exposure >= 2 times / reading 1-2 times / reading time 0-30 seconds;

[0082] Live broadcast: 2-10 times of entering live broadcast / 10-300 seconds of live broadcast viewing / card display times>=4 times;

[0083] E-commerce: Coupon exposure.

[0084] A3 Inquiry (ask):

[0085] Platform A (last 30 days)

[0086] Advertisement: Exposure>=15 times / Click>=2 times; for example, "Exposure>=15 times" or "Click>=2 times", "Stay on the new car product page for more than 10 seconds" or "Click to submit a guess price" or "Click to answer a question" or "Highlight module exposure" or "Click to book a test drive" or "PK comparison module exposure";

[0087] Content: Exposure>=3 times / Effective reading>=1 / Duration>=30 seconds / Repost, comment, like / Read other content; for example, "Read>=3 times" or "Effective reading>=1 time" or "Duration>=30 seconds" or "Like" or "Collect" or "Repost" or "Comment" or "Read XXX content";

[0088] Live broadcast: Number of times of entering live broadcast>=10 times / Live broadcast viewing duration>300 seconds / Repost, comment, like, etc.; for example, "Number of times of entering live broadcast>=10 times" or "Live broadcast viewing duration>=300 seconds" or "Like" or "Comment" or "Follow the host" or "Send a gift";

[0089] Search: >= 1 times;

[0090] E-commerce: "Click on the coupon" or "Superstore - click on the car series card" or "E-commerce - receive the coupon" or "Superstore - car series details page exposure";

[0091] And other related content, such as:

[0092] IM (Instant Messaging): IM consultation;

[0093] Car knowledge score: "Car knowledge score ranking" or "Enter car knowledge score car series details");

[0094] Model Library: "Go to the model page to learn about prices and dealers";

[0095] Car Friends Circle: "Enter the car car friends circle".

[0096] A4 Action (act):

[0097] Platform A (last 90 days)

[0098] Users who have submitted leads, for example, users who have submitted leads within the last 90 days.

[0099] A5 advocate

[0100] A platform

[0101] Brand certified car owner and the latest review / comment is non-negative.

[0102] It can be seen from the above description that corresponding statistical indicators are set for each behavior correlation, and for the same type of statistical indicators, the indicator values ​​of different behavior correlations are different. And the indicator types of statistical indicators corresponding to different behavior correlations are not exactly the same. Here, the indicator type of the statistical indicator of each behavior correlation and the indicator value of each statistical indicator can be set according to actual needs. At this time, the behavior data can be counted according to the statistical indicators of each behavior correlation, so as to determine the behavior correlation between the user and the target vehicle type according to the statistical results, that is, the degree of intention. Among them, for the above-mentioned behavior correlations A1-A5, the user's intention degree gradually increases, and the deeper the user's behavior depth, the closer to the decision.

[0103] In an optional implementation, step S101 determines the core parameters for the target vehicle type based on the user's behavior data, and specifically includes the following steps:

[0104] Step S13: determining basic behavior data from the user's behavior data for the target vehicle type; wherein the basic behavior data includes public opinion data and / or content data;

[0105] Step S14: Determine a preset number of basic behavior data whose correlation with the popularity index of the target vehicle type in the basic behavior data meets the requirements, and obtain the public opinion data and / or content data in the core parameters.

[0106] In the disclosed embodiments, multiple basic behavior data may be pre-screened from the full amount of behavior data. For example, M behavior data may be screened from public opinion data and content data as basic behavior data according to a sampling algorithm. For another example, the basic behavior data may also be behavior data pre-selected by relevant technical personnel from the full amount of behavior data. For another example, M strongly correlated behavior data may be screened as basic behavior data according to a correlation algorithm.

[0107] After screening the basic behavior data, the basic behavior data can be input into the deep learning model for processing, so as to output the correlation degree between the basic behavior data and the popular index; finally, according to the correlation degree, a preset number of basic behavior data with correlation degrees that meet the requirements are screened out from the basic behavior data, and the screened basic behavior data and user asset data are determined as core parameters.

