Crowd value research and judgment method and device based on equipment information and order information

Abstract

The invention discloses a crowd value research and judgment method and device based on equipment information and order information, and relates to the technical field of data processing, and the method comprises the steps: obtaining the order data of each user and the equipment information of equipment used by the user to purchase a commodity; performing adaptive RFM analysis on the order data of each user to obtain a value layering result of the user; according to the equipment information of each user and the order data, analyzing an association rule between equipment characteristics and commodity categories, and further estimating commodity category preferences of the users; according to a preset device price grade division rule, determining a device price grade of a device used by each user to purchase a commodity, and further estimating a consumption ability size and a commodity category grade preference of the user; and finally, according to the value layering result, the commodity category preference, the consumption ability size and the commodity category grade preference of each user, determining a value three-dimensional portrait of the user. According to the method, the value stereo portrait of the user is constructed from multiple dimensions, and the accuracy of user value research and judgment is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a people value research and judgment method and device based on device information and order information. BACKGROUND

[0002] In the existing user analysis system, user stratification is usually based on order data (such as consumption amount and purchase frequency), and there is a lack of deep mining of device information. For example, the RFM (recent purchase time, purchase frequency, and total purchase amount) model only relies on transaction data and cannot reflect the influence of user device attributes on consumption behavior, which leads to incomplete user portraits and difficulty in accurately identifying high-value people. In addition, the relationship between device information (such as device model, screen size, and network environment) and consumption behavior has not been systematically analyzed, limiting the optimization of personalized recommendations and marketing strategies.

[0003] Currently, the research and judgment of high-value users in the industry mainly relies on statistical analysis based on traditional RFM models. By calculating the user's last consumption time, consumption frequency, and consumption amount, and then setting thresholds or using simple binning methods, users are divided into "important value users", "general retention users", and other groups. The analysis results mainly revolve around transaction data itself, such as: which users have high total consumption, and which users have frequent repurchases. The existing research and judgment methods have single data dimensions and one-sided user portraits, resulting in inaccurate research and judgment results. SUMMARY

[0004] The purpose of the present application is to provide a people value research and judgment method and device based on device information and order information, which can construct a three-dimensional portrait of user value from multiple dimensions, and accurately research and judge user value.

[0005] To achieve the above purpose, the present application provides the following solutions: In a first aspect, the present application provides a people value research and judgment method based on device information and order information, comprising: obtaining order data of each user and device information of devices used by the user to purchase goods within a preset time period; performing adaptive RFM analysis on the order data of each user to obtain the value stratification result of each user; analyzing the association rules between device characteristics and product categories according to the device information and order data of each user; estimating the product category preference of each user according to the association rules between device characteristics and product categories and the device characteristics of the devices used by the user to purchase goods; determining the device price level of the devices used by each user to purchase goods according to a preset device price level division rule, and determining the consumption ability and product category level preference of each user according to the device price level. The value stereoscopic portrait of each user is determined according to the value stratification result of each user, the commodity category preference, the consumption capacity size and the commodity category grade preference.

[0006] In a second aspect, the present application provides a device for judging the value of a group of people based on equipment information and order information, comprising: An information acquisition module is configured to acquire order data of each user and equipment information of equipment used by the user to purchase commodities in a preset time period. An RFM analysis module is configured to perform adaptive RFM analysis on the order data of each user to obtain a value stratification result of each user. An association rule analysis module is configured to analyze an association rule between equipment characteristics and commodity categories according to the equipment information and the order data of each user. A commodity category preference estimation module is configured to estimate the commodity category preference of each user according to the association rule between the equipment characteristics and the commodity categories and the equipment characteristics of the equipment used by the user to purchase commodities. A consumption capacity estimation module is configured to determine the equipment price grade of the equipment used by each user to purchase commodities according to a preset equipment price grade division rule, and determine the consumption capacity size and the commodity category grade preference of each user according to the equipment price grade. A user value stereoscopic portrait construction module is configured to determine the value stereoscopic portrait of each user according to the value stratification result of each user, the commodity category preference, the consumption capacity size and the commodity category grade preference.

[0007] In a third aspect, the present application provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned method for judging the value of a group of people based on equipment information and order information.

[0008] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executable by a processor to implement the above-mentioned method for judging the value of a group of people based on equipment information and order information.

