Purchase management method and system based on business big data

Abstract

The invention discloses a purchase management method and system based on business big data, and relates to the technical field of purchase prediction and supplier management, and the method comprises the steps: collecting historical purchase records of a business database, carrying out the preprocessing, generating a preprocessing data set, and constructing and training a purchase prediction model through a long short-term memory (LSTM) network algorithm; the purchase demand is predicted through the purchase prediction model, a demand prediction result is output, and the suppliers are graded through the demand prediction result to form graded suppliers; and generating a purchase suggestion list through the graded suppliers, and sending the purchase suggestion list to the purchase management terminal. According to the invention, the purchase prediction model based on the multilayer LSTM neural network is established, so that the long and short term characteristics of the purchase data are effectively captured and accurately predicted; and in combination with the purchase decision matrix and the differentiation processing strategy, the system automatically generates purchase suggestions and pushes the purchase suggestions to the management terminal through the encryption channel, so that the automation level and the operation efficiency of enterprise purchase management are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of procurement forecasting and supplier management, and in particular to a procurement management method and system based on business big data. Background Art

[0002] Traditional procurement management systems mainly rely on the basic functional modules provided by the ERP system, and use manual experience to perform demand forecasting and supplier management. In recent years, with the rapid development of artificial intelligence technology, especially the breakthrough progress of deep learning algorithms in the field of time series forecasting, new technical support has been provided for the intelligent upgrade of procurement management. The procurement forecasting methods currently commonly used in the industry include traditional statistical methods such as moving average method and exponential smoothing method, as well as forecasting methods based on machine learning algorithms such as BP neural network and support vector machine. These methods have improved the accuracy of procurement forecasting to a certain extent, but there are still obvious limitations in dealing with long-term dependencies and dynamic time series characteristics.

[0003] The existing procurement management systems generally have the following technical problems: First, the procurement demand forecasting model does not adequately capture the long-term dependency characteristics and short-term volatility characteristics of historical data, resulting in unstable forecasting accuracy; second, the supplier grading evaluation relies too much on a static indicator system and lacks a dynamic correlation mechanism with the procurement forecast accuracy; third, the generation logic of procurement recommendations is relatively simple and fails to fully combine the supplier grading results for intelligent decision-making optimization. These technical problems lead to inefficient procurement management, insufficient supply chain coordination, and poor procurement cost control. Therefore, it is urgent to develop an intelligent procurement management method based on deep learning to improve procurement forecast accuracy and supplier management efficiency.

[0004] In response to the above problems, the present invention provides a procurement management method and system based on business big data. The present invention constructs a procurement forecasting model by introducing a long short-term memory network (LSTM) algorithm, establishes a dynamic association mechanism between forecast accuracy and supplier grading, realizes the coordinated optimization of procurement forecasting and supplier management, and effectively solves the technical problems in the prior art such as unstable procurement forecasting accuracy and lack of dynamic adjustment mechanism for supplier grading. Summary of the invention

[0005] In view of the fact that the existing procurement management system has the problems of insufficient model capture capability and unstable prediction accuracy in demand forecasting, and a single evaluation mechanism, static grading standards and lack of intelligent procurement decision support in supplier management, the present invention is proposed.

[0006] Therefore, the problem to be solved by the present invention is how to realize the dynamic association between supplier classification and prediction accuracy through a procurement forecasting model, and provide intelligent procurement suggestions based on this, thereby improving the automation level and decision-making efficiency of enterprise procurement management.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In the first aspect, an embodiment of the present invention provides a procurement management method based on business big data, which includes collecting historical procurement records from a business database, and preprocessing the historical procurement records to generate a preprocessed data set; based on the preprocessed data set, building and training a procurement prediction model using a long short-term memory network LSTM algorithm; predicting procurement demand through the procurement prediction model, outputting demand prediction results, and grading suppliers through the demand prediction results to form graded suppliers; generating a procurement recommendation list through the graded suppliers, and sending it to a procurement management terminal.