[0108] In the embodiment of the present disclosure, the following core parameters can be determined through the above processing method:

[0109] X1, with a scale of 0 to 11,000,000, is used to reflect the potential user base of the car series on the platform and the scope and intensity of advertising;

[0110] X2, 0-13,000,000, used to reflect the volume of interested users on the platform and the scope and intensity of advertising for this car series;

[0111] X3, with a scale of 0 to 6,500,000, is used to reflect the number of users who follow this car series on the platform, as well as the popularity of its content and advertisements;

[0112] X4, with a scale of 0 to 450,000, is used to reflect the volume of potential car buyers on the platform;

[0113] X5, with a scale of 0 to 25,000, is used to reflect the number of fans of this car series on the platform;

[0114] X6, with a scale of 0 to 30,000,000, is used to reflect the size of all user assets of this car series on the platform;

[0115] X7, the scale is 0 to 700, which is used to reflect the diversity of the platform content of this car series;

[0116] X8, with a scale of 0 to 200,000; X9, with a scale of 0 to 100,000, which are used to reflect the content consumption, content interaction, dissemination trend and coverage of the car series on the platform;

[0117] X 10 , ranging from 0 to 25,000, which is used to reflect the number of loyal user assets of this car series on the platform;

[0118] X 11 , ranging from 0 to 6,500,000, which is used to reflect the performance of the car series in the competitive circle;

[0119] X 12 , with a magnitude of 0 to 1,500,000, which is used to reflect the trend of user assets changes of the car series on the platform;

[0120] X 13 , with a scale of 0 to 10,000, is an indicator used to reflect the level of public opinion for the car series on the platform.

[0121] Here, the above core parameters X1-X 13 The correlation with the hot index can be obtained by Figure 2 To express. Figure 2 The horizontal axis in the middle is the various quantiles of the popularity index. Figure 2 The vertical axis in the middle represents the impact of the improvement of core parameters on the popularity index under different quantiles of the popularity index (or the contribution weight of the core parameters to the popularity index). Figure 2 It can be seen that when the percentile of the popularity index is low, X9 has a greater impact on the popularity index, but as the popularity index increases, the impact of X9 on the popularity index gradually decreases.

[0122] Prior to this, a deep learning model can be pre-trained. Specifically, a training sample can be constructed; wherein the training sample includes basic behavior data and label values. Here, the training sample can be input into the deep learning model to be trained for processing, so as to obtain a predicted value of the degree of association; thereafter, the loss function value can be calculated based on the predicted value and the true value (i.e., the label value), and then the model parameters of the deep learning model to be trained can be adjusted based on the loss function value, so as to obtain a trained deep learning model. Thereafter, the trained deep learning model can be used to screen a preset number of basic behavior data whose degree of association meets the requirements.

[0123] Through the above processing method, core parameters with strong correlation to model fitting can be screened out, and the correlation between each core parameter and the popularity index is greater than or equal to the preset correlation threshold. The core parameters screened out above can more accurately reflect user decision-making intentions and brand preferences, innovatively improving the representativeness and popularity of the popularity model for car brand assets.

[0124] In an optional implementation, the above step S102 performs fitting processing on the core parameters through a regression fitting model to obtain an initial fitting model of the target vehicle type, which specifically includes the following steps:

[0125] The core parameters are fitted by the regression fitting model according to preset constraints to obtain exponential parameters and linear parameters of each core parameter, wherein the preset constraints include a first condition and / or a second condition, the first condition being that each exponential item in the initial fitting model changes in the same direction as the hot index, and the second condition being that the marginal utility of each exponential item in the initial fitting model decreases.

