A system for precise matching and recommendation generation of commodities by fusing multi-dimensional label recognition

Abstract

This invention relates to the fields of intelligent recommendation and enterprise consumable management technology, and provides a product accurate matching and recommendation generation system that integrates multi-dimensional tag recognition. The system includes a request identification module, a cycle analysis module, a replacement range analysis module, a trend judgment module, and a correction coefficient generation module. The request identification module is used to obtain the historical procurement data, historical usage records, and current consumable product recommendation model of the corresponding consumables after the target enterprise submits a replacement request. By constructing a product accurate matching and recommendation generation system that integrates multi-dimensional tag recognition, this invention addresses the common pitfall of enterprises "blindly raising prices but reducing suitability" during the replacement process. It proposes an identification mechanism based on the downward trend of the suitability index referencing the procurement cycle, combined with a dynamic weight enhancement strategy when no effective downward trend is shown, to accurately adjust the influence of the suitability index in the current recommendation model.

Description

Technical Field

[0001] This invention belongs to the field of intelligent recommendation and enterprise consumable management technology, and in particular relates to a product accurate matching and recommendation generation system that integrates multi-dimensional tag recognition. Background Technology

[0002] In the online sales of consumable products, sales platforms typically offer recommendation services to enterprise users, helping them quickly select alternative products during the exchange or replenishment process. To improve recommendation efficiency and matching accuracy, existing sales platforms mostly build product ranking mechanisms based on recommendation models. These models combine multiple factors such as price, sales volume, reviews, and usage scope, assigning weighted values ​​to each factor to generate recommendation priorities, thereby achieving an intelligent and automated product recommendation process. This type of approach is widely used in traditional e-commerce scenarios and is gradually expanding into the field of enterprise-level consumable procurement, becoming one of the common technical means for supporting digital sales.

[0003] However, in the actual process of exchanging consumables in enterprises, some procurement decision-makers still lack experience and make subjective decisions regarding exchanges. This is especially true in new companies or when procurement is handled by non-professionals, where situations often arise where "high-priced consumables are blindly chosen in pursuit of quality or to avoid risk." This price-driven exchange strategy often ignores the specific usability of consumables, potentially leading to a decrease in usability and resulting in increased scrap rates and ineffective cost control. Existing recommendation models typically lack the ability to identify whether price increases genuinely improve usability, and they also lack a relevant dynamic response mechanism to adjust recommendation rankings. Summary of the Invention

[0004] The purpose of this invention is to provide a product accurate matching and recommendation generation system that integrates multi-dimensional tag recognition, in order to solve the problems mentioned in the background art.

[0005] This invention is implemented as follows: a product accurate matching and recommendation generation system integrating multi-dimensional tag recognition, the system including a request recognition module, a periodic analysis module, a return rate analysis module, a trend judgment module, and a correction coefficient generation module, wherein:

[0006] The request identification module is used to obtain the historical purchase data, historical usage records and current consumable product recommendation model of the corresponding consumables after the target enterprise submits a replacement request;

[0007] The cycle analysis module is used to identify several procurement cycles of consumables and determine whether there are specific replacement behaviors within each procurement cycle;

[0008] The exchange rate analysis module is used to filter several reference procurement cycles with consistent exchange frequency and consistent exchange scale, and calculate the average price increase of all exchange behaviors in each reference procurement cycle.

[0009] The trend judgment module is used to sort the reference procurement cycles according to the average price increase, extract the scrap rate of consumables in each reference procurement cycle, and determine whether the scrap rate shows an effective downward trend as the price increase rises.

[0010] The correction coefficient generation module is used to extract a specified reference procurement period that is consistent with the price increase of the current exchange request when no effective downward trend is detected. It obtains the decline of the consumable product adaptability index for each exchange behavior, generates an adaptability weight amplification coefficient, and enhances the weight of the adaptability index in the current consumable product recommendation model to improve the priority of consumable products with high adaptability in the recommendation ranking.

[0011] As a further limitation of the technical solution of the present invention, the specific exchange behavior refers to the replacement operation of low-priced consumables with high-priced consumables during the consumables exchange process.

[0012] As a further limitation of the technical solution of the present invention, the cycle analysis module is used to identify several procurement cycles from the historical procurement data of consumables, select procurement cycles with the same time length, and the time interval between adjacent procurement cycles is less than a preset time threshold.

[0013] As a further limitation of the technical solution of this embodiment of the invention, the exchange rate analysis module specifically includes:

[0014] The exchange cycle filtering unit is used to parse each procurement cycle in sequence and extract reference procurement cycles with consistent exchange behavior frequency and consistent scale of each exchange behavior. The consistent scale of the exchange behavior means that the quantity of consumables replaced in each exchange process is the same, or the quantity variation is within a preset tolerance range.