[0009] As a preferred solution of the procurement management method based on business big data described in the present invention, wherein: the method for generating the procurement suggestion list is to establish a procurement decision matrix based on the demand forecast results and the supplier grading processing results; adopt a weighted calculation method to prioritize the procurement suggestion list in the procurement decision matrix; if the priority score of the procurement suggestion list is greater than a first preset threshold, the procurement demand in this list is marked as an emergency procurement item; if the priority score of the procurement suggestion list is less than or equal to the first preset threshold and greater than a second preset threshold, the procurement demand in this list is marked as a regular procurement item; if the priority score of the procurement suggestion list is less than or equal to the second preset threshold, the procurement demand in this list is marked as a planned procurement item; for the emergency procurement item, a price inquiry request is sent to the first-level supplier, and when there is a continuous supplier responding, the price comparison process is started; for the regular procurement item, suppliers with similar commodity supply records are screened from the first-level suppliers and the second-level suppliers to generate a recommended supplier list; for the planned procurement item, similar commodity demands are integrated and a centralized procurement time window is set; the marked procurement suggestion list is sent to the procurement management terminal through an encrypted channel, and a reminder notification is sent to the procurement personnel in real time.

[0010] As a preferred solution of the procurement management method based on business big data described in the present invention, the method for forming the hierarchical suppliers is to input the historical procurement records into the procurement forecasting model to generate demand forecasting results for several time periods, wherein the demand forecasting results include procurement categories, procurement quantities and procurement time; perform deviation analysis on the demand forecasting results and actual procurement records to calculate the forecasting accuracy, and the specific formula is as follows:

[0011]

[0012] Among them, ACC is the prediction accuracy, n is the number of prediction cycles, and w i is the time weight coefficient of the i-th prediction period, is the predicted purchase quantity for the i-th period, is the actual purchase quantity in the ith period, ΔT i is the time deviation days of the i-th cycle, k is the number of category errors, C j is the forecast deviation cost coefficient of the jth category, M is the benchmark cost parameter, and α is the category penalty factor.

[0013] The deviation analysis includes: when the prediction accuracy rate is greater than the prediction threshold for multiple consecutive times, the supplier is graded; when the prediction accuracy rate is less than or equal to the prediction threshold for multiple consecutive times, the procurement prediction model is optimized.

[0014] As a preferred solution of the procurement management method based on business big data described in the present invention, the grading process includes the following steps: establishing a supplier rating system based on the supplier's historical transaction data, wherein the supplier rating system includes the on-time delivery rate, product quality qualification rate and price competitiveness index; using a weighted average algorithm to calculate the supplier's comprehensive score, and dividing the suppliers into first-tier suppliers, second-tier suppliers and third-tier suppliers according to the score range, wherein the first-tier suppliers are given priority in allocating purchase orders, the second-tier suppliers are used as alternative suppliers, and the third-tier suppliers need to formulate improvement plans; in the supplier grading process, a dynamic evaluation mechanism is introduced to regularly update the supplier ratings, and adjust the supplier grades according to changes in the supplier ratings.

[0015] As a preferred solution of the procurement management method based on business big data described in the present invention, wherein: the construction method of the procurement prediction model is to group the preprocessed data set according to the commodity code, and divide all historical data into a training set and a verification set in proportion; adopt a multi-layer LSTM neural network structure, wherein the neurons of the input layer establish a corresponding relationship with the feature indicators, the hidden layer enhances the feature extraction capability through a multi-layer architecture design, and the output layer designs a number of prediction time windows, which correspond to the demand prediction results of several time periods respectively; determine the time step of the long short-term memory network LSTM algorithm according to the feature indicators and business needs, and construct a procurement prediction model; adopt an adaptive optimization algorithm to train the procurement prediction model, set the initial learning rate parameter, and dynamically adjust the learning rate according to the changing trend of the loss function during the training process; evaluate and verify the demand prediction results of the procurement prediction model, and when the evaluation index reaches the preset target requirements, it is confirmed that the prediction performance of the procurement prediction model meets the actual application needs.

[0016] As a preferred solution of the procurement management method based on business big data described in the present invention, the characteristic indicators include procurement quantity, procurement frequency, procurement unit price, inventory level, supplier delivery time, market price index and seasonal index.