[0126] The amount of behavior indicates the depth of intention. Similarly, core parameters such as user assets calculated based on basic behavior data also reflect the user's intention or the popularity of the car series. All core parameters change in the same direction as the popularity index (increasing function). However, the popularity index should comprehensively reflect the performance of the car series in various dimensions on the car service platform, so the impact of the increase of a single core parameter on the car knowledge index should be constructed in a way of diminishing marginal utility.

[0127] Based on this, a constraint condition, namely a preset constraint condition, is pre-constructed, wherein the preset constraint condition includes a first condition and / or a second condition. The first condition is that each index item in the initial fitting model changes in the same direction as the hot index.

[0128] Assume that the calculation formula of the initial fitting model of the hot index is:

[0129]

[0130] In this initial fitted model, That is, an exponential term. The first condition requires that each exponential term in the initial fitting model changes in the same direction as the popular index, that is, an increasing function. The second condition requires that the marginal utility of each exponential term in the initial fitting model decreases.

[0131] Based on the above two conditions, the parameters to be adjusted are constructed for each core parameter, namely the exponential parameter and the linear parameter. In the above initial fitting model, X1-X 13 is the core parameter, C1-C 13 is the linear parameter of the core parameter, P1-P 14 The exponential parameter of the core parameter.

[0132] Here, C1-C 13 The magnitude range can be selected as 1-560, and the linear weight of the corresponding core parameter can be adjusted through the linear parameter.

[0133] P1-P 14 The magnitude of can be selected as 1 to 4. Since all core parameters and hot indexes are increasing functions with decreasing marginal utility, the impact of fluctuations in parameter values ​​on the hot index can be controlled by P.

[0134] In addition, the initial fitting model also includes a parameter ρ, the magnitude of which can be selected as 1000-7000. This parameter can be used to control the overall fluctuation range of the popularity index, so that the popularity index of most car series is within an index range that is easy to observe, and there is a certain degree of separation between the indexes of car series with different popularity.

[0135] In the embodiment of the present disclosure, the core parameters can be fitted according to the exponential fitting model and the linear regression model, so as to obtain the exponential parameters, linear parameters and parameter ρ of each core parameter. In the fitting process, each exponential term of the core parameter is required to satisfy the first condition and the second condition, so the fitting process can be repeatedly performed to obtain the exponential term that satisfies the first condition and the second condition, and then obtain the initial fitting model.

[0136] In an optional implementation, in the above step S103, when the evaluation of the initial fitting model fails, adjusting the parameters to be adjusted of the core parameters in the initial fitting model to obtain the target heat model specifically includes the following steps:

[0137] Step S21: performing a model evaluation on the initial fitting model to obtain an evaluation result; wherein the evaluation result includes a rationality evaluation result and / or a stress resistance evaluation result;

[0138] Step S22: when it is determined based on the evaluation result that the evaluation of the initial fitting model has not passed, adjusting the parameters to be adjusted of the core parameters in the initial fitting model to obtain a target heat model.

[0139] After fitting the linear parameters and exponential parameters of each core parameter, an initial fitting model can be obtained. Next, the initial fitting model needs to be evaluated for rationality and stress resistance to obtain evaluation results, including rationality evaluation results and / or stress resistance evaluation results.

[0140] Next, when it is determined that the evaluation of the initial fitting model has not passed based on the rationality evaluation results and / or the stress resistance evaluation results, the linear parameters and / or exponential parameters of the core parameters in the initial fitting model can be adjusted. Then, the adjusted initial fitting model is re-evaluated, wherein, if it is determined that the evaluation of the initial fitting model has passed according to the new evaluation results, the initial fitting model is determined as the target heat model; otherwise, the parameters to be adjusted of the core parameters in the initial fitting model are repeatedly adjusted until the stop condition is met. The stop condition may be that the evaluation of the initial fitting model has passed; or, the number of adjustments of the parameters to be adjusted in the initial fitting model meets the preset number threshold. If the number of adjustments of the parameters to be adjusted in the initial fitting model reaches the preset number threshold, and the first condition and the second condition are still not met, at this time, you can choose to re-screen the core parameters, and repeatedly perform the above steps S102 to S104; until an initial fitting model that has passed the evaluation is obtained.

