A dynamic pricing method and system based on weight grading

Abstract

The application discloses a dynamic pricing method and system based on weight classification, relates to the technical field of data processing, and automatically realizes classification of users, establishes price classification rules, automatically identifies time into a variable period, and automatically calculates a total price of target goods corresponding to actual weight in a classified manner.The application comprises the following steps: obtaining a price fluctuation trend of target goods in a specific historical time period before the current time, determining at least one time interval in which the price of the target goods changes with seasonal cycles according to the price fluctuation trend, determining at least one jump threshold value for controlling price changes according to a current user category, wherein the jump threshold value refers to at least one weight boundary value for updating the price after the purchase weight exceeds the weight boundary value, and calculating a total price of the target goods corresponding to the actual weight in a manner that the corresponding unit weight price is switched when the actual weight exceeds each jump threshold value when any specific time interval is identified.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a dynamic pricing method and system based on weight grading. Background Technology

[0002] Currently, the trading of bulk commodities such as fertilizers, cement, and sand and gravel, sold by weight, generally relies on offline weighing and manual calculation. Taking fertilizer wholesale as an example, because the final amount is highly correlated with the transaction weight, the transaction process is usually as follows: after the buyer determines the tonnage to be purchased, the seller needs to arrange for a vehicle to load the goods. The vehicle must drive into a weighbridge for empty weighing, and then be weighed again after loading. The difference between the two weights is the actual loaded weight. This process not only requires dedicated personnel to operate the weighbridge and record data, but also leads to disputes due to differences in cargo moisture and packaging tare weight, resulting in repeated weighings.

[0003] The entire transaction chain described above involves multiple weighings, manual form filling, and telephone inquiries, making it inefficient and prone to errors. Furthermore, the industry commonly employs a "one-size-fits-all" pricing method—for example, "one price for under 5 tons, one price for over 5 tons." How prices are determined, and when different weights require separate pricing, is often determined verbally by salespeople based on experience, followed by using a calculator to categorize weights according to price levels and calculate the total amount. This is prone to errors, especially since inexperienced salespeople are unaware of different customers' pricing habits, leading to inappropriate pricing. Secondly, due to a lack of quantitative analysis of historical transaction data and market fluctuations, this extensive pricing and trading model is no longer suitable for the high-frequency fluctuations and diversified customer needs of the commodity market. Summary of the Invention

[0004] In view of this, this application provides a dynamic pricing method and system based on weight grading. Through big data analysis, it automatically classifies users and establishes price grading rules. Based on the price changes with seasonal cycles, it determines the unit price corresponding to different levels in the current cycle. When the time enters the change cycle, it automatically calculates the total price corresponding to the target product based on the actual weight in a grading manner.

[0005] The first aspect of this application provides a dynamic pricing method based on weight grading, the method comprising:

[0006] Obtain the price fluctuation trend of the target commodity within a specific historical time period prior to the current time, and determine at least one time interval for the price of the target commodity to change with the seasonal cycle based on the price fluctuation trend;

[0007] Determine at least one jump threshold to control price changes based on the current user category; the jump threshold refers to at least one weight boundary value that the purchase weight exceeds before the price is updated.

[0008] When an item enters any specific time interval, the total price corresponding to the target product based on the actual weight is calculated by switching the corresponding unit weight price for each time the actual weight exceeds a threshold.

[0009] Optionally, the method further includes:

[0010] At least one pricing dividing line is generated based on the price fluctuation data corresponding to the time interval; the pricing dividing line is used to divide the price interval corresponding to the jump threshold.

[0011] Calculate the tiered unit price after the weight exceeds the current jump threshold based on the values ​​within the price range;

[0012] Upon detecting entry into any specific time interval, the total price corresponding to the target product based on the actual weight is calculated by switching the corresponding unit weight price for each time the actual weight exceeds a threshold, including:

[0013] When an item enters any specific time interval, the total price corresponding to the target item based on the actual weight is calculated by switching to the corresponding grade price for each time the actual weight exceeds a jump threshold.

[0014] Optionally, the process of obtaining the current user category includes:

[0015] Collect user habit characteristics for historical target products; the habit characteristics include: the number of transactions of the user in a period of time before the current time, the user's registration date, the user's address, the user's purchase frequency, and the interval between the user's most recent purchase and the current time;

[0016] Using the aforementioned habitual features as input variables, the K-Means clustering algorithm is executed to obtain multiple user grouping results;

[0017] Calculate the distance between the current user and each user group result to obtain the current user's user category.

