Potential Customer Mining Methods Based on Short Video Data

Abstract

This invention relates to the field of customer acquisition technology, specifically disclosing a method for potential customer acquisition based on short video data. The method includes: retrieving account information corresponding to various short videos viewed in the previous time period from a database; obtaining customer personal information corresponding to each account; identifying the content type corresponding to the viewed short videos based on the customer's personal information, and statistically analyzing the viewing time of each content type to generate viewing record information; matching the content type whose viewing time ranks highest among preset categories based on the viewing record information, and identifying the customer as a primary potential customer; obtaining the customer's basic personal information based on the account information of the identified primary potential customer; and identifying the content type corresponding to the customer as a final potential customer based on the customer's basic personal information, and generating corresponding final potential data.

Description

Technical Field

[0001] This invention relates to the field of customer mining technology, and more specifically to a method for mining potential customers based on short video data. Background Technology

[0002] With the development of network communication technology, more and more merchants are selling goods through online stores. In order to better promote their products through short videos, short video platforms have built relevant product transaction processes to attract more users to register on the short video platform. On the short video platform, a key concern for merchants during promotions is how to discover potential customers and accurately reach them.

[0003] Existing potential customer mining systems may rely solely on customer browsing time for judgment and mining. While this method has achieved the goal of mining some potential customers, its accuracy and effectiveness are relatively low.

[0004] Therefore, there is an urgent need for a potential customer mining method based on short video data, which can achieve accurate and effective mining of potential customers and greatly improve the reliability of the mining. Summary of the Invention

[0005] The present invention aims to provide a method for mining potential customers based on short video data, which can achieve accurate and effective mining of potential customers and greatly improve the reliability of the mining.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for potential customer mining based on short video data, comprising the following steps:

[0007] S1. After reaching the preset time period, retrieve the account information corresponding to each short video viewed in the previous time period from the database.

[0008] S2. Based on the corresponding account information, obtain the customer personal information corresponding to each account information, wherein the customer personal information includes the short videos viewed and the viewing duration of the corresponding short videos;

[0009] S3. Based on the short videos viewed by each customer and the corresponding viewing time, identify the content type of the viewed short videos, and based on the content type, count the viewing time of each customer for each content type and generate corresponding viewing record information.

[0010] S4. Based on the browsing history information, match the content type whose browsing time is ranked higher than the preset ranking for the customer, and identify the customer as the primary potential customer corresponding to the matched content type.

[0011] S5. Based on the customer account information identified as a primary potential customer, obtain the customer's basic personal information, which includes personal identity information and personal purchase data information;

[0012] S6. Based on the customer's personal identity information, personal purchase data information, and browsing history information, and using the potential customer identification model, identify the content type corresponding to the ultimate potential customer and generate the corresponding ultimate potential data.

[0013] The principle and advantages of this solution are as follows: Firstly, within a preset time period, it determines whether a customer is a potential customer. Specifically, after the preset time period, it acquires the customer's personal information for each account. This personal information includes the short videos viewed and the corresponding viewing duration. Based on this information, the customer's overall browsing habits can be statistically analyzed, resulting in their browsing history. This history reveals the proportion of each content type viewed by the customer, identifying the content types with the highest viewing duration. The short videos corresponding to these content types can be considered the customer's preferred viewing content. These short videos represent the initial potential customers for the customer, meaning they might be interested in the products featured in the videos and potentially make a purchase.

[0014] To better identify potential customers for products within short videos, when a customer is identified as a primary potential customer for a specific content type, their basic personal information is obtained. This information includes personal identification details and purchase history. Using this data, a potential customer identification model is employed to determine the content type corresponding to that customer as a potential final customer. This data is then saved, allowing for a clear understanding of the potential customers for each content type, thus facilitating better subsequent recommendations and processing. This solution employs multiple identification processes, resulting in high accuracy in identifying potential customers and significantly improving the reliability of the identification process.