[0017] As a preferred solution of the procurement management method based on business big data described in the present invention, historical procurement records of a business database are collected, and the historical procurement records are preprocessed to generate a preprocessed data set, including: grouping the historical procurement records according to the procurement date, calculating the total procurement amount and average unit price of each commodity in each group, and obtaining the procurement quantity change trend and price fluctuation characteristics of the commodity, wherein the historical procurement records include commodity identification, procurement quantity, procurement price, procurement date and supplier information; marking the historical procurement records of the procurement quantity change trend by a preset fluctuation threshold, wherein when the procurement quantity change trend value is greater than the preset fluctuation threshold, this group of procurement commodities is marked as fluctuating procurement commodities; when the procurement quantity change trend value is less than or equal to the preset fluctuation threshold, this procurement commodity is marked as stable procurement commodity; the procurement features of the fluctuating procurement commodities and the stable procurement commodities are categorized by the K-means clustering algorithm, and the procurement features and category information are stored in the preprocessed data set in the form of a structured data table.

[0018] In the second aspect, an embodiment of the present invention provides a procurement management system based on business big data, which includes: a preprocessing module, which is used to collect historical procurement records from a business database, and preprocess the historical procurement records to generate a preprocessed data set; a construction module, which, based on the preprocessed data set, uses a long short-term memory network (LSTM) algorithm to construct and train a procurement prediction model; a grading processing module, which is used to predict procurement demand through the procurement prediction model, output demand prediction results, and grade suppliers through the demand prediction results to form graded suppliers; a generation module, which is used to generate a procurement recommendation list through the graded suppliers and send it to a procurement management terminal.

[0019] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the procurement management method based on business big data as described in the first aspect of the present invention are implemented.

[0020] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the procurement management method based on business big data as described in the first aspect of the present invention are implemented.

[0021] The beneficial effects of the present invention are as follows: the present invention establishes a procurement prediction model based on a multi-layer LSTM neural network through structured preprocessing and feature extraction of historical procurement records, thereby effectively capturing and accurately predicting the long-term and short-term characteristics of procurement data; a dynamic association mechanism between prediction accuracy and supplier grading is introduced, and scientific grading management of suppliers is achieved through a scoring system and a weighted algorithm; combined with a procurement decision matrix and a differentiated processing strategy, the system automatically generates procurement recommendations and pushes them to the management terminal through an encrypted channel, effectively improving the automation level and operational efficiency of enterprise procurement management. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0023] Figure 1 This is a flowchart of a procurement management method based on business big data in Example 1;

[0024] Figure 2 This is a computer equipment diagram of Example 2, a procurement management method and system based on business big data. DETAILED DESCRIPTION

[0025] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0026] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0027] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0028] Example 1

[0029] Reference Figure 1 , which is the first embodiment of the present invention, and provides a procurement management method based on business big data, including:

[0030] S1: Collect historical purchase records from a business database, and preprocess the historical purchase records to generate a preprocessed data set.

[0031] Specifically, historical purchase records are collected by connecting to the business database through the standard database interface JDBC; the historical purchase records are grouped according to the purchase date, the total purchase volume and average unit price of each commodity in each group are calculated, and the purchase quantity change trend and price fluctuation characteristics of the commodities are obtained.

[0032] It should be noted that historical purchase records include product identification, purchase quantity, purchase price, purchase date and supplier information.

[0033] The historical purchase records of the purchase quantity change trend are further marked by a preset fluctuation threshold, wherein when the purchase quantity change trend value is greater than the preset fluctuation threshold, this group of purchased goods is marked as fluctuating purchase goods; when the purchase quantity change trend value is less than or equal to the preset fluctuation threshold, this group of purchased goods is marked as stable purchase goods.

[0034] It should be noted that the preset fluctuation threshold is a dynamic reference value determined based on the coefficient of variation (CV) of the commodity purchase quantity and the time series fluctuation characteristics. Specifically, reasonable fluctuation judgment standards are set by analyzing the ratio of the standard deviation of the commodity quantity in the historical purchase data to the average value, combined with seasonal factors and industry characteristics.

[0035] Furthermore, the purchasing characteristics of fluctuating purchasing commodities and stable purchasing commodities are subdivided into categories through the K-means clustering algorithm, and the purchasing characteristics and category information are stored in the preprocessed data set in the form of a structured data table.

[0036] S2: Based on the preprocessed data set, a procurement prediction model is constructed and trained using a long short-term memory network (LSTM) algorithm.