[0015] The replacement margin calculation unit is used to obtain the price increase of consumables for each replacement behavior in each reference procurement cycle based on historical procurement data, and to average the price increase of all replacement behaviors to obtain the average price increase for each reference procurement cycle.

[0016] As a further limitation of the technical solution of this embodiment of the invention, the trend judgment module specifically includes:

[0017] The sorting processing unit is used to sort the reference procurement cycles from smallest to largest according to the average price increase of each reference procurement cycle, and generate an ordered price increase sequence.

[0018] The scrap rate extraction unit is used to obtain the usage data of the corresponding consumables of the target enterprise in each reference procurement cycle based on historical usage records, and to obtain the scrap rate of consumables in the reference procurement cycle based on the ratio of scrapped quantity to input quantity.

[0019] The trend judgment unit is used to analyze whether the scrap rate corresponding to each reference procurement cycle in the ordered price increase sequence shows an effective downward trend as the price increase rises.

[0020] As a further limitation of the technical solution of the present invention, the effective downward trend refers to the average slope of the change trend of the corresponding scrap rate sequence in the ordered increase sequence being less than a preset negative threshold.

[0021] As a further limitation of the technical solution of this embodiment of the invention, the correction coefficient generation module specifically includes:

[0022] The exchange request parsing unit is used to parse the current exchange request and determine its corresponding expected price increase when no effective downward trend is detected.

[0023] The reference cycle matching unit is used to filter out a specified reference procurement cycle from historical data whose average price increase is consistent with the expected price increase of the current replacement request, and extract the compatibility index data related to the use effect of consumables for each replacement behavior in the specified reference procurement cycle, and determine the decrease of the compatibility index before and after the replacement.

[0024] The amplification factor calculation unit is used to generate the fit weight amplification factor based on the average of the absolute values ​​of the decrease in the fit index of all exchange behaviors.

[0025] The weight enhancement execution unit is used to obtain the initial weights assigned to the adaptability index in the current consumable product recommendation model, and to enhance and adjust them based on the adaptability weight amplification coefficient.

[0026] The ranking update unit is used to apply the enhanced and adjusted weighted weights to the current recommendation scoring model, re-rank the candidate consumable products, and thus improve the recommendation priority of consumable products with higher suitability index.

[0027] As a further limitation of the technical solution of the present invention, the adaptability index data refers to the multi-dimensional tag scoring results constructed based on dimensions such as functional attributes, usage habits, and compatibility;

[0028] The decrease in the compatibility index before and after the exchange is obtained by calculating the difference in the average score of the consumable before and after the exchange on the multidimensional label scoring results, or by the ratio of the relative deviation between the two.

[0029] As a further limitation of the technical solution of the present invention, after enhancing and correcting the initial weighted weight of the fit index based on the fit weight amplification coefficient, the weighted weights of other scoring factors already set in the current recommendation model should be reduced by an equal amount at the same time, so as to maintain the overall balance between the weighted weights and avoid imbalance of recommendation results due to the increase of the weight of a single factor.

[0030] As a further limitation of the technical solution of this embodiment of the invention, when enhancing and adjusting the fit degree weight amplification coefficient, a preset enhancement function is used, and the enhancement function is:

[0031] ;

[0032] in, This refers to the initial weighting of the adaptation index in the current product and consumable recommendation model. This refers to the corrected weighted average. This refers to the total number of exchanges during the specified reference procurement cycle. This refers to the first [number]th [period] in the specified reference procurement cycle. The absolute value of the decrease in the compatibility index of consumable products corresponding to each exchange behavior. This refers to the average of the absolute values ​​of the decrease in the compatibility index of consumable products for all exchange transactions. This refers to the amplitude control factor, and Greater than 0.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] This invention constructs a product precision matching and recommendation generation system that integrates multi-dimensional tag recognition. Addressing the common pitfall of enterprises "blindly raising prices but reducing suitability" during product exchange processes, it proposes an identification mechanism based on the declining trend of the suitability index during the procurement cycle. Combined with a dynamic weight enhancement strategy when no effective downward trend is observed, the influence of the suitability index in the current recommendation model is precisely adjusted, thereby increasing the priority of highly suitable products in the recommendation ranking. This method not only introduces a mature multi-dimensional scoring factor and tag modeling system but also incorporates historical usage feedback and price increase trends for optimization judgment, significantly enhancing the adaptability, accuracy, and user trust of consumable recommendation results. It has good intelligent optimization effects and promotional value. Attached Figure Description

[0035] Figure 1 Application architecture diagram of the system provided in the embodiments of the present invention;

[0036] Figure 2 This is a structural block diagram of the exchange rate analysis module in the system provided in the embodiments of the present invention;

[0037] Figure 3 This is a structural block diagram of the trend judgment module in the system provided in the embodiment of the present invention;

[0038] Figure 4 This is a structural block diagram of the correction coefficient generation module in the system provided in the embodiment of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0040] Figure 1 An application architecture diagram of the system provided in an embodiment of the present invention is shown.