[0018] Optionally, obtain the price fluctuation trend of the target commodity within a specific historical time period prior to the current time, including:

[0019] Price data of the target product within the specific historical time period can be retrieved from the data platform via API interface;

[0020] Fill in missing values ​​for the price data and set the time corresponding to each price data point as the index marker;

[0021] Using the index marker as the horizontal axis and the price as the vertical axis, a seasonal cycle curve is plotted to obtain the price fluctuation trend.

[0022] Optionally, determining at least one time interval for the price of the target commodity to change with the seasonal cycle based on the price fluctuation trend includes:

[0023] Define the peak and trough ranges of the seasonal cyclical variation curve;

[0024] For the peak range and the trough range, the start time and end time are extracted respectively to obtain the at least one time interval.

[0025] Optionally, at least one pricing dividing line is generated based on the price fluctuation data corresponding to the time interval, including:

[0026] Calculate the average price within the time interval, and generate a first straight line composed of the average price within the peak range or the trough range;

[0027] The first line is copied and shifted up and down by a specific unit to obtain the second and third lines; the specific unit refers to the standard deviation of the data within the time interval.

[0028] Optionally, after plotting a seasonal cyclical change curve with the index marker as the horizontal axis and the price as the vertical axis to obtain the price fluctuation trend, the method further includes:

[0029] Noise in the price data is detected, and a random noise curve is generated using the index marker as the horizontal axis.

[0030] When data exceeding a boundary threshold is detected in the random noise curve, the noise time corresponding to the data exceeding the boundary threshold is determined;

[0031] Search for historical weather data corresponding to the noise time on the data platform through the API interface;

[0032] Calculate the tiered unit price based on the values ​​within the price range after the weight exceeds the current tier threshold, including:

[0033] When no feature corresponding to the historical weather data is detected, the graded unit price after the weight exceeds the current jump threshold is calculated based on the value within the price range, and the price change value corresponding to random noise is subtracted from the graded unit price; the random noise value is obtained by reading the value of the random noise curve.

[0034] A second aspect of this application provides a dynamic pricing system based on weight grading, the system comprising:

[0035] The time determination module is used to obtain the price fluctuation trend of the target commodity within a specific historical time period before the current time, and to determine at least one time interval for the price of the target commodity to change with the seasonal cycle based on the price fluctuation trend.

[0036] The weight threshold determination module is used to determine at least one jump threshold for controlling price changes based on the current user category; the jump threshold refers to at least one weight boundary value at which the price is updated after the purchased weight exceeds the current user category.

[0037] The pricing module is used to calculate the total price corresponding to the target product of the actual weight by switching the corresponding unit weight price for each time the actual weight exceeds a certain threshold when it is detected that the product has entered any specific time interval.

[0038] A third aspect of this application provides a server, including: a processor and a memory, the processor and the memory being connected via a communication bus; wherein, the processor is used to call and execute a program stored in the memory; the memory is used to store the program, the program being used to implement the weight-based dynamic pricing method provided in the first aspect of this application.

[0039] A fourth aspect of this application provides a computer-readable storage medium storing computer-executable instructions for performing a weight-based dynamic pricing method as provided in the first aspect of this application.

[0040] Compared to existing technologies, the weight-based dynamic pricing method proposed in this application retrieves price data of the target product from the most recent historical time period. Through big data analysis, it obtains the long-term price trend, seasonal cyclical changes, and random noise of the target product over the historical time period. By analyzing and calculating the seasonal cyclical changes, it determines the time interval in which the price of the target product will objectively fluctuate cyclically. When entering the same time interval, it calculates the price according to a cyclical pricing strategy set for that time interval. The cyclical pricing strategy determines the user category through big data statistics and uses different unit price weight boundary values ​​as jump thresholds. On the other hand, it determines the specific price level based on the seasonal cyclical changes in prices. The entire cyclical pricing strategy is determined by big data statistical information, which uncovers the inherent connections between users and the objective laws of prices. It automates pricing while solving the problem that traditional fixed cost plus pricing is seriously lagging behind market changes. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0042] Figure 1This is a flowchart of the steps of the dynamic pricing method based on weight grading proposed in the embodiments of this application;

[0043] Figure 2 This is a schematic diagram of a seasonal cyclical variation curve generated as an example in this application;

[0044] Figure 3 This is a schematic diagram of an example interface for setting the price per unit weight according to this application;

[0045] Figure 4 This is a schematic diagram of a pricing segmentation line generated for the peak range corresponding to time region X, as described in this application.

[0046] Figure 5 This is a flowchart illustrating an example of a weight-based dynamic pricing method according to this application;

[0047] Figure 6 This is an architecture diagram of the dynamic pricing system based on weight grading proposed in the embodiments of this application. Detailed Implementation

[0048] 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.