[0037] Specifically, the method for constructing the procurement forecasting model is to group the preprocessed data set according to the commodity code, and divide all historical data into training set and validation set in proportion; a multi-layer LSTM neural network structure is adopted, in which the neurons in the input layer establish a corresponding relationship with the feature indicators, the hidden layer enhances the feature extraction capability through a multi-layer architecture design, and the output layer designs several prediction time windows, which correspond to the demand forecast results of several time periods.

[0038] It should be noted that the characteristic indicators include purchase quantity, purchase frequency, purchase unit price, inventory level, supplier delivery time, market price index and seasonal index.

[0039] Furthermore, the time step of the long short-term memory network (LSTM) algorithm is determined according to the characteristic indicators and business needs, and a procurement forecasting model is constructed. The procurement forecasting model is trained using an adaptive optimization algorithm, the initial learning rate parameters are set, and the learning rate is dynamically adjusted during the training process according to the changing trend of the loss function.

[0040] Furthermore, the demand forecast results of the procurement forecast model are evaluated and verified. When the evaluation indicators reach the preset target requirements, it is confirmed that the forecast performance of the procurement forecast model meets the actual application needs.

[0041] S3: Predicting procurement demand through the procurement forecasting model, outputting demand forecasting results, and grading suppliers based on the demand forecasting results to form graded suppliers.

[0042] Specifically, the method for forming a hierarchical supplier is to input historical procurement records into a procurement forecasting model to generate demand forecast results for several time periods, where the demand forecast results include procurement categories, procurement quantities, and procurement time; perform deviation analysis on the demand forecast results and actual procurement records, and calculate the forecast accuracy. The specific formula is as follows:

[0043]

[0044] Among them, ACC is the prediction accuracy, n is the number of prediction cycles, and w i is the time weight coefficient of the i-th prediction period, is the predicted purchase quantity for the i-th period, is the actual purchase quantity in the ith period, ΔT i is the time deviation days of the i-th cycle, k is the number of category errors, C j is the forecast deviation cost coefficient of the jth category, M is the benchmark cost parameter, and α is the category penalty factor.

[0045] Furthermore, the deviation analysis includes grading the suppliers when the prediction accuracy is greater than the prediction threshold for multiple consecutive times; and optimizing the procurement prediction model when the prediction accuracy is less than or equal to the prediction threshold for multiple consecutive times.

[0046] It should be noted that the prediction threshold is a critical value determined based on multi-dimensional factors such as historical procurement data analysis, business objective requirements and supply chain risk control.

[0047] Furthermore, the grading process includes the following steps: Based on the supplier's historical transaction data, a supplier rating system is established, where the supplier rating system includes the on-time delivery rate, product quality qualification rate and price competitiveness index. The weighted average algorithm is used to calculate the supplier's comprehensive score, and the suppliers are divided into first-tier suppliers, second-tier suppliers and third-tier suppliers according to the score range, where first-tier suppliers are given priority in allocating purchase orders, second-tier suppliers are used as alternative suppliers, and third-tier suppliers need to formulate improvement plans; in the supplier grading process, a dynamic evaluation mechanism is introduced to regularly update the supplier rating, and the supplier level is adjusted according to the changes in the supplier rating.

[0048] S4: Generate a purchase suggestion list through the graded suppliers and send it to the purchase management terminal.

[0049] Specifically, the method for generating the purchase suggestion list is to establish a purchase decision matrix based on the demand forecast results and the supplier classification processing results; and to prioritize the purchase suggestion list in the purchase decision matrix using a weighted calculation method. The specific formula is as follows:

[0050]

[0051] Among them, S is the final priority score of the procurement project, f i is the calculation function of each evaluation index, w i is the weight coefficient of each evaluation index, D is the demand urgency index, U is the supplier grade coefficient, I is the inventory level index, and B is the cost index.

[0052] Furthermore, if the priority score of the procurement suggestion list is greater than the first preset threshold, the procurement requirements in this list are marked as urgent procurement items; if the priority score of the procurement suggestion list is less than or equal to the first preset threshold and greater than the second preset threshold, the procurement requirements in this list are marked as regular procurement items; if the priority score of the procurement suggestion list is less than or equal to the second preset threshold, the procurement requirements in this list are marked as planned procurement items.

[0053] It should be noted that the first preset threshold is a high-priority critical value determined based on historical procurement data statistics, which is used to identify procurement projects that require urgent processing. The second preset threshold is a distinguishing boundary set by analyzing the characteristics of regular procurement processes and planned procurement, and is used to classify regular procurement from procurement projects that can be planned for the long term.