[0009] According to the specific embodiments provided by the present application, the following technical effects are disclosed: This application provides a method and apparatus for assessing user value based on device and order information. The method involves acquiring order data and device information of the devices used by users to purchase goods; performing adaptive RFM analysis on the order data to derive user value stratification results; analyzing the association rules between device characteristics and product categories based on the device information and order data to estimate user product category preferences; determining the device price tier for each user's purchased goods based on preset device price tier classification rules to estimate user spending power and product category tier preferences; and finally, determining a comprehensive user value profile based on each user's value stratification results, product category preferences, spending power, and product category tier preferences. This application constructs a comprehensive user value profile from multiple dimensions, improving the accuracy of user value assessment. Furthermore, the use of adaptive RFM analysis enhances the accuracy of RFM analysis, thereby improving the accuracy of user value assessment. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is an application environment diagram of a method for assessing the value of a group of people based on device information and order information, as described in one embodiment of this application. Figure 2 A flowchart illustrating a method for assessing the value of a group of people based on device information and order information, provided as an embodiment of this application; Figure 3 A schematic diagram illustrating the technical concept of a method for assessing the value of a group of people based on device information and order information, provided in an embodiment of this application; Figure 4 A schematic flowchart of adaptive RFM analysis provided in an embodiment of this application; Figure 5 A schematic diagram illustrating the correlation analysis process between equipment characteristics and product categories provided in one embodiment of this application; Figure 6 A flowchart illustrating the analysis of equipment price levels and consumer behavior provided in an embodiment of this application; Figure 7 A schematic diagram of the functional modules of a crowd value assessment device based on device information and order information provided in an embodiment of this application; Figure 8This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] The purpose of this invention is to provide a method and apparatus for assessing user value based on device and order information, solving the problems of single-dimensional user value analysis, reliance on human experience, and low identification accuracy in existing technologies. By integrating multi-source data and applying machine learning algorithms, it achieves automated, accurate, and intelligent identification of high-value users, providing data support for enterprise decision-making.

[0014] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0015] The audience value assessment method based on device information and order information provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, the terminal communicates with the server via a network. The data storage system stores the data the server needs to process. The data storage system can be set up independently, integrated into the server, or placed in the cloud or on another server. The terminal can send each user's order data and the device information of the equipment used by the user to purchase goods to the server. After receiving the order data and device information of each user within a preset time period, the server performs adaptive RFM analysis on each user's order data to obtain the value stratification result for each user; analyzes the association rules between device characteristics and product categories based on each user's device information and order data; estimates each user's product category preference based on the association rules between device characteristics and product categories and the device characteristics of the equipment used by the user to purchase goods; determines the device price level of each user's purchased goods based on preset device price level classification rules; determines each user's purchasing power and product category preference based on the device price level; and determines a comprehensive value profile of each user based on the value stratification result, product category preference, purchasing power, and product category preference. The server can then feed back the comprehensive value profile of the user to the terminal. Furthermore, in some embodiments, the audience value assessment method based on device information and order information can also be implemented by the server or the terminal alone. For example, the terminal can directly perform audience value assessment based on the order data of each user and the device information of the device used by the user to purchase the goods. Alternatively, the server can obtain the order data of each user and the device information of the device used by the user to purchase the goods from the data storage system and perform audience value assessment based on the device information and order information.

[0016] The terminal can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. The server can be implemented using a standalone server, a server cluster consisting of multiple servers, or a cloud server.

[0017] In one exemplary embodiment, such as Figure 2 and Figure 3 As shown, a method for assessing the value of a customer base based on device information and order information is provided. This method is executed by a computer device, specifically a terminal or server, or both. In this embodiment, the method is applied to... Figure 1 The following steps, 101 to 106, are used as an example to illustrate the process of using a server in the example.

[0018] Step 101: Obtain order data for each user within a preset time period and device information for the devices used by the users to purchase goods.

[0019] Step 102: Perform adaptive RFM analysis on the order data of each user to obtain the value stratification results for each user.

[0020] Step 103: Analyze the association rules between device characteristics and product categories based on each user's device information and order data.

[0021] Step 104: Estimate each user's product category preference based on the association rules between device characteristics and product categories and the device characteristics of the device used by the user to purchase the product.

[0022] Step 105: Determine the price level of the equipment used by each user to purchase goods according to the preset equipment price level classification rules, and determine the user's spending power and product category preference based on the equipment price level.

[0023] Step 106: Determine the three-dimensional value profile of each user based on the value stratification results, product category preferences, spending power, and product category level preferences.

[0024] By implementing steps 101 to 106 above, this application obtains order data for each user and equipment information of the devices used by the users to purchase goods; it performs adaptive RFM analysis on the order data of each user to obtain the user's value stratification results; it analyzes the association rules between device characteristics and product categories based on the device information and order data of each user, and then estimates the user's product category preference based on the association rules; it also determines the device price level of the equipment used by each user to purchase goods based on the preset device price level classification rules, and then estimates the user's spending power and product category level preference; finally, it filters high-value users based on the value stratification results, product category preference, spending power, and product category level preference of each user. This application adopts adaptive RFM analysis, overcoming the problems of static rigidity and lack of adaptability of statistical analysis methods based on traditional RFM models; this application also estimates the user's product category preference by analyzing the association rules between device characteristics and product categories based on the device information and order data of each user, which has limited accuracy and cannot explore deep associations in existing methods; this application also uses device information to construct a three-dimensional user profile, overcoming the problem of superficial use of device information and insufficient value mining in existing methods. This application constructs a three-dimensional user profile from multiple dimensions to improve the accuracy of identifying high-value users.