[0049] In this application, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0050] The applicant has identified the following problems in the current industry of selling bulk commodities by weight:

[0051] 1. Since transactions are based on weight, relying on manual pricing is inefficient and prone to errors. Merchants need to frequently query historical and real-time price data from multiple public platforms, manually compile the data into tables, and then rely on personal experience to set prices for different purchase weight ranges (e.g., 1-5 tons, 5-10 tons). When the variety of goods or the amount of price data is large, this task becomes extremely tedious and prone to calculation or recording errors.

[0052] 2. Pricing strategies cannot dynamically adapt to the market. Traditional pricing methods are usually based on fixed costs plus a markup, making it difficult to effectively capture and utilize the long-term trends, seasonal cycles, and random fluctuations implied in price data. This results in pricing strategies lagging significantly behind market changes; for example, they cannot automatically raise fertilizer prices before the peak demand period for spring planting, or offer timely discounts during the off-season.

[0053] 3. When prices need to be adjusted for weights exceeding a certain threshold, pricing lacks data support: Existing methods often rely on experience to determine "wholesale prices for kilograms and above," lacking quantitative analysis of historical purchase volume distribution and the number of customers in different price ranges. The setting of pricing ranges is highly subjective, making it difficult to achieve automatic and accurate tiered pricing based on purchase weight, potentially leading to profit losses and reduced customer satisfaction.

[0054] Example 1

[0055] In view of the above problems, this application proposes a dynamic pricing method based on weight grading. It automatically classifies users and establishes price grading rules through big data. It determines the unit price corresponding to different levels in the current cycle based on the price changes with the seasonal cycle. It automatically identifies when the time enters the change cycle and automatically calculates the total price corresponding to the target product of actual weight in a grading manner.

[0056] Figure 1 This is a flowchart illustrating the steps of the dynamic pricing method based on weight grading proposed in this application. Figure 1 As shown, the steps include:

[0057] S110: Obtain the price fluctuation trend of the target commodity within a specific historical time period prior to the current time, and determine at least one time interval in which the price of the target commodity changes with the seasonal cycle based on the price fluctuation trend.

[0058] The specific historical time period can be one month prior to the current time, three years prior to the current time, etc., set according to the actual situation of the target commodity. The target commodity can be fertilizer, cement, sand, etc., which are traded by weight. For example, if the price of cement has not fluctuated significantly or broken down in the last N years, the historical time period can be set to the last N years. Assuming the current time is XX year XX month XX day, the specific historical time period can be set to {XX year XX month XX day, XX-N year XX month XX-1}.

[0059] Due to the natural changes in weather and social activities, the prices of fertilizers, cement, and sand may experience multiple cycles of change within a specific historical period or within a specific historical period. Therefore, the time intervals that vary with seasonal cycles may be one or more.

[0060] Within a natural time period, this refers to a year, a month, a quarter, etc.

[0061] This application embodiment also provides a specific process for executing S110, which includes sub-steps:

[0062] S1101: Retrieve price data of the target product within the specified historical time period from the data platform via API interface.

[0063] Taking the urea price data as an example, after registering on the data platform, set the parameters (commodities) to urea through the API interface, set "Historical Data Range (months)" to 36, and after running, download a CSV or Excel file containing 17 fields such as historical price, average price, and monthly / annual changes to obtain the urea price data for the 36 months prior to the current time.

[0064] S1102: Fill in missing values ​​for price data and set the time corresponding to each price data as an index marker.

[0065] Data preprocessing is performed by reading the data into memory using the Pandas library, checking the price data, automatically scanning the file using the missing value detection function, counting missing values, and filling the missing values ​​using specified strategies such as filling with a fixed value (df.fillna(0)), filling with the mean (df.fillna(df.mean())), and filling with the median (df.fillna(df.median())).

[0066] Parse the price data, convert the time information into a time format, set the time format data as an index marker, set the time interval, check the data for missing dates according to the time interval, and fill in the missing dates if they are found.

[0067] For example, the index frequency is set to DAY, and indexes are automatically generated for January 1, January 2, January 4, and January 5. January 3 is detected as NaN (null value), and January 3 is filled in. The average value is then filled in for the data on January 3.

[0068] S1103: Using the index marker as the horizontal axis and the price as the vertical axis, plot a seasonal cycle change curve to obtain the price fluctuation trend.

[0069] By using the methods described above, a complete and standardized time series index can be obtained, providing a foundation for comprehensive data analysis.

[0070] After obtaining the seasonal cycle change curve as described above, the following sub-steps are further performed: S110 is executed to process the seasonal cycle change curve and determine at least one time interval in which the price of the target commodity changes with the seasonal cycle.

[0071] S1104: Define the peak and trough ranges of the seasonal cyclical variation curve.