[0054] Furthermore, for urgent purchase items, a price inquiry request is sent to the first-tier suppliers. When a continuous supplier responds, the price comparison process is initiated. For regular purchase items, suppliers with a record of supplying similar goods are screened from the first-tier and second-tier suppliers to generate a list of recommended suppliers. For planned purchase items, the demand for similar goods is integrated and a centralized procurement time window is set.

[0055] Specifically, the price comparison process achieves the best decision in supplier selection through a systematic evaluation mechanism; collects standard quotations from responding suppliers and requires them to provide complete business terms including price, delivery date and payment method within 24 hours; establishes a price comparison matrix, unifies the pricing unit and considers various additional costs, and uses a weighted scoring method to conduct a comprehensive evaluation of suppliers, in which the quoted price accounts for 40%, the delivery cycle accounts for 30%, the payment terms account for 20%, and the historical performance accounts for 10%; selects the supplier with the best comprehensive score based on the evaluation results, and initiates a bargaining procedure of no more than 10% of the original quotation when necessary, and finally forms a procurement decision after approval by the department head. This process ensures that the evaluation and decision are completed within 48 hours, ensuring the quality of procurement while meeting the timeliness requirements of urgent needs.

[0056] Furthermore, the marked purchase suggestion list is sent to the purchase management terminal through an encrypted channel, and reminder notifications are sent to purchase personnel in real time.

[0057] Furthermore, the present embodiment also provides a procurement management system based on business big data, including: a preprocessing module, which is used to collect historical procurement records from a business database, and preprocess the historical procurement records to generate a preprocessed data set; a construction module, which, based on the preprocessed data set, uses a long short-term memory network (LSTM) algorithm to construct and train a procurement prediction model; a grading processing module, which is used to predict procurement demand through the procurement prediction model, output demand prediction results, and grade suppliers through the demand prediction results to form graded suppliers; a generation module, which is used to generate a procurement recommendation list through the graded suppliers and send it to a procurement management terminal.

[0058] In summary, the present invention establishes a procurement prediction model based on a multi-layer LSTM neural network through structured preprocessing and feature extraction of historical procurement records, thereby effectively capturing and accurately predicting the long-term and short-term characteristics of procurement data; a dynamic association mechanism between prediction accuracy and supplier grading is introduced, and scientific grading management of suppliers is achieved through a scoring system and a weighted algorithm; combined with the procurement decision matrix and differentiated processing strategy, the system automatically generates procurement recommendations and pushes them to the management terminal through an encrypted channel, effectively improving the automation level and operational efficiency of enterprise procurement management.

[0059] Example 2

[0060] Reference Figure 2 , is an embodiment of the present invention, which is different from the previous embodiment in that:

[0061] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0062] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0063] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0064] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0065] Example 3

[0066] Referring to Table 1, which is the third embodiment of the present invention, this embodiment provides a procurement management method based on business big data. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0067] Specifically, the raw material procurement data of a large manufacturing enterprise was selected for verification; historical procurement records were extracted from the Oracle business database through the JDBC interface, including procurement information of 15 key raw materials, totaling 3650 records; when preprocessing the procurement data, the Python data analysis library pandas was used for data cleaning, outliers and duplicate records were removed, and the data format was standardized; by calculating the coefficient of variation (CV) of the total monthly procurement volume, the fluctuation threshold was set to 0.35, and the commodities with CV>0.35 were marked as fluctuating procurement categories, and those with CV≤0.35 were marked as stable procurement categories; the K-means clustering algorithm (k=4) was used to segment the commodities, and the preprocessed data set was divided into training set and validation set in a ratio of 7:3.

[0068] Furthermore, in the multi-layer LSTM neural network structure, a network structure including 1 input layer, 2 hidden layers and 1 output layer is designed, and each hidden layer contains 128 neurons; the input features include: 7 dimensional indicators such as purchase quantity, purchase frequency, purchase unit price, inventory level, supplier delivery time, market price index and seasonal index in the previous 12 months; the Adam optimizer is used, the initial learning rate is set to 0.001, the batch_size is set to 32, and the training rounds are 200 rounds; through the dynamic evaluation mechanism, when the prediction accuracy exceeds 85% for 5 consecutive times, the supplier grading process is triggered; when it is lower than 85%, the model parameters are automatically adjusted for optimization.