[0025] In another exemplary embodiment of this application, in step 102, adaptive RFM analysis is performed on the order data of each user to obtain the value stratification result for each user, such as... Figure 4 As shown, it specifically includes: (1) For each user, calculate the R value, F value and M value of each user based on the order data within the user's preset time period.

[0026] The user order data includes fields such as order ID, user ID, store type, order status (completed, valid, etc.), payment confirmation time, order start time, order end time, order cancellation time, province, city, district, parent order ID, payment type, order price, store ID, total order amount, order sales amount, order payment amount, product ID, product quantity, product price, category ID, creation time, membership card activation time, membership level, whether to cancel membership, membership level start time, membership level end time (if canceled, it is the cancellation time), etc.

[0027] Calculate R (number of days since last purchase), F (total number of purchases), and M (total spending amount) for each user. Then, use the K-means clustering algorithm on each of the three dimensions of R, F, and M to divide the users in each dimension into 3-5 groups.

[0028] (2) Cluster the R values ​​of each user to obtain multiple R clusters, and determine the R value score for each R cluster based on the user's R value.

[0029] Let's take the R dimension (recent purchase time) as an example for a detailed explanation: Suppose there are 8 users, and their original R values ​​(in days) are shown in Table 1 below. The smaller the R value, the more recent the purchase time, and theoretically the higher the value.

[0030] Table 1. Original R values ​​of users

[0031] The clustering process of the K-means clustering algorithm is as follows: Setting parameters: Set the number of clusters K=3, which means that we want to automatically divide users into 3 groups according to their R value.

[0032] The K-means algorithm performs the following operations: 1) Randomly initialize 3 center points (e.g., around R values ​​of 2, 60, and 120).

[0033] 2) Assign all users to the nearest center point to form 3 temporary groups.

[0034] 3) Recalculate the center point of each group (take the average value).

[0035] 4) Repeat steps 2) and 3) until the center point no longer changes significantly, at which point the algorithm converges.

[0036] Ultimately, the algorithm will find three optimal center points to distinguish users, assuming the final center points are 5 days, 20 days, and 90 days.

[0037] Users are assigned to three different groups based on their distance from the central point, as shown in Table 2.

[0038] Table 2 Clustering and Binning Results

[0039] Next, these cluster bins (clusters) need to be assigned a value score R_label. The core principle is: the smaller the R value (the more recent the purchase time), the higher the value score. As shown in Table 3.

[0040] Table 3 Value scores assigned to cluster binning

[0041] For the clustering process of F and M dimensions, please refer to the R clustering process described above.

[0042] (3) Cluster the F values ​​of each user to obtain multiple F clusters, and determine the F value score for each F cluster based on the user's F value.

[0043] (4) Cluster the M values ​​of each user to obtain multiple M clusters, and determine the M value score for each M cluster based on the user's M value.

[0044] (5) Based on the data distribution results after R-value clustering, F-value clustering and M-value clustering, the entropy weight method is applied to determine the weights (W_R, W_F, W_m) of R-value, F-value and M-value.

[0045] (6) Calculate the RFM comprehensive score for each user based on the R value score, F value score, M value score and the weights of the R value, F value and M value.

[0046] Calculate the overall RFM score for each user:

[0047] Assume that the user's R-value score (R_label), F-value score (F_label), and M-value score (M_label) have been obtained through clustering binning and entropy weighting, as shown in Table 4.

[0048] Table 4. R-value score, F-value score, and M-value score for each user

[0049] Here, we assume the calculated weights are: w_R=0.5; w_F=0.3; w_M=0.2; the total weight is 1.

[0050] The RFM_Score calculation formula described above is used to calculate the overall RFM score for each user.

[0051] User A: ; User B: ; User C: ; User D: ; User E: ; User F: .

[0052] The final results are summarized in Table 5.

[0053] Table 5. Overall RFM Score for Each User

[0054] The final RFM_Score is the final value quantification score; the higher the score, the higher the user value.

[0055] (7) Cluster the RFM comprehensive scores of each user to obtain RFM clusters; the user value of different RFM clusters is different.

[0056] (8) Determine the value stratification results of each user based on the RFM cluster to which each user belongs.

[0057] K-means clustering is applied again to the RFM_Score of all users to automatically divide users into 4 core value groups (e.g., important value, important development, important maintenance, and general users).