[0072] Taking the definition of peak range as an example, in the peak detection function such as scipy.signal.find_peaks, filter conditions such as height threshold, prominence threshold, minimum horizontal distance, and width range of the peaks can be set to filter out multiple peak ranges and multiple valley ranges.

[0073] S1105: Extract the start time and end time from the peak range and the trough range respectively to obtain the at least one time interval.

[0074] The start and end points of the peaks can be detected using the `width` parameter of `scipy.signal.find_peaks` or a change-point detection algorithm, thus obtaining the time interval {start point, end point}.

[0075] One example of this application employs a LOESS-based seasonal trend decomposition (STL) to break down prices into three parts: long-term trend, seasonal cycle, and random noise.

[0076] Figure 2 This is a schematic diagram of a seasonal cyclical change curve generated as an example of this application. The retrieved price data is the monthly wholesale price data of urea in a certain region from year X to year X+2. The generated cyclical curve is as follows: Figure 2 As shown, the long-term trend curve shows that urea prices generally increased from year X to year X+2. The seasonal cycle curve shows that prices peaked in January-March of the previous year before spring planting, with a secondary peak in June-August during the summer fertilization period, and then fell back to a low point in demand after the autumn harvest in September-October. The random noise curve is generated after removing the influence of trends and seasons, and abnormally strong noise corresponds to occasional events such as extreme weather.

[0077] right Figure 2 The seasonal cycle curve shown is used to calculate the range of two peaks and one trough during the X+2 year period, resulting in three time intervals. The time intervals corresponding to the two peak ranges are {x1,x2} and {x3,x4}.

[0078] S120: Determine at least one jump threshold for controlling price changes based on the current user category; the jump threshold refers to at least one weight boundary value at which the purchased weight exceeds the price update.

[0079] The tiered threshold refers to the weight boundary value that marks a price change. For example, assuming the tiered thresholds are 50, 60, and 80, then a user purchasing 0-50kg would use unit price A, 50-60kg would use unit price B, and 60-80kg would use unit price C; if the tiered thresholds are 20, 70, and 80, then a user purchasing 0-20kg would use unit price A, 20-70kg would use unit price B, and 70-80kg would use unit price C.

[0080] S120 determines different price tier thresholds based on user categories, objectively analyzes differences in user behavior, avoids incentive misalignment caused by uniform thresholds, automatically sets thresholds through data-driven approaches, eliminates subjective biases from human experience, and can dynamically adjust according to changes in user behavior to achieve a long-term optimized tiered pricing strategy.

[0081] This application also provides a method for obtaining the current user category:

[0082] S200: Collect user habit characteristics for historical target products; the habit characteristics include: user transaction data within a period of time before the current time, user registration date, user address, user purchase frequency, and the interval between the user's most recent purchase and the current time.

[0083] Collect transaction data to objectively represent the overall scale of the user base; collect user registration dates to determine if a user is a new customer; collect user addresses to assess the ease of product pickup; collect user purchase frequency to objectively represent the user's market potential; and collect the interval between a user's most recent purchase and the current time to assess the user's activity level.

[0084] First, relevant data is acquired from the system, such as the total number of transactions completed by the user within a preset time period before the current time, the date the user registered an account in the system, the average number of purchases per unit time, etc. The collected data is then transformed into features, and all features are concatenated into a feature vector in a pre-set order. Since K-Means clustering is used, the calculation result is related to the comprehensive distance of all features. Therefore, this application embodiment does not limit the concatenation order of features.

[0085] S210: Using the aforementioned habitual features as input variables, execute the K-Means clustering algorithm to obtain multiple user grouping results.

[0086] Set the number of clusters K, select the initial cluster center as the initialization method, calculate the features to assign each user to the nearest cluster center, and then recalculate the center of each cluster until the centers converge or the maximum number of iterations is reached.

[0087] Users making transactions in the system also have records of each transaction, including the relationship between the weight of the target product purchased and the unit price. For example, in user A's historical transactions, purchases under Y kilograms are usually settled at the higher unit price, and purchases over Y kilograms are settled at the lower unit price, where Y is the weight boundary value. In one case, there may be multiple weight boundary values. Therefore, for each cluster obtained from clustering, the median R of the weight boundary values ​​of all users in that cluster is calculated, and the cluster is labeled R.

[0088] By using the clustering method described above, users with objectively inherent connections are grouped together. This allows for the analysis of historical transactions and the patterns in price adjustments based on purchase weight for different user categories. This helps identify the different purchase volumes that different users receive when they achieve a lower unit price. For example, during customer acquisition, a lower weight threshold can be set for new users to attract them to purchase more; a higher weight threshold can be set for users with lower purchase frequency to balance revenue. However, in reality, the factors influencing the weight threshold are complex, and if experienced staff are not set according to established practices, it can lead to losses.