[0069] Furthermore, the supplier rating adopts a comprehensive weighting method, in which the weight of timely delivery rate is 0.4, the weight of product quality qualification rate is 0.35, and the weight of price competitiveness index is 0.25. Suppliers are divided into three levels according to the scores: 90 points or more are first-level suppliers, 75-90 points are second-level suppliers, and below 75 points are third-level suppliers.

[0070] Specifically, as shown in Table 1, by comparing the experimental data, it can be seen that the scheme of the present invention has achieved significant improvements in various key indicators, and the procurement forecast accuracy has increased from 75.3% to 92.8%, an increase of 17.5%, thanks to the efficient prediction ability of the LSTM deep learning algorithm. The procurement process processing time was shortened from 48.6 hours to 12.4 hours, and the optimization effect reached 74.5%, reflecting the advantages of the present invention in process automation.

[0071] Table 1 Comparison between the present invention and the traditional method

[0072] Evaluation Metrics Prior art Solution of the present invention Optimization effect Procurement forecast accuracy (%) 75.3 92.8 17.5% improvement Procurement process processing time (h) 48.6 12.4 74.5% shorter Supplier response time (h) 36.2 8.6 76.2% shorter Procurement cost reduction rate (%) 6.8 15.2 8.4% increase Supply chain risk identification rate (%) 72.5 93.6 21.1% increase

[0073] Furthermore, the supplier response time was greatly shortened from 36.2 hours to 8.6 hours, and the efficiency was improved by 76.2%, which was attributed to the effective implementation of the supplier hierarchical management mechanism. The procurement cost reduction rate increased by 8.4 percentage points to 15.2%, showing the outstanding effectiveness of the invention in cost control. The supply chain risk identification rate increased from 72.5% to 93.6%, an increase of 21.1%, indicating that the prediction model of the invention has strong risk prevention and control capabilities. The procurement plan execution rate increased to 94.7%, an increase of 16.3% over the prior art, fully verifying the feasibility and effectiveness of the invention in practical applications.

[0074] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A procurement management method based on business big data, characterized by: include, Collecting historical purchase records from a business database, and preprocessing the historical purchase records to generate a preprocessed data set; Based on the preprocessed data set, a procurement prediction model is constructed and trained using a long short-term memory network (LSTM) algorithm; Predicting procurement demand through the procurement forecasting model, outputting demand forecasting results, and grading suppliers through the demand forecasting results to form graded suppliers; A purchase suggestion list is generated through the graded suppliers and sent to the purchase management terminal.

2. The procurement management method based on business big data according to claim 1, characterized in that: The method for generating the purchase suggestion list is as follows: Establishing a purchasing decision matrix based on the demand forecasting results and the supplier grading processing results; Using a weighted calculation method to prioritize the purchase suggestion list in the purchase decision matrix; If the priority score of the purchase suggestion list is greater than the first preset threshold, the purchase demand in the list is marked as an urgent purchase item; If the priority score of the purchase suggestion list is less than or equal to the first preset threshold and greater than the second preset threshold, the purchase demand in this list is marked as a regular purchase item; If the priority score of the purchase suggestion list is less than or equal to the second preset threshold, the purchase demand in the list is marked as a planned purchase item; For the emergency purchase items, a price inquiry request is sent to the first-tier supplier, and when a continuous supplier responds, the price comparison process is initiated; For the conventional procurement items, suppliers with supply records of similar products are screened from first-tier and second-tier suppliers to generate a list of recommended suppliers; for the planned procurement items, the demand for similar products is integrated and a centralized procurement time window is set; The marked purchase suggestion list is sent to the purchase management terminal through an encrypted channel, and reminder notifications are sent to purchasers in real time.