[0015] Preferably, as an improvement, S5 includes the following steps:

[0016] S50. Based on the account information of customers identified as primary potential customers, determine whether the customer has commented or sent private messages in the short video. If so, obtain the basic personal information corresponding to the customer's account information. The basic personal information includes personal identity information and personal purchase data information. If not, determine that the customer is not a primary potential customer.

[0017] S51. After obtaining the customer's personal purchase data, analyze the personal purchase data to determine whether the customer has purchased the product corresponding to the matched content type within a preset time period. If so, the personal purchase data is determined to be invalid data; otherwise, the personal purchase data is determined to be valid data.

[0018] Beneficial Effects: In this solution, upon identifying initial potential customers, the system checks whether the customer has commented on short videos or sent private messages. If so, the system obtains the customer's basic personal information; otherwise, it does not. This significantly protects customer privacy and enhances the protection of customer information. Furthermore, after obtaining personal purchase data, the system assesses its validity, further confirming the data's accuracy and providing more accurate data for identifying subsequent potential customers, thus avoiding ineffective judgments.

[0019] Preferably, as an improvement, the method further includes the following steps:

[0020] S7. Based on the generated ultimate potential data, count the number of potential customers corresponding to each content type and calculate the weight ratio corresponding to each content type.

[0021] S8. Determine whether the weight ratio corresponding to each content type is greater than or equal to the preset weight value. If so, determine that the content type is a key promotion type; otherwise, the content type is not a key promotion type.

[0022] S9. When it is determined that a certain content type is a key promotion type, the promotion focus corresponding to the content type is analyzed based on the basic personal information of the customer corresponding to the content type, and the feedback is given to the short video creator corresponding to the content type.

[0023] Beneficial effects: In this solution, the number of potential customers in the ultimate potential data is statistically analyzed to determine the weight ratio of the number of customers corresponding to each content type. Once the weight ratio of the number of customers for a certain content type exceeds the preset weight value, it is determined that the content type is a key promotion type, that is, the content type is more popular with customers. Then, based on the basic personal information of the customers corresponding to the content type, the promotion focus of the content type is analyzed to better provide feedback to the creators, so that the short videos produced by the creators are closer to the customers' preferences, and greatly improve the conversion rate of products in the short videos.

[0024] Preferably, as an improvement, S6 includes the following steps:

[0025] S60. Obtain historical customer data and other customer information from the database, wherein the historical customer data includes the customer's historical personal identity information, personal purchase data information, and browsing history information;

[0026] S61. Based on the neural network algorithm, construct a potential customer identification model, select several pieces of information from other customer information, input them into the potential customer identification model synchronously with historical customer data, and output the corresponding identification results;

[0027] S62. Based on the output recognition results, calculate the accuracy of each recognition result, select the recognition result with the highest accuracy, and determine the input data type corresponding to other customer information in the potential customer recognition model based on the selected recognition result.

[0028] S63. Based on the input data type of other customer information in the determined potential customer identification model, match the other customer information corresponding to the current customer with the same input data type from the database;

[0029] Based on the matched customer information, as well as the customer's personal identity information, personal purchase data, and browsing history, the system identifies the content type corresponding to the ultimate potential customer based on the potential customer identification model, and generates the corresponding ultimate potential data.

[0030] Beneficial effects: In this solution, when constructing the potential customer identification model, it is considered that the more data input, the slower the processing efficiency. In order to ensure the accuracy as much as possible with limited data, other customer information is also introduced. However, in actual use, not all of the other customer information will be input, as this would inevitably increase the processing burden of the system. Therefore, some types of this other customer information will be selected, and the most representative types will be selected, so as to achieve the greatest possible accuracy with limited data. Attached Figure Description

[0031] Figure 1 This is a flowchart of the potential customer mining method based on short video data in Embodiment 1 of the present invention. Detailed Implementation

[0032] The following detailed description illustrates the specific implementation method:

[0033] The basic implementation examples are as follows: Figure 1 As shown: A method for potential customer mining based on short video data, including the following steps:

[0034] S1. After reaching the preset time period, retrieve the account information corresponding to each short video viewed in the previous time period from the database.