[0058] As an example, a final clustering (e.g., K=4) is performed based on the RFM composite scores in Table 5 to automatically complete the final user segmentation: Cluster Center 2.0: High-Value Users (User A); Cluster Center 1.5: Key Development Users (User B, User C); Cluster Center 1.0: Important user retention (User D); Cluster center 0.5: General value users (user E, user F).

[0059] The entire process of adaptive RFM analysis in this application, from cluster binning to entropy weighting and finally cluster stratification, is entirely data-driven and requires no manual threshold setting. Compared with statistical analysis methods based on traditional RFM models, this enables the model to adapt to different business scenarios and data distributions and intelligently identify truly high-value individuals.

[0060] In another exemplary embodiment of this application, in step 103, the association rules between device characteristics and product categories are analyzed based on each user's device information and order data, such as... Figure 5 As shown, it specifically includes: (1) Create an item transaction set for each user based on the device information and order data of each user; the item transaction set includes the device characteristics of the device used by the user to purchase the goods and the categories of goods purchased by the user.

[0061] User device characteristics include screen size, pixel ratio, smartphone, 4K display, device price range, network environment, etc. Continuous device characteristics (such as screen size) need to be converted into categorical variables (e.g., large, medium, small screen).

[0062] Example of an item transaction set: {User A:[dev_large screen, dev_high pixel, cat_smartphone, cat_4k monitor]}. When creating tags for device characteristics, an explicit prefix is ​​added, such as "dev" (abbreviation for device). Similarly, the prefix "cat_" (abbreviation for category) is added for product categories. In subsequent processing, whether an item is device data can be determined by whether it begins with "dev". Table 6 shows the user's item transaction set data.

[0063] Table 6 User's Item Transaction Set Data

[0064] (2) For each user, multiple item combinations are determined based on the item transaction set; the item combination includes any equipment feature and any product category.

[0065] To reduce computational load, after obtaining the item transaction set, the FP-Growth algorithm is used to quickly scan all users' item transaction sets to find frequently occurring "device characteristic-product category" combinations. The FP-Growth algorithm's workflow is as follows: Step 1: Constructing an FP-tree (Frequent Pattern Tree): The algorithm scans the database only twice. The first scan identifies the frequency (support) of all individual products and discards infrequent products. The second scan sorts the remaining products in each transaction record by frequency from high to low, and then inserts these products into the FP-tree like adding paths to a tree structure. Products with the same prefix share the same path, which greatly compresses data storage. Step 2: Mining Frequent Itemsets from the FP-tree: For each product, a conditional pattern base (i.e., all prefix paths containing that product) is generated from the FP-tree. Based on these conditional pattern bases, a smaller "conditional FP-tree" is constructed. Recursively mining on this conditional FP-tree easily finds all frequent itemsets containing that product, i.e., frequently occurring "device characteristic-product category" combinations. For each frequently occurring combination of items, the support and confidence scores are further calculated.

[0066] (3) Calculate the support and confidence of each item combination.

[0067] The formula for calculating support is as follows:

[0068] In the formula, Represents a combination of items Support level; This indicates a characteristic of a device; Indicates a product category; This represents the number of trading users who simultaneously include both X and Y; N represents the total number of trading users. Higher support indicates that this combination is more common; if support is low, even a strong rule may not be very valuable.

[0069] The confidence level is calculated using the following formula:

[0070] In the formula, Represents a combination of items Confidence level; This indicates a characteristic of a device; Indicates a product category; This indicates the number of users who include both X and Y. This represents the number of users who have X. The higher the confidence level, the greater the probability that Y will occur when X occurs, and the more reliable the rule is.

[0071] As an example, ① calculate the support of the item combination: {dev_large screen} -> {cat_dress}.

[0072] First, to identify transactions that occur simultaneously, we need to find transactions that simultaneously contain both dev_bigscreen and cat_dress. We check the item transaction set for each user: User 1: contains; User 2: contains; User 3: does not contain; User 4: does not contain; User 5: contains.

[0073] Next, calculate the numerator count(X∪Y) and the denominator N. Transactions involving both X and Y are: User 1, User 2, and User 5, so count(X∪Y) = 3. The total number of transactions in Table 6 is for 5 users, so N = 5.

[0074] Finally, the support score is calculated. =3 / 5=0.6.

[0075] Interpretation of results: Among all users, 60% purchased dresses while using large-screen devices, indicating that this item combination is a fairly common phenomenon.

[0076] ② Calculate the confidence level of the item combination: {dev_large screen} -> {cat_dress} First, calculate count(X) to find all transactions that contain dev_large screen. Since user 1, user 2, user 4, and user 5 all contain dev_large screen, count(X) = 4.

[0077] Secondly, calculate count(X∪Y). Based on the aforementioned support calculation, we know that count(X∪Y) = 3.

[0078] Finally, the confidence level was calculated: Confidence = 3 / 4 = 0.75.

[0079] Interpretation of results: Among users using large screens, 75% purchased dresses, indicating a strong correlation between these items.