[0089] S220: Calculate the distance between the current user and each user grouping result to obtain the current user's user category.

[0090] Calculate the distance between the current user and the center of each cluster obtained from the above clustering, and assign the user to the cluster with the closest distance.

[0091] Based on the above method of determining user categories through clustering, the process of determining at least one jump threshold for controlling price changes according to the current user category includes:

[0092] Read the cluster's tag R and set the skip threshold based on tag R.

[0093] For example, suppose the R label of a cluster is: , , , < < Then set the skip threshold to , , .

[0094] In one example of this application, the price per unit weight for each jump threshold can be pre-set in the system. Figure 3 This is a schematic diagram of an example interface for setting the price per unit weight according to this application, such as... Figure 3As shown, the jump thresholds are respectively , , Staff can set the corresponding unit weight price based on experience for each tier. , , When a user does not belong to this category, they purchase... For target items with a +3 rating, the system will automatically calculate the price. × +( - )× +( +3- )× .

[0095] Another example in this application illustrates the process of automatically obtaining the unit weight price corresponding to different jump thresholds using big data.

[0096] S300: Generate at least one pricing dividing line based on the price fluctuation data corresponding to the time interval; the pricing dividing line is used to divide the price interval corresponding to the jump threshold.

[0097] Since price fluctuation data reflects price changes over a short period of time and corresponds to similar change cycles, using price fluctuation data corresponding to the time interval as a reference makes the calculation of pricing more accurate.

[0098] This application also provides an example of the execution process for generating at least one pricing dividing line based on price fluctuation data corresponding to the time interval:

[0099] S3001: Calculate the average price within the time interval, and generate a first straight line composed of the average price within the peak range or the trough range.

[0100] For example, you can calculate the total price within a time interval, divide it by the number of days to get the average price, and then generate a straight line representing the average price in the curve area corresponding to the time interval.

[0101] S3002: Copy the first straight line and shift it up and down by a specific unit respectively to obtain the second and third straight lines; the specific unit refers to the standard deviation of the data within the time interval.

[0102] Using the first straight line as the center line (SMA), calculate the standard deviation of the grid sequence within the time interval. The second straight line is the upper track = SMA + 2 × standard deviation, and the third straight line is the lower track = SMA - 2 × standard deviation.

[0103] Figure 4This is a schematic diagram of a pricing segmentation line generated for a time region X corresponding to a peak range, as described in this application. Time region X is... Figure 2 The peak range corresponding to the time interval {x5, x6} in the seasonal cyclical change curve shown is used to generate the first straight line D1. This line is then copied and shifted upwards and downwards by specific units to obtain the second straight line D2 and the third straight line D3. In this way, the price fluctuation data within the time interval is subdivided into several different parts, corresponding to low-price areas, middle-price areas, and high-price areas. The median price is selected from each corresponding area as the actual unit weight price. Because the reference interval for pricing is smaller, the pricing is more accurate and corresponds to the original price range.

[0104] S310: Calculate the tiered unit price after the weight exceeds the current jump threshold based on the values ​​within the price range.

[0105] Upon detecting entry into any specific time interval, the total price corresponding to the target product based on the actual weight is calculated by switching the corresponding unit weight price for each time the actual weight exceeds a threshold, including:

[0106] When an item enters any specific time interval, the total price corresponding to the target item based on the actual weight is calculated by switching to the corresponding grade price for each time the actual weight exceeds a jump threshold.

[0107] For example, price zones can be flexibly set according to pricing division lines, and the median can be selected based on the price data of the D1-D3 interval to obtain the transaction weight. The previous tiered unit price was calculated by selecting the median of the price data from the D1-D2 range, resulting in a transaction weight of [value missing]. And not reached The tiered unit price is determined by selecting the median from the price data of the D2-peak fixed-point range, resulting in a transaction weight reaching [a certain value]. And not reached The unit price is tiered, and so on.

[0108] S120: When it is detected that the target product has entered any specific time interval, the total price corresponding to the actual weight is calculated by switching the corresponding unit weight price for each time the actual weight exceeds a jump threshold.

[0109] The above method can automatically acquire and process time-series price data, intelligently identify price change cycles and key nodes, dynamically calculate tiered pricing based on purchase weight, calculate the unit price of purchase quantity in different weight ranges based on historical price data, and automatically calculate the price calculation strategy for different users under different purchase quantities. It does not rely on human experience and solves the current technical problems of cumbersome manual operation, slow response, and crude tiering.