[0035] S2. Based on the corresponding account information, obtain the customer personal information corresponding to each account information, wherein the customer personal information includes the short videos viewed and the viewing duration of the corresponding short videos;

[0036] S3. Based on the short videos viewed by each customer and the corresponding viewing time, identify the content type of the viewed short videos, and based on the content type, count the viewing time of each customer for each content type and generate corresponding viewing record information.

[0037] S4. Based on the browsing history information, match the content type whose browsing time is ranked higher than the preset ranking for the customer, and identify the customer as the primary potential customer corresponding to the matched content type.

[0038] S5. Based on the customer account information identified as a primary potential customer, obtain the customer's basic personal information, which includes personal identity information and personal purchase data information; the personal identity information includes the customer's name, age, date of birth, etc.

[0039] S5 includes the following steps:

[0040] S50. Based on the customer account information identified as a primary potential customer, determine whether the customer has commented or sent a private message in the short video. If so, obtain the basic personal information corresponding to the customer account information. The basic personal information includes personal identity information and personal purchase data information. If not, determine that the customer is not a primary potential customer. In this embodiment, the basic personal information of the primary potential customer is not directly obtained. Instead, the basic personal information of the customer is obtained based on the customer's motivation to take an action, such as commenting or sending a private message. This greatly improves the confidentiality of customer information and better protects customer privacy.

[0041] S51. After obtaining the customer's personal purchase data, analyze the personal purchase data to determine whether the customer has purchased the product corresponding to the matched content type within a preset time period. If so, the personal purchase data is determined to be invalid data; otherwise, the personal purchase data is determined to be valid data.

[0042] S6. Based on the customer's personal identity information, personal purchase data information, and browsing history information, and using the potential customer identification model, identify the content type corresponding to the ultimate potential customer and generate the corresponding ultimate potential data.

[0043] S6 includes the following steps:

[0044] S60. Obtain historical customer data and other customer information from the database, wherein the historical customer data includes the customer's historical personal identity information, personal purchase data information, and browsing history information;

[0045] S61. Based on the neural network algorithm, construct a potential customer identification model, select several pieces of information from other customer information, input them into the potential customer identification model synchronously with historical customer data, and output the corresponding identification results;

[0046] S62. Based on the output recognition results, calculate the accuracy of each recognition result, select the recognition result with the highest accuracy, and determine the input data type corresponding to other customer information in the potential customer recognition model based on the selected recognition result.

[0047] S63. Based on the input data type of other customer information in the determined potential customer identification model, match the other customer information corresponding to the current customer with the same input data type from the database;

[0048] Based on the matched customer information, as well as the obtained customer's personal identity information, personal purchase data, and browsing history, the system uses a potential customer identification model to identify the content type corresponding to the ultimate potential customer and generates the corresponding ultimate potential data. In this embodiment, the potential customer identification model is a BP neural network model. Specifically, a three-layer BP neural network model is first constructed, including an input layer, a hidden layer, and an output layer. In this embodiment, the customer's historical personal identity information, personal purchase data, browsing history, and the information of the input data type corresponding to the identified customer's other information are used as the input to the input layer. The preset number is 4, so the input layer has 7 nodes. The output is whether the customer is a potential customer of the corresponding content type; therefore, it uses 1 node. For the hidden layer, this embodiment uses the following formula to determine the number of hidden layer nodes: Where l is the number of nodes in the hidden layer, n is the number of nodes in the input layer, m is the number of nodes in the output layer, and a is a number between 1 and 10, which is taken as 6 in this embodiment. Therefore, the hidden layer has a total of 8 nodes. Backpropagation (BP) neural networks typically use the sigmoid differentiable function and linear functions as the network's activation functions. This paper selects the sigmoid tangent function (tansig) as the activation function for the hidden layer neurons. The prediction model selects the sigmoid logarithmic function (tansig) as the activation function for the output layer neurons.