3. The procurement management method based on business big data according to claim 2, characterized in that: The formation method of the graded suppliers is as follows: Inputting the historical purchase records into the purchase forecasting model to generate demand forecasting results for several time periods, wherein the demand forecasting results include purchase categories, purchase quantities, and purchase times; Deviation analysis is performed on the demand forecast results and actual purchase records to calculate the forecast accuracy. The specific formula is as follows: Among them, ACC is the prediction accuracy, n is the number of prediction cycles, and w i is the time weight coefficient of the i-th forecast period, is the predicted purchase quantity for the i-th period, is the actual purchase quantity in the ith period, ΔT i is the time deviation days of the i-th cycle, k is the number of category errors, C j is the forecast deviation cost coefficient of the jth category, M is the benchmark cost parameter, and α is the category penalty factor; The deviation analysis includes: when the prediction accuracy rate is greater than the prediction threshold for multiple consecutive times, the supplier is graded; when the prediction accuracy rate is less than or equal to the prediction threshold for multiple consecutive times, the procurement prediction model is optimized.

4. The procurement management method based on business big data as claimed in claim 3, characterized in that: The classification process comprises the following steps: Establishing a supplier rating system based on the supplier's historical transaction data, wherein the supplier rating system includes the on-time delivery rate, product quality qualification rate and price competitiveness index; A weighted average algorithm is used to calculate the comprehensive score of suppliers, and suppliers are divided into first-tier suppliers, second-tier suppliers and third-tier suppliers according to the score range. First-tier suppliers are given priority in allocating purchase orders, second-tier suppliers are used as alternative suppliers, and third-tier suppliers need to develop improvement plans; In the supplier grading process, a dynamic evaluation mechanism is introduced to regularly update supplier ratings and adjust supplier grades based on changes in supplier ratings.

5. The procurement management method based on business big data according to claim 3, characterized in that: The construction method of the procurement forecast model is as follows: The preprocessed data set is grouped according to the commodity code, and all historical data is divided into a training set and a validation set in proportion; A multi-layer LSTM neural network structure is adopted, in which the neurons in the input layer establish a corresponding relationship with the feature indicators, the hidden layer enhances the feature extraction capability through a multi-layer architecture design, and the output layer designs several prediction time windows, which correspond to the demand prediction results of several time periods respectively; Determine the time step of the LSTM algorithm based on feature indicators and business needs, and build a procurement forecasting model; The procurement prediction model is trained using an adaptive optimization algorithm, an initial learning rate parameter is set, and the learning rate is dynamically adjusted according to a change trend of a loss function during the training process; The demand forecast results of the procurement forecast model are evaluated and verified. When the evaluation indicators reach the preset target requirements, it is confirmed that the forecast performance of the procurement forecast model meets the actual application needs.

6. The procurement management method based on business big data according to claim 5, characterized in that: The characteristic indicators include purchase quantity, purchase frequency, purchase unit price, inventory level, supplier delivery time, market price index and seasonal index.

7. The procurement management method based on business big data according to claim 6, characterized in that: Collect historical purchase records from a business database, and preprocess the historical purchase records to generate a preprocessed data set, including: The historical purchase records are grouped according to the purchase date, and the total purchase amount and average unit price of each commodity in each group are calculated to obtain the purchase quantity change trend and price fluctuation characteristics of the commodities, wherein the historical purchase records include commodity identification, purchase quantity, purchase price, purchase date and supplier information; The historical purchase records of the purchase quantity change trend are marked by a preset fluctuation threshold, wherein when the purchase quantity change trend value is greater than the preset fluctuation threshold, this group of purchased goods is marked as fluctuating purchase goods; when the purchase quantity change trend value is less than or equal to the preset fluctuation threshold, this group of purchased goods is marked as stable purchase goods; The purchasing features of the fluctuating purchasing commodities and the stable purchasing commodities are subdivided into categories by using a K-means clustering algorithm, and the purchasing features and category information are stored in a preprocessed data set in the form of a structured data table.

8. A procurement management system based on business big data, based on the procurement management method based on business big data according to any one of claims 1 to 7, characterized in that: Also includes, A preprocessing module, used for collecting historical purchase records from a business database, and preprocessing the historical purchase records to generate a preprocessing data set; A construction module, based on the preprocessed data set, uses a long short-term memory network (LSTM) algorithm to construct and train a purchase prediction model; A grading processing module, used to predict the procurement demand through the procurement forecasting model, output the demand forecasting result, and grade the suppliers according to the demand forecasting result to form graded suppliers; A generation module is used to generate a purchase suggestion list through the graded suppliers and send it to the purchase management terminal.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the procurement management method based on business big data described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the procurement management method based on business big data described in any one of claims 1 to 7 are implemented.

Images