[0041] In another preferred embodiment of the present invention, a product accurate matching and recommendation generation system integrating multi-dimensional tag recognition includes:

[0042] The request identification module 100 is used to obtain the historical purchase data, historical usage records and current consumable product recommendation model of the corresponding consumables after the target enterprise submits a replacement request.

[0043] In this embodiment of the invention, the target enterprise mainly refers to enterprises and institutions that have a certain frequency of consumable procurement and regularly update the configuration of consumable products, such as testing laboratories, manufacturing enterprises, and other unit users with bulk consumable usage needs. These target enterprises typically initiate exchange requests through their internal ERP systems, procurement management systems, or collaborative platforms with suppliers.

[0044] The exchange request refers to the target enterprise's intention to replace or upgrade its current consumable products based on actual usage feedback, inventory status, or procurement plans. The system can obtain exchange request data in real time by connecting with the enterprise's system or directly with online shopping platforms. Information can also be collected through manual input, API calls, etc. Exchange requests typically include parameters such as the name of the product to be replaced, historical purchase price, desired price range for the replacement, usage scenario, and acceptable compatibility range.

[0045] The consumables described in this invention refer to materials that are continuously consumed, have a limited lifespan, and require periodic replenishment or replacement within the target enterprise. Specific examples include: industrial consumables such as ink cartridges, cutting fluids, and cleaning cloths used for printing, packaging, or process support in manufacturing enterprises; and testing consumables such as reagent bottles, centrifuge tubes, and filter membranes in laboratories. The types of consumables should correspond to the industry attributes of the target enterprise.

[0046] Current consumable product recommendation models are typically based on widely deployed intelligent recommendation systems. These systems comprehensively evaluate various replaceable consumable products using multiple scoring factors and generate a recommendation list based on the scoring results. These scoring factors may include, but are not limited to, price indicators, historical usage evaluations, functional attribute suitability (i.e., suitability index), procurement cycle stability, and supplier service level. This invention specifically focuses on the "suitability index" scoring factor, which reflects the degree of matching between the recommended consumables and the actual usage environment and user needs. In existing recommendation models, each scoring factor typically has a weighted weight to reflect its influence in the final comprehensive score. The recommendation system calculates a comprehensive score based on each consumable product's performance across all scoring factors, combined with the corresponding weighted weights, and then ranks and recommends several priority candidate products to the target company accordingly.

[0047] It should be noted that the technical solution of the present invention is based on the currently widely used consumable product recommendation model, and the key weight parameters are perturbed and optimized to improve its recommendation rationality and adaptability in the exchange scenario.

[0048] This invention addresses a common misconception among corporate procurement personnel regarding product replacement decisions: the assumption that "higher price equals better results." In reality, target companies often choose higher-priced products as replacements without fully understanding compatibility and suitability. However, such blind upgrades do not always yield better results and may instead lead to compatibility issues, wasted costs, or increased scrap rates. This invention establishes a dynamic correlation mechanism between the procurement cycle and the suitability index, guiding the system to adjust weight parameters based on historical performance. This generates more effective recommendations, enhancing the scientific rigor and practical value of the recommendation system.

[0049] Regarding data sources:

[0050] 1. Historical procurement data is usually obtained from the enterprise's ERP system, supply chain collaboration platform or contract management platform. The data content includes, but is not limited to: procurement time, unit price, quantity, supplier information, receipt confirmation records, and corresponding procurement cycle divisions for each consumable product.

[0051] 2. Historical usage records can be obtained from the warehouse system's inbound and outbound logs, consumable issuance ledgers, usage work order feedback, IoT device sensing data, or manual registration systems. The data content should include: the actual usage period of the consumables, usage description, usage quantity, discarded quantity and reason for discarding, user department or position identification, key evaluation feedback (such as performance stability, compatibility issues), etc.

[0052] Furthermore, the product accurate matching and recommendation generation system that integrates multi-dimensional tag recognition also includes:

[0053] The cycle analysis module 200 is used to identify several procurement cycles of consumables and determine whether a specific exchange behavior exists within each procurement cycle. The specific exchange behavior refers to the replacement operation from low-priced consumables to high-priced consumables during the consumables exchange process.

[0054] The cycle analysis module is used to identify several procurement cycles from the historical procurement data of consumables, select procurement cycles with the same time length, and ensure that the time interval between adjacent procurement cycles is less than a preset time threshold.

[0055] In this embodiment of the invention, the specific exchange behavior refers to the replacement of consumables from lower-priced to higher-priced items during the procurement process. This behavior is widespread in the consumables management practices of many enterprises, especially among procurement personnel with short tenures or those newly hired. Due to a lack of in-depth understanding of product compatibility, actual usage effects, and procurement history, these personnel often tend to believe that higher-priced products necessarily represent better performance, thus prioritizing higher-priced products as alternatives when faced with exchange decisions.