[0141] In the above implementation, by evaluating the initial fitting model and adjusting the core parameters if the evaluation fails, the accuracy of the popularity index can be further guaranteed, thereby providing users with more accurate recommendations for popular car series.

[0142] In an optional implementation, the above step performs model evaluation on the initial fitting model to obtain an evaluation result, specifically including:

[0143] First, based on the ranking of the hot index determined by the initial fitting model and / or the change in the ranking of the hot index, the rationality of the initial fitting model is evaluated to obtain the rationality evaluation result;

[0144] Secondly, based on the influence of the network delivery data for the target vehicle type on the popularity index determined based on the initial fitting model, the initial fitting model is evaluated for stress resistance to obtain the stress resistance evaluation result.

[0145] In the disclosed embodiment, the popularity index of each target vehicle type can be determined through the initial fitting model, and the popularity index of each target vehicle type can be ranked; in addition, the ranking change of the popularity index of the target vehicle type can also be determined. For example, the ranking change of the popularity index of a benchmark vehicle series during its life cycle can be determined; or the change of the popularity index of a specified vehicle series under a key event can be determined.

[0146] Next, the rationality of the initial fitting model can be evaluated based on the ranking of the hot index and / or the change in the ranking of the hot index.

[0147] For example, the popularity index can be compared with information such as car sales, retention frequency, and user asset ranking, so as to determine the rationality of the popularity index based on the comparison results. For example, if the popularity index is proportional to the car sales, retention frequency, and user asset ranking, then the popularity index is reasonable.

[0148] For example, a rationality assessment can also be conducted based on the absolute value and ranking changes of the popularity index of the benchmark vehicle series during its life cycle.

[0149] For example, we can determine whether the ranking changes of the car knowledge index after a key event are reasonable. For example, negative public opinion will cause the popularity index of a car series to drop.

[0150] In the disclosed embodiment, the influence of network delivery data on the popularity index can also be analyzed. Specifically, the influence of the delivery of commercial products and advertisements on the popularity index can be counted. Generally speaking, the delivery of a large number of advertisements will increase the popularity index of the corresponding car series. In addition, the initial fitting model can also be evaluated for its stress resistance by the influence of non-natural behaviors such as a large number of low-quality posts on the popularity index.

[0151] In the above implementation, through rationality evaluation and stress resistance evaluation, the initial fitting model can be verified and tested from multiple angles, thereby further ensuring the accuracy and effectiveness of the target heat model.

[0152] In the disclosed embodiment, after the target popularity model is obtained, the popularity index of each car series can be calculated according to the target popularity model. For example, the popularity index can be calculated at a preset interval, for example, the popularity index can be calculated daily, and the daily popularity index is used as the car series popularity ranking to display the car series popularity list.

[0153] In the disclosed embodiment, the popularity index of the car series can also be calculated according to the specified dimensions, wherein the specified dimensions include user attributes (gender, age), geographic location, etc. For example, the basic behavior data can be split based on the user's location information, and the popularity index of the specified area and the ranking of the popularity index can be calculated based on the split basic behavior data.

[0154] After the popular car series are calculated in the manner described above, the recommended car series may be filtered for the user according to the ranking of the popularity index.

[0155] In the disclosed embodiment, after determining the popularity index of the target vehicle type according to the target popularity model, the method further includes the following steps:

[0156] Step S105: determining the historical key events of the target vehicle type and the change data of the hot index at the historical moment corresponding to the historical key events;

[0157] Step S106: training a neural network model according to the historical key events and the change data;

[0158] Step S107: predicting the change information of the hot index corresponding to the key event to be occurred according to the trained neural network model.