[0049] It also includes the following steps:

[0050] S7. Based on the generated ultimate potential data, count the number of potential customers corresponding to each content type and calculate the weight ratio corresponding to each content type.

[0051] S8. Determine whether the weight ratio corresponding to each content type is greater than or equal to the preset weight value. If so, determine that the content type is a key promotion type; otherwise, the content type is not a key promotion type.

[0052] S9. When a content type is determined to be a key promotional type, the promotional focus corresponding to that content type is analyzed based on the customer's basic personal information, and this information is fed back to the short video creator corresponding to that content type. In this embodiment, for example, if the weight ratio of a certain content type is lower than a preset weight value, it indicates that the content type is not a key promotional type, thus greatly improving the accuracy of promotion.

[0053] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for identifying potential customers based on short video data, characterized by: Includes the following steps: S1. After reaching the preset time period, retrieve the account information corresponding to each short video viewed in the previous time period from the database. S2. Based on the corresponding account information, obtain the customer personal information corresponding to each account information, wherein the customer personal information includes the short videos viewed and the viewing duration of the corresponding short videos; S3. Based on the short videos viewed by each customer and the corresponding viewing time, identify the content type of the viewed short videos, and based on the content type, count the viewing time of each customer for each content type and generate corresponding viewing record information. S4. Based on the browsing history information, match the content type whose browsing time is ranked higher than the preset ranking for the customer, and identify the customer as the primary potential customer corresponding to the matched content type. S5. Based on the customer account information identified as a primary potential customer, obtain the customer's basic personal information, which includes personal identity information and personal purchase data information; S6. Based on the customer's personal identity information, personal purchase data information, and browsing history information, and using the potential customer identification model, identify the content type corresponding to the ultimate potential customer and generate the corresponding ultimate potential data. S6 includes the following steps: S60. Obtain historical customer data and other customer information from the database, wherein the historical customer data includes the customer's historical personal identity information, personal purchase data information, and browsing history information; S61. Based on the neural network algorithm, construct a potential customer identification model, select several pieces of information from other customer information, input them into the potential customer identification model synchronously with historical customer data, and output the corresponding identification results; S62. Based on the output recognition results, calculate the accuracy of each recognition result, select the recognition result with the highest accuracy, and determine the input data type corresponding to other customer information in the potential customer recognition model based on the selected recognition result. S63. Based on the input data type of other customer information in the determined potential customer identification model, match the other customer information corresponding to the current customer with the same input data type from the database; Based on the matched customer information, as well as the customer's personal identity information, personal purchase data, and browsing history, the system identifies the content type corresponding to the ultimate potential customer based on the potential customer identification model, and generates the corresponding ultimate potential data.

2. The method for potential customer mining based on short video data according to claim 1, characterized in that: S5 includes the following steps: S50. Based on the account information of customers identified as primary potential customers, determine whether the customer has commented or sent private messages in the short video. If so, obtain the basic personal information corresponding to the customer's account information. The basic personal information includes personal identity information and personal purchase data information. If not, determine that the customer is not a primary potential customer. S51. After obtaining the customer's personal purchase data, analyze the personal purchase data to determine whether the customer has purchased the product corresponding to the matched content type within a preset time period. If so, the personal purchase data is determined to be invalid data; otherwise, the personal purchase data is determined to be valid data.

3. The potential customer mining method based on short video data according to claim 2, characterized in that: It also includes the following steps: S7. Based on the generated ultimate potential data, count the number of potential customers corresponding to each content type and calculate the weight ratio corresponding to each content type. S8. Determine whether the weight ratio corresponding to each content type is greater than or equal to the preset weight value. If so, determine that the content type is a key promotion type; otherwise, the content type is not a key promotion type. S9. When it is determined that a certain content type is a key promotion type, the promotion focus corresponding to the content type is analyzed based on the basic personal information of the customer corresponding to the content type, and the feedback is given to the short video creator corresponding to the content type.

Images