[0056] However, this price-driven exchange method easily overlooks the differences in compatibility of different consumables in specific application scenarios. In actual use, it may lead to problems such as decreased compatibility, increased waste rate, and increased unit cost, not only failing to bring the expected performance improvement but also affecting the overall efficiency of consumable usage. Therefore, this specific exchange behavior is the core technical pain point targeted by this invention, and its identification and analysis are of great significance for the dynamic optimization of the subsequent recommendation model.

[0057] In the specific implementation process, the cycle analysis module divides the procurement time series of each consumable item based on historical procurement data, identifying several procurement cycles with good continuity and consistent time structure. In this invention, a procurement cycle refers to the period during which a certain type of consumable exhibits relatively stable procurement behavior within the target enterprise, typically measured in monthly, quarterly, or actual procurement batches. For example, if a certain type of reagent consumable is procured centrally once a month, then each month constitutes a procurement cycle.

[0058] To ensure the rationality of subsequent comparisons and trend analyses, the selected procurement cycles must meet the following two conditions: First, the duration of each cycle must be consistent, i.e., cycles with the same time span should be selected to avoid data deviations caused by changes in procurement frequency; second, the time interval between adjacent procurement cycles should be less than a preset time threshold to eliminate discontinuous cycles caused by long-term gaps or abnormal procurement behaviors, thereby improving the representativeness and comparability of the analyzed behaviors.

[0059] After completing the procurement cycle segmentation, the cycle analysis module further extracts product replacement records within each procurement cycle. By comparing the unit price of each item before and after the replacement, it identifies whether there has been a replacement behavior from low-priced to high-priced products. The data required for identifying this type of price replacement path includes: the unique identifier of the purchased product, the purchase unit price, the product replacement order, and the purchase time, all of which can be extracted from historical procurement data. Therefore, the identification of the specific replacement behavior does not rely on additional data sources and can be efficiently implemented under the current data structure.

[0060] Through the above identification mechanism, this invention can effectively locate and quantify typical purchasing behaviors within target enterprises caused by "price-oriented exchange misjudgment," providing an accurate and traceable analytical basis for subsequent compatibility analysis, scrap rate assessment, and recommendation weight adjustment.

[0061] Furthermore, the product accurate matching and recommendation generation system that integrates multi-dimensional tag recognition also includes:

[0062] The Exchange Rate Analysis Module 300 is used to filter several reference procurement cycles with consistent exchange frequency and consistent exchange scale, and calculate the average price increase of all exchange behaviors in each reference procurement cycle.

[0063] Specifically, Figure 2 The diagram shows a structural block diagram of the exchange rate analysis module 300 in the system provided in an embodiment of the present invention.

[0064] In a preferred embodiment of the present invention, the exchange rate analysis module 300 specifically includes:

[0065] The exchange cycle filtering unit 301 is used to sequentially analyze each procurement cycle and extract reference procurement cycles with consistent exchange behavior frequency and consistent exchange behavior scale. The consistent exchange behavior scale means that the quantity of consumables replaced in each exchange process is the same, or the quantity variation is within a preset tolerance range.

[0066] The replacement margin calculation unit 302 is used to obtain the price increase of consumables corresponding to each replacement behavior in each reference procurement cycle based on historical procurement data, and to average the price increase of all replacement behaviors to obtain the average price increase corresponding to each reference procurement cycle.

[0067] In this embodiment of the invention, the purpose of setting up the exchange cycle screening unit 301 is to improve the comparability and analytical effectiveness of exchange behaviors between different procurement cycles. Specifically, by screening out reference procurement cycles with consistent exchange frequency and scale, the interference caused by differences in behavior patterns can be minimized in subsequent analysis, thereby more accurately assessing the correlation between price increases and usage effects (such as scrap rate).

[0068] Among the above screening criteria, consistent number of exchange transactions means that the number of exchange operations recorded within each reference procurement cycle is the same; while consistent scale of exchange transactions further requires that the quantities of consumables replaced in each exchange process be comparable. By default, consistent quantity of goods can directly reflect the consistency of the scale of transactions. However, in practical applications, to improve the adaptability of the screening, this embodiment allows for a certain degree of fluctuation in the quantity of goods, provided that the fluctuation is within the set tolerance range. In addition to changes in the quantity of consumables, the scale of exchange transactions can also be comprehensively judged in combination with other dimensions, such as: 1. The total procurement amount corresponding to the exchange operation; 2. The number of types of goods involved in the replacement; 3. The business scope affected by the exchange transaction (such as the number of applicable departments or the number of positions covered); 4. The number of order items or line items corresponding to each exchange transaction.

[0069] By integrating these factors, we can more comprehensively depict the actual scale of exchange behavior and thus select reference procurement cycles with similar behavioral structures and stronger comparative significance.