[0080] (4) Compare the support and confidence of each item combination with their respective thresholds to determine the item combinations with a relationship, i.e. the association rules between equipment characteristics and product categories.

[0081] If the support of an item combination is greater than or equal to the corresponding threshold, and the confidence is also greater than or equal to the corresponding threshold, then the current item combination is an item combination with a relationship; otherwise, it is an invalid rule and should be discarded.

[0082] A minimum threshold is typically set to filter meaningful rules. The minimum support threshold, Min_support=0.1, ensures that the support is at least 10% to guarantee that the rules are sufficiently representative; the minimum confidence threshold, Min_confidence=0.6, ensures that the confidence is at least 60% to guarantee that the rules are reliable.

[0083] In the example above, the support is 0.6, which is greater than 0.1, and the confidence is 0.75, which is greater than 0.6. Therefore, the conclusion is that this is a high-quality rule and should be adopted.

[0084] In another exemplary embodiment of this application, in step 105, a mapping table is established to map device models to three tiers: "high-end," "mid-range," and "low-end," based on the device's retail price and retail time (considering time depreciation). This table needs to be updated periodically. Specifically, in step 105, the device price tier for each user's purchased goods is determined according to a preset device price tier classification rule, including: (1) Determine the initial price range based on the sales price of the equipment used by the user to purchase the goods.

[0085] As an example, we first divide the devices into initial price tiers based on their official retail prices, as shown in Table 7.

[0086] Table 7 Preliminary Tier Classification Based on Official Retail Price of Equipment

[0087] (2) Determine whether the sales duration of the device used by the user to purchase the product is greater than the preset duration.

[0088] If not, the initial price tier will be the final price tier.

[0089] If so, the degraded price of the device used by the user to purchase the product will be determined based on the duration of the sale, and the final price tier will be determined based on the degraded price.

[0090] The time decay adjustment rule adopts a segmented decay model for the value of equipment over time. For example, the value of equipment in the current year does not decay; after 1 year, the value decays by 30%; after 2 years, the value decays by 50%; after 3 years, the value decays by 70%; and after 3 years or more, the value decays by 80%.

[0091] The grade is re-determined based on the current value after decay, as shown in Table 8.

[0092] Table 8 Final Grade Determination Rules

[0093] Specific example: Determining the iPhone 17's class Let's analyze several scenarios to determine the iPhone 17's perceived value: Scenario 1: iPhone 17 is released in 2025. Price at launch: Assuming a price of 8999 yuan, release date: 2025, current date: 2025, time decay: 0 years → no decay, current value: 8999 yuan. Price tier: Based on launch price of 8999 yuan. 6000 yuan → High-end, valued at 8999 yuan after depreciation 4000 yuan → High-end. The final level is high-end.

[0094] Scenario 2: One year after the release of the iPhone 17 (2026) Release Price: 8999 RMB, Release Date: 2025, Current Date: 2026, Time Decrease: 1 year → 30% decrease, Current Value: Yuan. Grade assessment: Value after depreciation is 6299.3 yuan. 4000 yuan → High-end. The final level is high-end.

[0095] Scenario 3: 3 years after the release of the iPhone 17 (2026) Release Price: 8999 RMB, Release Date: 2025, Current Date: 2028, Time Decrease: 3 years → 70% decrease, Current Value: Yuan. Tier determination: After attenuation, the value is 1500 < 2699.7 yuan < 3999 yuan → mid-range. The final tier is mid-range.

[0096] Table 9 below shows an example of the constructed equipment price tiers.

[0097] Table 9 Equipment Price Tiers

[0098] In step 105, determining each user's spending power and product category preferences based on equipment price tiers requires first verifying whether there are differences in spending power among users at different equipment price tiers. For example... Figure 6 As shown.

[0099] (1) Data Association: Users' device models are associated with their purchase order data, thus assigning each user a "device price range" label. Specific steps: Step 1): Match the price range to the equipment model. Based on the established equipment price tier table, find the corresponding high-end, mid-range, or low-end label for each equipment model in the equipment information table.

[0100] Step 2): Link consumption data by user ID Using the user ID as the unique connection key, a database table connection is established between the device information table (which now includes grade tags) and the order information table.

[0101] Operation: The database searches for records with the same user ID in two tables and combines them into one record. This is an exact match operation.

[0102] Result: A temporary table is generated, where each record clearly indicates: a specific user ID, a specific level of device used, and a specific consumption record generated. Step 3): Data aggregation – forming the final analysis unit Since a user may have multiple orders, it is necessary to aggregate the records based on orders into records based on users.

[0103] Operation: Use the SQL group by statement to group by user ID and equipment price tier, and then summarize the orders for each user.

[0104] Results: The final analysis wide table was obtained, with each user occupying one row, and including their device tier label and summarized consumption behavior indicators.