[0159] In the disclosed embodiment, key events include key nodes of the vehicle series, such as new car launches, vehicle generation changes, mid-term facelifts, and other nodes, and also include major events of the vehicle series, such as price cuts, auto shows, new configurations, and other events.

[0160] Here, we can count the key events that occurred at historical moments (i.e., historical key events) and the change data of the popularity index of the corresponding car series under the historical key events; then, train the neural network through the historical key events and change data to enable the neural network to learn the impact of various types of key events on the popularity index.

[0161] Next, the trained neural network can be used to predict the impact of the key event to be occurred on the hot index, and then the change information of the hot index after the key event occurs can be predicted based on the impact. Here, the prediction method selected by the neural network model can be the time prediction method, the pattern prediction method and the multi-line prediction method.

[0162] From the above description, we can see that the hot index can help brands achieve diagnostic analysis. Through the hot index, you can drill down to the relevant parameters of a single car series to achieve detailed analysis and diagnosis, providing support, basis and rapid feedback for the brand's long-term operation strategy.

[0163] The disclosed technical solution is based on the car knowledge index that is representative of the brand assets and popularity of the car series, which is fitted based on the sales volume of the entire network and the deep behavior of users on the terminal. It has newly introduced multi-dimensional parameters such as the user's stay time, key touchpoint behavior, frequency, the competitive landscape of the competing products, the decision-making cycle, and the semantics of public opinion. Through multiple rounds of verification, fitting and adjustment of external sales data and car series sampling data, the model has been comprehensively and innovatively improved in terms of the representativeness and popularity of the car series brand assets. At the same time, through the disassembly of the model itself, it has the ability to be diagnosed. Through in-depth analysis, it can discover the strengths and weaknesses of the car series brand and provide effective improvement strategy suggestions.

[0164] Those skilled in the art will appreciate that, in the above method of specific implementation, the order in which the steps are written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of the steps should be determined by their functions and possible internal logic.

[0165] Based on the same inventive concept, a heat calculation device corresponding to the heat calculation method is also provided in the embodiment of the present disclosure. Since the principle of solving the problem by the device in the embodiment of the present disclosure is similar to the above-mentioned heat calculation method in the embodiment of the present disclosure, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0166] Reference Figure 3 , which is a schematic diagram of a heat calculation device provided by an embodiment of the present disclosure, the device comprises: a first determination unit 10, a construction unit 20, an adjustment unit 30 and a second determination unit 40; wherein,

[0167] A first determination unit is used to determine core parameters for a target vehicle type based on user behavior data; the core parameters include at least one of the following: user asset data, public opinion data, and content data;

[0168] A construction unit, configured to perform fitting processing on the core parameters through a regression fitting model to obtain an initial fitting model of the target vehicle type;

[0169] An adjusting unit, configured to adjust the to-be-adjusted parameters of the core parameters in the initial fitting model to obtain a target heat model when the initial fitting model fails to pass the evaluation;

[0170] The second determining unit is used to determine the popularity index of the target vehicle type according to the target popularity model.

[0171] In a possible implementation manner, the first determining unit is further configured to:

[0172] Acquire the user's behavior data for the target vehicle type, and calculate the behavior correlation between the user and the target vehicle type based on the behavior data;

[0173] The intention degree of each user for the target vehicle type is determined based on the behavior correlation, and users with different intention degrees are counted as user asset data of different levels for the target vehicle type.

[0174] In a possible implementation manner, the first determining unit is further configured to:

[0175] Determining basic behavior data from the user's behavior data for the target vehicle type; wherein the basic behavior data includes public opinion data and / or content data;

[0176] A preset number of basic behavior data whose correlation with the popularity index of the target vehicle type meets the requirement is determined, and public opinion data and / or content data in the core parameters are obtained.