[0070] In the exchange rate calculation unit 302, the specific implementation method includes the following steps:

[0071] 1. Within each selected reference procurement cycle, identify the exchange behavior item by item and determine the unit price of the goods before and after each exchange;

[0072] 2. For each exchange, the price will be based on the unit price of the consumable item before the exchange. The unit price of consumables after replacement Calculate the price increase value ;

[0073] 3. Within each procurement cycle, calculate the price increase of all exchange transactions during that cycle, and then take their arithmetic average to obtain the average price increase for that reference procurement cycle. ;

[0074] 4. Use the average price increase value corresponding to all reference procurement cycles as the input variable for the subsequent trend judgment module to construct the increase sequence and analyze the trend of the scrap rate change.

[0075] Through the above calculation process, the present invention can construct a set of reference periods with stable structure and consistent behavioral characteristics, and quantify the overall increase level of exchange behavior in each period, providing an objective quantitative basis for subsequent judgment on whether exchange behavior brings about actual improvement in usage effect.

[0076] Furthermore, the product accurate matching and recommendation generation system that integrates multi-dimensional tag recognition also includes:

[0077] The trend judgment module 400 is used to sort the reference procurement cycles according to the average price increase, extract the scrap rate of consumables in each reference procurement cycle, and determine whether the scrap rate shows an effective downward trend as the price increase rises.

[0078] Specifically, Figure 3 The diagram shows a structural block diagram of the trend judgment module 400 in the system provided in an embodiment of the present invention.

[0079] In a preferred embodiment provided by the present invention, the trend judgment module 400 specifically includes:

[0080] The sorting processing unit 401 is used to sort the reference procurement cycles from small to large according to the average price increase of each reference procurement cycle, and generate an ordered price increase sequence.

[0081] The scrap rate extraction unit 402 is used to obtain the usage data of the corresponding consumables of the target enterprise in each reference procurement cycle based on historical usage records, and to obtain the scrap rate of consumables in the reference procurement cycle based on the ratio of scrapped quantity to input quantity.

[0082] The trend judgment unit 403 is used to analyze whether the scrap rate corresponding to each reference procurement cycle in the ordered price increase sequence shows an effective downward trend as the price increase rises.

[0083] The effective downward trend refers to the average slope of the change trend of the corresponding scrap rate sequence in the ordered increase sequence being less than a preset negative threshold.

[0084] In this embodiment of the invention, the specific implementation process of the scrap rate extraction unit 402 is based on the historical usage records stored by the target enterprise. These records typically include the actual input and scrap quantities of different consumables within each procurement cycle. To achieve automated extraction and analysis of this data, the following technical means are preferably adopted:

[0085] 1. Integrate with the company's existing consumables usage management system or ERP platform, and call the inventory inbound and outbound records, requisition registration data and usage loss records within the procurement cycle through the data interface;

[0086] 2. Take all consumables involved in each reference procurement cycle as the analysis object, and extract the cumulative input quantity and cumulative waste quantity in that cycle accordingly;

[0087] 3. Use configurable scripting tools or data processing programs to calculate the scrap rate of each type of consumable, and combine it with the weighted sum of the total input within the cycle to obtain the overall scrap rate of consumables for that procurement cycle.

[0088] 4. To ensure data quality and result stability, the extracted data can be further filtered and confirmed through outlier removal mechanisms, data synchronization time verification, and cross-departmental consistency review.

[0089] The above processing flow can be implemented based on commonly used data processing platforms, such as Python data analysis scripts with the Pandas library, or by directly using SQL statements in the database to perform data filtering and ratio calculations, ensuring that the system has good implementation and scalability.

[0090] The trend assessment unit 403 is designed to evaluate whether there is a significant positive correlation between price increases and improved performance. By analyzing the changing trends of the scrap rate corresponding to each reference procurement cycle in the ordered price increase sequence, it can be revealed whether the company has achieved the desired improvement in consumable quality through price increases.

[0091] If the average slope of this trend is less than the preset negative threshold, it indicates that as prices gradually increase, the waste rate of consumables decreases significantly. This means that the price increase effectively improves the performance of consumables; in other words, the increase in procurement costs is worthwhile. This trend is verified as an "effective downward trend," which can be regarded as objective evidence that the company's replacement decision is reasonable and the quality of consumables has improved significantly.

[0092] Conversely, if a valid downward trend is not detected, meaning the slope does not reach the preset negative threshold, or even shows no significant change or reverse growth, it indicates that the price increase has not led to a synchronous decrease in the scrap rate. This may mean:

[0093] 1. The high-priced consumables chosen by enterprises do not actually improve the degree of adaptability or compatibility; 2. The procurement behavior is affected by irrational factors such as brand orientation and subjective judgment; 3. In the actual use scenario of enterprises, the core variable affecting the scrap rate is not determined by the price of the consumables themselves.