[0110] This application also provides a process for implementing a weight-based dynamic pricing method, which plots a seasonal cycle curve with the index marker as the horizontal axis and the price as the vertical axis, and obtains a random noise curve to adjust the price when random noise is detected, in order to cope with price fluctuations caused by extreme weather.

[0111] S400: Retrieve price data of the target product within the specified historical time period from the data platform via API interface.

[0112] S410: Fill in missing values ​​for price data and set the time corresponding to each price data point as an index marker.

[0113] S420: Using the index marker as the horizontal axis and the price as the vertical axis, plot a seasonal cycle change curve to obtain the price fluctuation trend.

[0114] S430: Detect noise in the price data and generate a random noise curve using the index marker as the horizontal axis.

[0115] S440: When data exceeding the boundary threshold is detected in the random noise curve, the noise time corresponding to the data exceeding the boundary threshold is determined.

[0116] S450: Use the API interface to search for historical weather data corresponding to the noise time on the data platform.

[0117] S460: Generate at least one pricing dividing line based on the price fluctuation data corresponding to the time interval; the pricing dividing line is used to divide the price interval corresponding to the jump threshold.

[0118] S470: When entering any specific time interval and no feature corresponding to the historical weather data is detected, calculate the graded unit price after the weight exceeds the current jump threshold based on the value within the price interval, and subtract the price change value corresponding to random noise from the graded unit price; the random noise value is obtained by reading the value of the random noise curve.

[0119] If the long-term trend price corresponding to time T is The seasonal price cycle corresponding to time T is Find the seasonal period price corresponding to time T in the price data. The price change corresponding to random noise = - - This indicates price fluctuations caused by extreme weather, in addition to normal trends and seasonal effects.

[0120] S480: Determine at least one jump threshold to control price changes based on the current user category; the jump threshold refers to at least one weight boundary value at which the purchased weight exceeds the price update.

[0121] S490: When entering any specific time interval, calculate the total price corresponding to the target product for the actual weight by switching the corresponding unit weight price for each time the actual weight exceeds a jump threshold.

[0122] refer to Figure 2 By comparing the random noise curve and the seasonal cycle variation curve, information is obtained: the data in the time interval {x1,x2} contains random noise exceeding the boundary threshold. The noise time T corresponding to the data exceeding the boundary threshold is determined. Historical weather data corresponding to the noise time is searched in the data platform through the API interface. If the weather feature corresponding to the historical weather data is detected, the price is calculated using a tiered unit price. If the weather feature corresponding to the historical weather data is not detected, it indicates that the current price will not be affected by extreme weather that occurred in the historical period. Therefore, the impact of the weather on the price is subtracted, thereby eliminating the price deviation caused by occasional extreme weather events.

[0123] Figure 5 This is a flowchart illustrating an example of a dynamic pricing method based on weight grading according to this application. Upon receiving a price calculation request from a user, the system first checks if a specific time interval has been entered. If not, the system calculates the price according to basic calculation rules. If a specific time interval has been entered, steps S100-S120 are executed to determine the pricing rules corresponding to that time interval. These rules include the threshold for entering each grade and the unit weight price for each grade. The entire price calculation process utilizes big data analysis to interpret the periodic changes in the target commodity price and dynamically determines the unit weight price based on these periodic changes, adapting to the high-frequency fluctuations in the commodity market.

[0124] The weight-based dynamic pricing method proposed in this application retrieves price data of the target product within the most recent historical time period. Through big data analysis, it obtains the long-term price trend, seasonal cyclical changes, and random noise of the target product over the historical time period. By analyzing and calculating the seasonal cyclical changes, it determines the time interval in which the price of the target product will objectively fluctuate cyclically. When entering the same time interval, it calculates the price according to a cyclical pricing strategy set for that time interval. The cyclical pricing strategy determines the user category through big data statistics and uses different unit price weight boundary values ​​as jump thresholds. On the other hand, it determines the specific price level based on the seasonal cyclical changes in prices. The entire cyclical pricing strategy is determined by big data statistical information, which uncovers the inherent connections between users and the objective laws of prices. It automates pricing while solving the problem that traditional fixed cost plus pricing is seriously lagging behind market changes.

[0125] Example 2

[0126] Based on the weight-based dynamic pricing method provided in Embodiment 1 of this application, correspondingly, Embodiment 2 of this application also provides a weight-based dynamic pricing system. Figure 6 This is an architecture diagram of the dynamic pricing system based on weight grading proposed in the embodiments of this application, as follows: Figure 6 As shown, the system includes:

[0127] The time determination module 601 is used to obtain the price fluctuation trend of the target commodity within a specific historical time period before the current time, and to determine at least one time interval for the price of the target commodity to change with the seasonal cycle based on the price fluctuation trend.