[0177] In a possible implementation manner, the construction unit is further used for:

[0178] The core parameters are fitted by the regression fitting model according to preset constraints to obtain exponential parameters and linear parameters of each core parameter, wherein the preset constraints include a first condition and / or a second condition, the first condition being that each exponential item in the initial fitting model changes in the same direction as the hot index, and the second condition being that the marginal utility of each exponential item in the initial fitting model decreases.

[0179] In a possible implementation manner, the adjustment unit is further configured to:

[0180] Performing model evaluation on the initial fitting model to obtain an evaluation result; wherein the evaluation result includes a rationality evaluation result and / or a stress resistance evaluation result;

[0181] When it is determined based on the evaluation result that the evaluation of the initial fitting model fails, the parameters to be adjusted of the core parameters in the initial fitting model are adjusted to obtain a target heat model.

[0182] In a possible implementation manner, the adjustment unit is further configured to:

[0183] Based on the ranking of the hot index determined by the initial fitting model and / or the change in the ranking of the hot index, the rationality of the initial fitting model is evaluated to obtain the rationality evaluation result;

[0184] Based on the degree of influence of the network delivery data for the target vehicle type on the popularity index determined based on the initial fitting model, the initial fitting model is evaluated for stress resistance to obtain the stress resistance evaluation result.

[0185] In a possible implementation manner, the device is also used for:

[0186] After determining the popularity index of the target vehicle type according to the target heat model, determining historical key events of the target vehicle type and change data of the popularity index at historical moments corresponding to the historical key events;

[0187] Training a neural network model based on the historical key events and the change data;

[0188] According to the trained neural network model, the change information of the hot index corresponding to the key events to be occurred is predicted.

[0189] For descriptions of the processing flow of each module in the device and the interaction flow between each module, reference may be made to the relevant descriptions in the above method embodiment, which will not be described in detail here.

[0190] Corresponds to Figure 1 The heat calculation method in the present disclosure also provides an electronic device 400, such as Figure 4 FIG. 4 is a schematic diagram of the structure of an electronic device 400 provided in an embodiment of the present disclosure, including:

[0191] Processor 41, memory 42, and bus 43; memory 42 is used to store execution instructions, including internal memory 421 and external memory 422; the internal memory 421 is also called internal memory, which is used to temporarily store the operation data in the processor 41 and the data exchanged with the external memory 422 such as a hard disk. The processor 41 exchanges data with the external memory 422 through the internal memory 421. When the electronic device 400 is running, the processor 41 communicates with the memory 42 through the bus 43, so that the processor 41 executes the following instructions:

[0192] Determine core parameters for the target vehicle type based on the user's behavior data; the core parameters include at least one of the following: user asset data, public opinion data, and content data;

[0193] Fitting the core parameters through a regression fitting model to obtain an initial fitting model of the popularity index of the target vehicle type;

[0194] When the evaluation of the initial fitting model fails, adjusting the parameters to be adjusted of the core parameters in the initial fitting model to obtain a target heat model;

[0195] A popularity index of the target vehicle type is determined according to the target popularity model.

[0196] The present disclosure also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the heat calculation method described in the above method embodiment are executed. The storage medium can be a volatile or non-volatile computer-readable storage medium.

[0197] The present disclosure also provides a computer program product that carries a program code. The program code includes instructions that can be used to execute the steps of the heat calculation method described in the above method embodiment. For details, please refer to the above method embodiment, which will not be repeated here.

[0198] The computer program product may be implemented in hardware, software or a combination thereof. In one optional embodiment, the computer program product is implemented as a computer storage medium. In another optional embodiment, the computer program product is implemented as a software product, such as a software development kit (SDK).

[0199] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. In the several embodiments provided in the present disclosure, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of the device or unit can be electrical, mechanical or other forms.