[0094] Furthermore, the product accurate matching and recommendation generation system that integrates multi-dimensional tag recognition also includes:

[0095] The correction coefficient generation module 500 is used to extract a specified reference procurement period that is consistent with the price increase of the current exchange request when no effective downward trend is detected. It obtains the decline of the consumable product adaptability index for each exchange behavior, generates an adaptability weight amplification coefficient, and enhances the weight of the adaptability index in the current consumable product recommendation model to improve the priority of consumable products with high adaptability in the recommendation ranking.

[0096] Specifically, Figure 4 The diagram shows a structural block diagram of the correction coefficient generation module 500 in the system provided in an embodiment of the present invention.

[0097] In a preferred embodiment provided by the present invention, the correction coefficient generation module 500 specifically includes:

[0098] The exchange request parsing unit 501 is used to parse the current exchange request and determine its corresponding expected price increase when no effective downward trend is detected.

[0099] The reference cycle matching unit 502 is used to filter out a specified reference procurement cycle from historical data whose average price increase is consistent with the expected price increase of the current replacement request, and extract the adaptability index data related to the consumable usage effect corresponding to each replacement behavior in the specified reference procurement cycle, and determine the decrease of the adaptability index before and after the replacement.

[0100] The amplification factor calculation unit 503 is used to generate the fit degree weight amplification factor based on the average of the absolute values ​​of the decrease in the fit index of all exchange behaviors.

[0101] The weight enhancement execution unit 504 is used to obtain the initial weights assigned to the adaptability index in the current consumable product recommendation model, and to enhance and adjust them based on the adaptability weight amplification coefficient.

[0102] The ranking update unit 505 is used to apply the enhanced and adjusted weighted weights to the current recommendation scoring model, re-ranking candidate consumables to increase the recommendation priority of consumables with higher suitability indices. After enhancing and correcting the initial weighted weights of the suitability index based on the suitability weight amplification coefficient, the weighted weights of other scoring factors already set in the current recommendation model should be reduced by an equal amount to maintain the overall balance between the weighted weights and avoid imbalance in the recommendation results due to the increase of a single factor weight.

[0103] The adaptability index data refers to the multi-dimensional tag scoring results constructed based on dimensions such as functional attributes, usage habits, and compatibility.

[0104] The decrease in the compatibility index before and after the exchange is obtained by calculating the difference in the average score of the consumable before and after the exchange on the multidimensional label scoring results, or by the ratio of the relative deviation between the two.

[0105] In this embodiment of the invention, the correction coefficient generation module 500 is a key component of the system, designed to address the core technical pain points encountered by target enterprises during the product exchange process: when enterprise purchasing personnel blindly replace products with high-priced consumables based on subjective experience or price preferences, and the actual usage effect (such as the scrap rate) does not improve significantly, the system can identify such situations and make adaptation-oriented corrections to the recommendation model to improve the rationality and scientific nature of the recommendation results.

[0106] The activation prerequisite for this module is "no effective downward trend detected," meaning the trend judgment module fails to detect a significant decrease in the waste rate of consumables as prices rise. This is precisely the problem situation that this case focuses on. In this situation, the price increase fails to bring about an improvement in value, indicating that the replacement behavior lacks rational decision-making support based on suitability. Therefore, a remedial correction mechanism needs to be introduced.

[0107] The exchange request parsing unit 501 is responsible for obtaining the exchange request from the current target enterprise and determining the expected price increase. This increase can be automatically calculated based on the ratio between the current pre-exchange price of consumables recorded in the enterprise's procurement system and the price of the target product selected by the user. This price increase is deterministic data with a clear numerical expression and can be used as a reference standard for subsequent adjustment and matching.

[0108] Based on this, the reference cycle matching unit 502 filters out reference procurement cycles with similar price increase levels from historical data. The significance of this filtering is that it extracts real cases from the company's own past replacement cycles with similar price change magnitudes to evaluate the adaptation performance after replacing consumables under similar conditions, thereby establishing a data-supported correction baseline.

[0109] Within these designated reference procurement cycles, the system extracts the compatibility index data corresponding to each exchange behavior. This index is derived from the multi-dimensional product label scoring method widely used in existing technologies, specifically including but not limited to the following three key dimensions:

[0110] 1. Functional attributes: Whether the basic functions of the consumables meet the usage requirements of the corresponding equipment or scenario;

[0111] 2. Usage habits: Does this consumable match the usage preferences and maintenance methods of the operators in the company?

[0112] 3. Compatibility: Are there any compatibility issues when the consumables are used in the target device, such as the interface and the interface wear rate?

[0113] The adaptability index is a fusion score of the aforementioned multi-dimensional tags. Its generation typically relies on historical user feedback data, product specification documents, and expert evaluation models, and it participates in the overall scoring within the recommendation system using a weighted approach. The generation mechanism of this type of index is a mature existing technology, usually based on algorithms such as natural language processing, tag vector modeling, user behavior analysis, and multi-factor scoring fusion. It can structurally integrate information from heterogeneous sources and assign quantitative evaluation values. It is already widely used in e-commerce recommendations, digital device replacement suggestions, and intelligent consumable recommendations, possessing good reusability and a solid foundation for implementation.