[0128] The weight threshold determination module 602 is used to determine at least one jump threshold for controlling price changes based on the current user category; the jump threshold refers to at least one weight boundary value after the purchased weight exceeds the updated price.

[0129] The pricing module 603 is used to calculate the total price corresponding to the target product of the actual weight when it is detected that the product has entered any specific time interval, by switching the corresponding unit weight price for each time the actual weight exceeds a jump threshold.

[0130] Optionally, the system further includes:

[0131] The unit price determination module is used to generate at least one pricing dividing line based on the price fluctuation data corresponding to the time interval; the pricing dividing line is used to divide the price interval corresponding to the jump threshold; and to calculate the tiered unit price after the weight exceeds the current jump threshold based on the value within the price interval.

[0132] The pricing module is specifically used to calculate the total price corresponding to the target product of the actual weight when it detects that the product has entered any specific time interval, by switching to the corresponding grade unit price for each time the actual weight exceeds a jump threshold.

[0133] Optionally, the system further includes a feature acquisition module and a classification module;

[0134] The feature acquisition module is used to collect user habit characteristics of target products in the history of data collection; the habit characteristics include: the number of transactions of the user in a period of time before the current time, the user's registration date, the user's address, the user's purchase frequency, and the interval between the user's most recent purchase and the current time;

[0135] The classification module uses the habitual features as input variables, executes the K-Means clustering algorithm to obtain multiple user clustering results, calculates the distance between the current user and each user clustering result, and obtains the user category of the current user.

[0136] Optionally, the time determination module includes:

[0137] The retrieval submodule is used to retrieve price data of the target product within the specified historical time period through the API interface on the data platform;

[0138] The settings submodule is used to fill in missing values ​​in the price data and set the time corresponding to each price data as an index marker.

[0139] The first drawing submodule is used to draw a seasonal cycle change curve with the index mark as the horizontal axis and the price as the vertical axis to obtain the price fluctuation trend.

[0140] Optionally, the time determination module also includes:

[0141] The range calibration submodule is used to define the peak and trough ranges of the seasonal cyclical variation curve;

[0142] The time extraction submodule is used to extract the start time and end time from the peak range and the trough range, respectively, to obtain the at least one time interval.

[0143] Optionally, the unit price determination module is specifically used to calculate the average price within the time interval, generate a first straight line composed of the average price within the peak range or the trough range; copy the first straight line and shift it up and down by a specific unit respectively to obtain a second straight line and a third straight line; the specific unit refers to the standard deviation of the data within the time interval.

[0144] Optionally, the system further includes:

[0145] The second drawing submodule detects noise in the price data and generates a random noise curve using the index marker as the horizontal axis.

[0146] The detection submodule is used to determine the noise time corresponding to the data exceeding the boundary threshold when the random noise curve is detected to contain data exceeding the boundary threshold.

[0147] The search submodule is used to search for historical weather data corresponding to noise times on the data platform via API interfaces;

[0148] The unit price determination module is specifically used to calculate the graded unit price after the weight exceeds the current jump threshold based on the value within the price range when no feature corresponding to the historical weather data is detected, and to subtract the price change value corresponding to random noise from the graded unit price; the random noise value is obtained by reading the value of the random noise curve.

[0149] The specific principles and execution processes of each unit in the weight-based dynamic pricing system disclosed in Embodiment 2 of this application can be found in the corresponding parts of the weight-based dynamic pricing method disclosed in Embodiment 1 of this application, and will not be repeated here.

[0150] Example 3

[0151] Embodiment 3 of this application provides a server, including: a processor and a memory, the processor and the memory being connected via a communication bus; wherein, the processor is used to call and execute a program stored in the memory; the memory is used to store the program, the program being used to implement the weight-based dynamic pricing method provided in Embodiment 1 of this application.

[0152] Example 4

[0153] Embodiment 4 of this application provides a computer-readable storage medium storing computer-executable instructions for executing the weight-based dynamic pricing method provided in Embodiment 1 of this application.

[0154] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computing software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0155] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0156] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A dynamic pricing method based on weight grading, characterized in that, The method includes: Obtain the price fluctuation trend of the target commodity within a specific historical time period prior to the current time, and determine at least one time interval for the price of the target commodity to change with the seasonal cycle based on the price fluctuation trend; Determine at least one jump threshold to control price changes based on the current user category; the jump threshold refers to at least one weight boundary value that the purchase weight exceeds before the price is updated. When an item enters any specific time interval, the total price corresponding to the target product based on the actual weight is calculated by switching the corresponding unit weight price for each time the actual weight exceeds a threshold.