[0105] (2) Consumption Capacity Analysis: Using analysis of variance (ANOYA), we examined whether there were differences in average order amount, total consumption amount, and other indicators among user groups at different equipment price levels. Detailed analysis steps: Step 1): Formulate a hypothesis Null hypothesis: The mean order amount is exactly the same for users of high-end, mid-range, and low-end devices.

[0106] :

[0107] Alternative hypothesis: At least one group has a different mean of orders from the other groups.

[0108] Step 2): Data Preparation and Grouping: Extract the device price tier and average order amount for each user from the correlated data. Divide the data into three independent groups: Group 1 (High-end): A list of average order amounts for all users with high-end devices.

[0109] Group 2 (Mid-range): A list of average order amounts for all users with mid-range devices.

[0110] Group 3 (Low-end): A list of average order amounts for all users with low-end devices.

[0111] Step 3): Calculate key variance: The core is to calculate and compare the between-group variance and the within-group variance.

[0112] Within-group variance: Measures the degree of variation within a group. It calculates the difference between the order amount of each user within each group and the group average. This represents random error; a small within-group variance indicates that the spending power of group members is very similar.

[0113] Between-group variance: measures the degree of variation between different groups. It calculates the difference between the mean of each group and the overall mean of all data. This represents the treatment effect (i.e., the impact of the equipment price range). A large between-group variance indicates a significant difference in consumer spending power across different equipment price ranges, suggesting that, generally, consumers with higher-priced equipment also have greater spending power.

[0114] Secondly, such as Figure 6 As shown, this application also conducted a preference analysis of different product categories for different equipment price ranges: a chi-square test was used to analyze whether there was a significant correlation between equipment price ranges and preferred product category ranges. Detailed analysis steps: Step 1): Formulate a hypothesis Null hypothesis: The price range of equipment and the category of goods are independent of each other (unrelated).

[0115] Alternative hypothesis: The price range of equipment is not independent of the price range of the product category (there is a significant correlation).

[0116] Step 2): Construct a contingency table Generate a table in the following format to show the observed frequency distribution: Table 10. Product Categories Purchased at Different Equipment Price Levels

[0117] Step 3): Calculate the expected frequency. If the null hypothesis holds (neither of the two is relevant), then the theoretical frequency of each cell should be:

[0118] For example: The expected frequency of purchasing high-end equipment and high-end digital products is:

[0119] This means that if the price range and product category of the device were truly unrelated, one should expect to see approximately 76 high-end users purchasing premium digital products, but in reality, 150 were observed.

[0120] Step 4): Calculate the chi-square statistic. For each cell in the contingency table, calculate:

[0121] in, It is the frequency of observations. This is the expected frequency. This formula measures the overall deviation between the observed values ​​and the expected values. If the null hypothesis is true, the observed values ​​should be close to the expected values, and the chi-square value will be small. If there is a correlation between the two, the observed values ​​will deviate significantly from the expected values, and the chi-square value will be large.

[0122] Step 5): Decision The calculated chi-square value is compared with the critical value obtained by looking up the table based on the degrees of freedom.

[0123] Degrees of freedom = (number of rows - 1) (Number of columns - 1) = (3 - 1) (3-1)=4 If the calculated chi-square value is greater than the critical value, then the null hypothesis is rejected.

[0124] Conclusion: Statistical evidence shows a significant correlation between equipment price range and product category preference. For example, those with higher-priced equipment tend to prefer higher-end product categories. Similarly, people with higher-priced equipment are more likely to purchase luxury goods.

[0125] In an exemplary embodiment, in step 106, a comprehensive value profile of each user is determined based on their value stratification results, product category preferences, spending power, and product category level preferences. An example scenario will be used to illustrate this: Scenario: A user is found to be an iPhone 15 Pro Max (high-end device) user.

[0126] RFM analysis output: This user has been marked as a high-value user.

[0127] Equipment price tier analysis output: It confirmed that they are high-end equipment users and verified that these users have a high average order value and prefer higher-end products.

[0128] Device characteristic analysis output: Provides association rules: {high-pixel device}->{photographic equipment}.

[0129] Conclusion: This is no longer a simple high-spending user, but a top-value user who uses high-end, high-pixel equipment, has a strong interest in photography, has extremely strong purchasing power, and prefers higher-end products.

[0130] This application, through the deep integration of equipment information and order information, and the intelligent processing of data, can significantly improve the accuracy of high-value user identification, reduce missed judgments (missing potential high-value users) and misjudgments (over-investing resources in low-value users), directly improve the marketing return on investment, reduce the long-term maintenance cost of the system, and maintain continuous accuracy.

[0131] Based on the above comprehensive value profile, precise strategies can be formulated: Core strategy: Push high-end photography equipment to this user.