[0114] The amplification factor calculation unit 503 generates an adaptation weight amplification factor based on the average absolute value of the change in the adaptability index before and after the replacement. The core function of this factor is to quantify "the actual degree of deterioration in adaptability in similar replacement behaviors". If there is a significant decline in the adaptability index in multiple reference periods, it indicates that under similar price increases, consumable replacement often sacrifices adaptability, thus requiring emphasis on the adaptability index in the current recommendation model.

[0115] The weight enhancement execution unit 504 obtains the original weights allocated to the fit index in the current recommendation model and enhances them. The enhancement function is:

[0116] ;

[0117] in, This refers to the initial weighting of the adaptation index in the current product and consumable recommendation model. This refers to the corrected weighted average. This refers to the total number of exchanges during the specified reference procurement cycle. This refers to the first [number]th [period] in the specified reference procurement cycle. The absolute value of the decrease in the compatibility index of consumable products corresponding to each exchange behavior. This refers to the average of the absolute values ​​of the decrease in the compatibility index of consumable products for all exchange transactions. This refers to the amplitude control factor, and Greater than 0.

[0118] This approach has three advantages: 1. It avoids the system failing to detect the decline in usability caused by low compatibility; 2. It strengthens the weight of compatibility in the overall score, making the recommendations closer to actual usage needs; 3. It provides objective decision-making assistance for procurement personnel who have not yet accumulated sufficient experience.

[0119] The ranking update unit 505 updates the enhanced weighted weights to the current recommendation model in real time, recalculates the recommendation scores of all candidate consumables, and prioritizes them based on the new scores. For example, suppose in a certain exchange request, the suitability indices of candidate products A, B, and C are 90, 80, and 70, respectively. Before participating in the recommendation model calculation, these indices are normalized to 0.9, 0.8, and 0.7, with an initial weight of 0.2. The combined scores of other scoring factors are: Product A = 0.6, B = 0.65, and C = 0.7. If the suitability index weight is increased to 0.35 after enhancement, the new total scores might be: A = 0.6 + 0.35 × 0.9 = 0.915, B = 0.65 + 0.35 × 0.8 = 0.93, and C = 0.7 + 0.35 × 0.7 = 0.945. Although product C has a higher price, its recommendation priority will decrease accordingly because its suitability is significantly worse than products A and B.

[0120] To prevent the model from becoming unbalanced due to an increase in the weight of a single factor, the system also performs a weighted weight balancing adjustment mechanism after the weight is increased. This means that the weights of other scoring factors in the model are reduced proportionally to maintain the overall balance of the recommendation model.

[0121] In summary, this module effectively suppresses irrational price-driven exchange behavior by referencing period analysis, introducing suitability indicators, and weighted dynamic adjustments, thereby improving the actual adaptability and intelligence level of the recommendation system.

[0122] The enhancement function provided by this invention is an intuitive and effective weighted adjustment method that dynamically enhances the initial weights based on the average decrease in the fit index, thereby increasing the recommendation priority of highly fit consumable products. However, it should be noted that this function is not limited to a single form. In practical applications, various methods such as weighted average, weighted median, standard deviation correction, and piecewise mapping can be used to calculate the fit weight amplification coefficient based on the individual values, distribution characteristics, or extreme values ​​of the fit index in each exchange behavior. This further enhances the model's responsiveness to abnormal situations or specific exchange patterns, demonstrating strong scalability and adaptability.

[0123] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0124] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0125] 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.

[0126] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0127] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A product accurate matching and recommendation generation system integrating multi-dimensional tag recognition, characterized in that, The system includes a request identification module, a periodic analysis module, a return rate analysis module, a trend judgment module, and a correction coefficient generation module, wherein: The request identification module is used to obtain the historical purchase data, historical usage records and current consumable product recommendation model of the corresponding consumables after the target enterprise submits a replacement request; The cycle analysis module is used to identify several procurement cycles of consumables and determine whether there are specific replacement behaviors within each procurement cycle; The exchange rate analysis module is used to filter several reference procurement cycles with consistent exchange frequency and consistent exchange scale, and calculate the average price increase of all exchange behaviors in each reference procurement cycle. The trend judgment module is used to sort the reference procurement cycles according to the average price increase, extract the scrap rate of consumables in each reference procurement cycle, and determine whether the scrap rate shows an effective downward trend as the price increase rises. The correction coefficient generation module is used to extract a specified reference procurement period that is consistent with the price increase of the current exchange request when no effective downward trend is detected. It obtains the decline of the consumable product adaptability index for each exchange behavior, generates an adaptability weight amplification coefficient, and enhances the weight of the adaptability index in the current consumable product recommendation model to improve the priority of consumable products with high adaptability in the recommendation ranking.