2. The method of claim 1, wherein, The method further includes: At least one pricing dividing line is generated based on the price fluctuation data corresponding to the time interval; the pricing dividing line is used to divide the price interval corresponding to the jump threshold. Calculate the tiered unit price after the weight exceeds the current jump threshold based on the values ​​within the price range; Upon detecting entry into any specific time interval, the total price corresponding to the target product based on the actual weight is calculated by switching the corresponding unit weight price for each time the actual weight exceeds a threshold, including: When an item enters any specific time interval, the total price corresponding to the target item based on the actual weight is calculated by switching to the corresponding grade price for each time the actual weight exceeds a jump threshold.

3. The method of claim 1, wherein, The process of obtaining the current user category includes: Collect user habit characteristics for historical target products; the habit characteristics include: the number of transactions of the user in a period of time before the current time, the user's registration date, the user's address, the user's purchase frequency, and the interval between the user's most recent purchase and the current time; Using the aforementioned habitual features as input variables, the K-Means clustering algorithm is executed to obtain multiple user grouping results; Calculate the distance between the current user and each user group result to obtain the current user's user category.

4. The method of claim 2, wherein, To obtain the price fluctuation trend of a target commodity within a specific historical time period prior to the current time, including: Price data of the target product within the specific historical time period can be retrieved from the data platform via API interface; Fill in missing values ​​for the price data and set the time corresponding to each price data point as the index marker; Using the index marker as the horizontal axis and the price as the vertical axis, a seasonal cycle curve is plotted to obtain the price fluctuation trend.

5. The method of claim 4, wherein, Determining at least one time interval for the price of the target commodity to change with the seasonal cycle based on the price fluctuation trend includes: Define the peak and trough ranges of the seasonal cyclical variation curve; For the peak range and the trough range, the start time and end time are extracted respectively to obtain the at least one time interval.

6. The method of claim 5, wherein, At least one pricing dividing line is generated based on the price fluctuation data corresponding to the time interval, including: Calculate the average price within the time interval, and generate a first straight line composed of the average price within the peak range or the trough range; The first line is copied and shifted up and down by a specific unit to obtain the second and third lines; the specific unit refers to the standard deviation of the data within the time interval.

7. The method according to claim 4, characterized in that, Using the index marker as the x-axis and the price as the y-axis, a seasonal cyclical change curve is plotted to obtain the price fluctuation trend. The method further includes: Noise in the price data is detected, and a random noise curve is generated using the index marker as the horizontal axis. When data exceeding a boundary threshold is detected in the random noise curve, the noise time corresponding to the data exceeding the boundary threshold is determined; Search for historical weather data corresponding to the noise time on the data platform through the API interface; Calculate the tiered unit price based on the values ​​within the price range after the weight exceeds the current tier threshold, including: When no feature corresponding to the historical weather data is detected, the graded unit price after the weight exceeds the current jump threshold is calculated based on the value within the price range, and the price change value corresponding to random noise is subtracted from the graded unit price; the random noise value is obtained by reading the value of the random noise curve.

8. A dynamic pricing system based on weight grading, characterized in that, The system includes: The time determination module is used to obtain the price fluctuation trend of the target commodity within a specific historical time period before the current time, and to determine at least one time interval for the price of the target commodity to change with the seasonal cycle based on the price fluctuation trend. The weight threshold determination module is used to determine at least one jump threshold for controlling price changes based on the current user category; the jump threshold refers to at least one weight boundary value at which the price is updated after the purchased weight exceeds the current user category. The pricing module is used to calculate the total price corresponding to the target product of the actual weight by switching the corresponding unit weight price for each time the actual weight exceeds a certain threshold when it is detected that the product has entered any specific time interval.

9. The dynamic pricing system based on weight grading according to claim 8, characterized in that, The system also includes: The unit price determination module is used to generate at least one pricing dividing line based on the price fluctuation data corresponding to the time interval; the pricing dividing line is used to divide the price interval corresponding to the jump threshold; and to calculate the tiered unit price after the weight exceeds the current jump threshold based on the value within the price interval. The pricing module is specifically used to calculate the total price corresponding to the target product of the actual weight when it detects that the product has entered any specific time interval, by switching to the corresponding grade unit price for each time the actual weight exceeds a jump threshold.

10. The dynamic pricing system based on weight grading according to claim 8, characterized in that, The system also includes a feature acquisition module and a classification module; The feature acquisition module is used to collect user habit characteristics of target products in the history of data collection; the habit characteristics include: the number of transactions of the user in a period of time before the current time, the user's registration date, the user's address, the user's purchase frequency, and the interval between the user's most recent purchase and the current time; The classification module uses the habitual features as input variables, executes the K-Means clustering algorithm to obtain multiple user clustering results, calculates the distance between the current user and each user clustering result, and obtains the user category of the current user.

Images