[0132] Strategy Enhancement 1: Because he is a high-value user, he should not be pushed discounted products, but new products, flagship models, and limited editions.

[0133] Strategy Enhancement 2: Since he belongs to the high-end device user group, which has a high membership activation rate, the advertisement can highlight exclusive membership benefits and services to guide him to activate premium membership.

[0134] Based on the aforementioned three-dimensional value profile, dynamic management of the user lifecycle is implemented.

[0135] Combining the time dimension of RFM with other analyses enables dynamic operations.

[0136] Scenario: A former high-value user is using high-end devices, but their R value has deteriorated.

[0137] Traditional approach: classify them as lost users and may initiate a general recall process.

[0138] Key insight: He uses high-end equipment, and his potential spending power still exists.

[0139] Precise targeting: Based on device characteristics and category association (such as previous purchase of high-end headphones), push the latest similar or upgraded products.

[0140] Incentive strategy: Since he is a high-value user, he can be given one-on-one VIP exclusive return treatment instead of ordinary vouchers.

[0141] Therefore, based on the aforementioned comprehensive value profile, a targeted push strategy for this user can be derived. The final output is not a lengthy analysis report, but a strategy library and user profile library that can be directly invoked and executed by business systems. For example, a visual intelligent analysis report can be generated for operations personnel, and data can be sent to machine systems via structured data interfaces. The system will automatically generate an enhanced user profile table containing the following fields: User ID: 1001; RFM Tier: High-Value User; Device Price Range: High-End; Preferred Product Categories (from association rules): Digital Products, Luxury Goods; Potential Interest Categories: Photography Equipment, Smart Home; Operational Strategy Priority: P0; Recommended Product Categories: Latest Flagship Mobile Phones, Drones.

[0142] Ultimately, an automated marketing strategy rule base can be built, and the system will pre-configure the following based on the integrated insights: If the user equipment price tier = "high-end" AND RFM tier = "high-value user" then push the new product launch channel and assign VIP exclusive customer service.

[0143] If the user's device features include "large screen" AND the preferred category includes "clothing", then the homepage will prioritize displaying detailed videos and high-definition images of the "clothing" category.

[0144] If the user's equipment level = "low-end" AND RFM tier = "general user" BUT the network environment is usually "WiFi" THEN, we will try to push "high-value membership" to increase its value.

[0145] This application constructs a three-dimensional value profile of users, which can integrate applications to generate precise strategies, and system integration to achieve automated operation, ultimately forming a self-evolving intelligent judgment process for high-value groups.

[0146] This application also provides an application scenario in which the aforementioned method for assessing audience value based on device information and order information is applied. Specifically, the method for assessing audience value based on device information and order information provided in this embodiment can be applied to a personalized advertising push scenario. This scenario includes a user profile construction stage and a personalized information recommendation stage; the user profile construction stage is used to construct a three-dimensional value profile of the user based on the user's device information and order information; the personalized information recommendation stage is used to push relevant information in a targeted manner based on the constructed three-dimensional value profile of the user. The method for assessing audience value based on device information and order information provided in this embodiment belongs to the user profile construction stage.

[0147] Based on the same inventive concept, this application also provides a device for assessing the value of a group based on device information and order information, which implements the aforementioned method for assessing the value of a group based on device information and order information. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the device for assessing the value of a group based on device information and order information provided below can be found in the limitations of the method for assessing the value of a group based on device information and order information described above, and will not be repeated here.

[0148] In one exemplary embodiment, such as Figure 7 As shown, a device for assessing the value of a customer base based on device information and order information is provided, comprising: The information acquisition module M1 is used to acquire order data of each user and equipment information of the devices used by the user to purchase goods within a preset time period.

[0149] The RFM analysis module M2 is used to perform adaptive RFM analysis on the order data of each user to obtain the value stratification results for each user.

[0150] The association rule analysis module M3 is used to analyze the association rules between device characteristics and product categories based on each user's device information and order data.

[0151] The product category preference estimation module M4 is used to estimate each user's product category preference based on the association rules between device characteristics and product categories and the device characteristics of the device used by the user to purchase the product.

[0152] The consumer spending power estimation module M5 is used to determine the price level of the equipment used by each user to purchase goods based on the preset equipment price level classification rules, and to determine the consumer spending power and product category preference of each user based on the equipment price level.

[0153] The User Value Profile Building Module M6 is used to determine the user's value profile based on each user's value stratification results, product category preferences, spending power, and product category level preferences.

[0154] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 8 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores customer value assessment data based on device information and order information. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a customer value assessment method based on device information and order information.