2. The product accurate matching and recommendation generation system integrating multi-dimensional tag recognition as described in claim 1, characterized in that, The specific exchange behavior refers to the replacement operation of low-priced consumables with high-priced consumables during the consumables exchange process.

3. The product accurate matching and recommendation generation system integrating multi-dimensional tag recognition according to claim 2, characterized in that, The cycle analysis module is used to identify several procurement cycles from the historical procurement data of consumables, select procurement cycles with the same time length, and ensure that the time interval between adjacent procurement cycles is less than a preset time threshold.

4. The product accurate matching and recommendation generation system integrating multi-dimensional tag recognition according to claim 2, characterized in that, The exchange rate analysis module specifically includes: The exchange cycle filtering unit is used to parse each procurement cycle in sequence and extract reference procurement cycles with consistent exchange behavior frequency and consistent exchange behavior scale. The consistent exchange behavior scale means that the quantity of consumables replaced in each exchange process is the same, or the quantity variation is within a preset tolerance range. The replacement margin calculation unit is used to obtain the price increase of consumables for each replacement behavior in each reference procurement cycle based on historical procurement data, and to average the price increase of all replacement behaviors to obtain the average price increase for each reference procurement cycle.

5. The product accurate matching and recommendation generation system integrating multi-dimensional tag recognition according to claim 2, characterized in that, The trend judgment module specifically includes: The sorting processing unit is used to sort the reference procurement cycles from smallest to largest according to the average price increase of each reference procurement cycle, and generate an ordered price increase sequence. The scrap rate extraction unit is used to obtain the usage data of the corresponding consumables of the target enterprise in each reference procurement cycle based on historical usage records, and to obtain the scrap rate of consumables in the reference procurement cycle based on the ratio of scrapped quantity to input quantity. The trend judgment unit is used to analyze whether the scrap rate corresponding to each reference procurement cycle in the ordered price increase sequence shows an effective downward trend as the price increase rises.

6. The product accurate matching and recommendation generation system integrating multi-dimensional tag recognition according to claim 5, characterized in that, The effective downward trend refers to the average slope of the change trend of the corresponding scrap rate sequence in the ordered increase sequence being less than a preset negative threshold.

7. The product accurate matching and recommendation generation system integrating multi-dimensional tag recognition according to claim 5, characterized in that, The correction coefficient generation module specifically includes: The exchange request parsing unit is used to parse the current exchange request and determine its corresponding expected price increase when no effective downward trend is detected. The reference cycle matching unit is used to filter out a specified reference procurement cycle from historical data whose average price increase is consistent with the expected price increase of the current replacement request, and extract the compatibility index data related to the use effect of consumables for each replacement behavior in the specified reference procurement cycle, and determine the decrease of the compatibility index before and after the replacement. The amplification factor calculation unit is used to generate the fit weight amplification factor based on the average of the absolute values ​​of the decrease in the fit index of all exchange behaviors. The weight enhancement execution unit is used to obtain the initial weights assigned to the adaptability index in the current consumable product recommendation model, and to enhance and adjust them based on the adaptability weight amplification coefficient. The ranking update unit is used to apply the enhanced and adjusted weighted weights to the current recommendation scoring model, re-rank the candidate consumable products, and thus improve the recommendation priority of consumable products with higher suitability index.

8. The product accurate matching and recommendation generation system integrating multi-dimensional tag recognition according to claim 7, characterized in that, The adaptability index data refers to the multi-dimensional tag scoring results constructed based on dimensions such as functional attributes, usage habits, and compatibility. The decrease in the compatibility index before and after the exchange is obtained by calculating the difference in the average score of the consumable before and after the exchange on the multidimensional label scoring results, or by the ratio of the relative deviation between the two.

9. The product accurate matching and recommendation generation system integrating multi-dimensional tag recognition according to claim 7, characterized in that, After enhancing and correcting the initial weighting of the fit index based on the fit weight amplification factor, the weighting of other scoring factors already set in the current recommendation model should be reduced by the same amount to maintain the overall balance between the weightings and avoid imbalance in the recommendation results due to the increase of the weight of a single factor.

10. The product accurate matching and recommendation generation system integrating multi-dimensional tag recognition according to claim 7, characterized in that, When enhancing and adjusting the fitness weight amplification coefficient, a preset enhancement function is used, which is: ； in, This refers to the initial weighting of the adaptation index in the current product and consumable recommendation model. This refers to the corrected weighted average. This refers to the total number of exchanges during the specified reference procurement cycle. This refers to the first [number]th [period] in the specified reference procurement cycle. The absolute value of the decrease in the compatibility index of consumable products corresponding to each exchange behavior. This refers to the average of the absolute values ​​of the decrease in the compatibility index of consumable products for all exchange transactions. This refers to the amplitude control factor, and Greater than 0.

Images