[0200] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0201] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0202] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present disclosure, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0203] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present disclosure, which are used to illustrate the technical solutions of the present disclosure, rather than to limit them. The protection scope of the present disclosure is not limited thereto. Although the present disclosure is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed in the present disclosure, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure, and should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A heat calculation method, characterized in that: include: Determine core parameters for the target vehicle type based on the user's behavior data; the core parameters include at least one of the following: user asset data, public opinion data, and content data; Fitting the core parameters through a regression fitting model to obtain an initial fitting model of the popularity index of the target vehicle type; When the evaluation of the initial fitting model fails, adjusting the parameters to be adjusted of the core parameters in the initial fitting model to obtain a target heat model; A popularity index of the target vehicle type is determined according to the target popularity model.

2. The method according to claim 1, characterized in that The core parameters for the target vehicle type are determined based on the user's behavior data, including: Acquire the user's behavior data for the target vehicle type, and calculate the behavior correlation between the user and the target vehicle type based on the behavior data; The intention degree of each user for the target vehicle type is determined based on the behavior correlation, and users with different intention degrees are counted as user asset data of different levels for the target vehicle type.

3. The method according to claim 1, characterized in that: The core parameters for the target vehicle type are determined based on the user's behavior data, including: Determining basic behavior data from the user's behavior data for the target vehicle type; wherein the basic behavior data includes public opinion data and / or content data; A preset number of basic behavior data whose correlation with the popularity index of the target vehicle type meets the requirement is determined, and public opinion data and / or content data in the core parameters are obtained.

4. The method according to claim 1, characterized in that The step of fitting the core parameters by using a regression fitting model to obtain an initial fitting model of the target vehicle type includes: The core parameters are fitted by the regression fitting model according to preset constraints to obtain exponential parameters and linear parameters of each core parameter, wherein the preset constraints include a first condition and / or a second condition, the first condition being that each exponential item in the initial fitting model changes in the same direction as the hot index, and the second condition being that the marginal utility of each exponential item in the initial fitting model decreases.

5. The method according to claim 1, characterized in that When the evaluation of the initial fitting model fails, adjusting the parameters to be adjusted of the core parameters in the initial fitting model to obtain a target heat model includes: Performing model evaluation on the initial fitting model to obtain an evaluation result; wherein the evaluation result includes a rationality evaluation result and / or a stress resistance evaluation result; When it is determined based on the evaluation result that the evaluation of the initial fitting model fails, the parameters to be adjusted of the core parameters in the initial fitting model are adjusted to obtain a target heat model.

6. The method according to claim 5, characterized in that The performing model evaluation on the initial fitting model to obtain an evaluation result includes: Based on the ranking of the hot index determined by the initial fitting model and / or the change in the ranking of the hot index, the rationality of the initial fitting model is evaluated to obtain the rationality evaluation result; Based on the degree of influence of the network delivery data for the target vehicle type on the popularity index determined based on the initial fitting model, the initial fitting model is evaluated for stress resistance to obtain the stress resistance evaluation result.

7. The method according to claim 1, characterized in that After determining the popularity index of the target vehicle type according to the target popularity model, the method further includes: Determine the historical key events of the target vehicle type and the change data of the popularity index at the historical moment corresponding to the historical key events; Training a neural network model based on the historical key events and the change data; According to the trained neural network model, the change information of the hot index corresponding to the key events to be occurred is predicted.

8. A heat calculation device, characterized in that: include: A first determination unit is used to determine core parameters for a target vehicle type based on user behavior data; the core parameters include at least one of the following: user asset data, public opinion data, and content data; A construction unit, configured to perform fitting processing on the core parameters through a regression fitting model to obtain an initial fitting model of the target vehicle type; An adjusting unit, configured to adjust the to-be-adjusted parameters of the core parameters in the initial fitting model to obtain a target heat model when the initial fitting model fails to pass the evaluation; The second determining unit is used to determine the popularity index of the target vehicle type according to the target popularity model.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate via the bus. When the machine-readable instructions are executed by the processor, the steps of the heat calculation method as described in any one of claims 1 to 7 are performed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of the heat calculation method as described in any one of claims 1 to 7.

Images