[0155] Those skilled in the art will understand that Figure 8The structures shown are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0156] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0157] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0158] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0159] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0160] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0161] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0162] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for assessing the value of a customer base based on device information and order information, characterized in that, include: Obtain order data for each user within a preset time period and device information for the devices used by users to purchase goods; Adaptive RFM analysis is performed on the order data of each user to obtain the value stratification results for each user. Analyze the association rules between device characteristics and product categories based on each user's device information and order data; Estimate each user’s product category preference based on the association rules between device characteristics and product categories and the device characteristics of the device used by the user to purchase the product. The price tier of the equipment used by each user to purchase goods is determined according to the preset equipment price tier classification rules, and the user's spending power and product category preference are determined according to the equipment price tier. A comprehensive value profile of each user is determined based on their value stratification results, product category preferences, spending power, and product category level preferences.

2. The method for assessing the value of a customer base based on device information and order information according to claim 1, characterized in that, Adaptive RFM analysis was performed on each user's order data to derive value stratification results for each user, specifically including: For each user, calculate the R value, F value, and M value based on the user's order data; R refers to the number of days since the last purchase, F refers to the total number of purchases, and M refers to the total amount spent. Cluster the R values ​​of each user to obtain multiple R clusters, and determine the R value score for each R cluster based on the user's R value; Cluster the F-values ​​of each user to obtain multiple F-clusters, and determine the F-value score for each F-cluster based on the user's F-value. Cluster the M values ​​of each user to obtain multiple M clusters, and determine the M value score for each M cluster based on the user's M value; Based on the data distribution results after R-value clustering, F-value clustering, and M-value clustering, the entropy weight method is applied to determine the weights of R-value, F-value, and M-value; Each user's RFM composite score is calculated based on their R value score, F value score, M value score, and the weights of their R, F, and M values. The RFM scores of each user are clustered to obtain RFM clusters; users in different RFM clusters have different values. The value stratification results for each user are determined based on the RFM cluster to which each user belongs.

3. The method for assessing the value of a customer base based on device information and order information according to claim 1, characterized in that, Based on each user's device information and order data, we analyze the association rules between device characteristics and product categories, specifically including: Each user's item transaction set is created based on their device information and order data; the item transaction set includes the device characteristics of the device used by the user to purchase the goods and the categories of goods purchased by the user. For each user, multiple item combinations are determined based on the item transaction set; the item combination includes any device feature and any product category; Calculate the support and confidence of each item combination; The support and confidence of each item combination are compared with their respective thresholds to determine the item combinations with a relationship, that is, the association rules between equipment characteristics and product categories.

4. The method for assessing the value of a customer base based on device information and order information according to claim 3, characterized in that, The support and confidence scores of each item combination are compared with their respective thresholds to determine item combinations with a correlation, specifically including: If the support of an item combination is greater than or equal to the corresponding threshold, and the confidence level is greater than or equal to the corresponding threshold, then the current item combination is an item combination with an association relationship.

5. The method for assessing the value of a customer base based on device information and order information according to claim 3 or 4, characterized in that, The formula for calculating support is: ； in, Represents a combination of items Support level; This indicates a characteristic of a device; Indicates a product category; This represents the number of trading users who simultaneously include both X and Y; N represents the total number of trading users.

6. The method for assessing the value of a customer base based on device information and order information according to claim 3 or 4, characterized in that, The formula for calculating confidence level is: ； in, Represents a combination of items Confidence level; This indicates a characteristic of a device; Indicates a product category; This indicates the number of users who include both X and Y. This represents the number of users including X.

7. The method for assessing the value of a customer base based on device information and order information according to claim 1, characterized in that, The price tier for the equipment purchased by each user is determined based on a pre-defined set of equipment price tiers, specifically including: The initial price range will be determined based on the retail price of the equipment used by the user when purchasing the product. Determine whether the device used by the user to purchase the product has been sold for a longer period than a preset duration; If not, the initial price tier will be the final price tier; If so, the degraded price of the device used by the user to purchase the product will be determined based on the duration of the sale, and the final price tier will be determined based on the degraded price.

8. A device for assessing the value of a customer base based on device information and order information, characterized in that, include: The information acquisition module is used to acquire order data of each user and equipment information of the devices used by the user to purchase goods within a preset time period; The RFM analysis module is used to perform adaptive RFM analysis on the order data of each user to obtain the value stratification results for each user. The association rule analysis module is used to analyze the association rules between device characteristics and product categories based on each user's device information and order data. The product category preference estimation module is used to estimate each user's product category preference based on the association rules between device characteristics and product categories and the device characteristics of the device used by the user to purchase the product; The spending power estimation module is used to determine the price level of the equipment used by each user to purchase goods based on the preset equipment price level classification rules, and to determine the spending power and product category preference of each user based on the equipment price level. The user value profile building module is used to determine the user's value profile based on each user's value stratification results, product category preferences, spending power, and product category level preferences.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the crowd value assessment method based on device information and order information as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the crowd value assessment method based on device information and order information as described in any one of claims 1-7.

Images