Service analysis and decision support method and system

Abstract

The invention relates to the technical field of data analysis, and discloses a business analysis and decision support method and system, and the method comprises the steps: collecting purchase-related business data from a plurality of data sources, and carrying out the preprocessing; storing the preprocessed data in a centralized data warehouse, and performing management based on a preset data governance strategy; performing multi-dimensional analysis on the structured data based on a statistical analysis model and a machine learning algorithm, and generating a key performance indicator KP I and a trend prediction result; providing a customized visual display interface according to a user role, and presenting the KP I and a trend prediction result through an interactive instrument panel; and generating a purchase optimization suggestion based on the analysis result, evaluating the influence of different purchase strategies through analogue simulation, and outputting corresponding decision support information. The system effectively solves the problems of weak analysis capability, insufficient decision support and the like in an existing purchase decision system, and is suitable for multiple application scenes such as enterprise purchase management and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, in particular to a business analysis and decision support method and system. BACKGROUND

[0002] With the increasing complexity of enterprise procurement business and the acceleration of the globalization trend of the supply chain, enterprises are faced with massive, heterogeneous, and multi-source procurement-related business data. These data are usually derived from internal systems (such as ERP systems), external supplier platforms, and market quotation databases, and have various data formats, different update frequencies, and a large amount of noise and redundant information, which are difficult to directly use for analysis and decision-making.

[0003] Traditional enterprise procurement decisions rely on manual experience or simple report analysis, lacking systematic, real-time, and intelligent analysis methods. Although some enterprises have introduced data warehouses and BI tools, there are still many deficiencies in data integration, model analysis, and strategy simulation, resulting in low decision-making efficiency, delayed response, and difficulty in adapting to rapidly changing market environments.

[0004] Therefore, a business analysis and decision support method and system are provided. SUMMARY

[0005] In view of this, the present application provides a business analysis and decision support method and system, which effectively solves the problems of data integration difficulty, weak analysis capability, and insufficient decision support in existing procurement decision systems through systematic data integration, intelligent modeling analysis, visual display, and strategy simulation, and is suitable for multiple application scenarios such as enterprise procurement management and supply chain optimization.

[0006] To achieve the above purpose, in a first aspect, the present application provides a business analysis and decision support method, comprising:

[0007] Collecting procurement-related business data from multiple data sources and preprocessing to generate structured data in a unified format;

[0008] Storing the preprocessed data into a centralized data warehouse and managing based on a pre-set data governance strategy;

[0009] Performing multi-dimensional analysis on the structured data based on statistical analysis models and machine learning algorithms to generate key performance indicators (KPIs) and trend prediction results;

[0010] Providing a customized visual display interface according to user roles and presenting the KPIs and trend prediction results through an interactive dashboard;

[0011] Based on the analysis results, procurement optimization suggestions are generated, and the impact of different procurement strategies is evaluated through simulation, and corresponding decision support information is output.

[0012] Preferably, the procurement-related business data from multiple data sources is collected and preprocessed, specifically:

[0013] The data collection method includes one or more combinations of API interface call, message queue transmission, distributed crawler crawling, and IoT device real-time upload;

[0014] The data sources include internal enterprise resource planning systems, external supplier platforms, and market quotation databases;

[0015] The preprocessing includes cleaning, standardizing, and integrating the collected data, the cleaning process includes outlier detection, missing value filling, duplicate record removal, and unstructured data analysis, the unstructured data analysis extracts key fields through natural language processing technology.

[0016] Preferably, the preprocessed data is stored in a centralized data warehouse and managed based on a pre-set data governance strategy, specifically:

[0017] The data warehouse adopts a hierarchical architecture, including an operational data layer ODS, a detailed data layer DWD, and a summary data layer DWS, and combines data lake to store raw unstructured data;

[0018] The data governance strategy includes metadata management strategy, permission control strategy, and data lifecycle management;

[0019] The metadata management strategy defines data lineage, data dictionary, and field description through a metadata management system;

[0020] The permission control strategy is based on a role-based access control mechanism to ensure that sensitive data is only accessible to authorized users;

[0021] The data lifecycle management is based on a pre-set data archiving strategy to automatically clean invalid data.

[0022] Preferably, the structured data is analyzed in multiple dimensions based on statistical analysis models and machine learning algorithms, specifically:

[0023] Multiple business features are extracted from the structured data, including supplier compliance rate, price fluctuation coefficient, delivery cycle standard deviation, and inventory turnover rate;

[0024] Price trends are predicted through time series prediction models, supplier categories are identified based on clustering analysis models, and transaction anomalies are identified using anomaly detection models;

[0025] Based on historical data, dynamically calculate procurement cost saving rate, supplier delivery on-time rate and inventory turnover rate, and combine external market factors to predict KPI trends.

[0026] Preferably, the customized visualization interface is provided according to the user role, and the KPI and trend prediction results are presented through an interactive dashboard, including:

[0027] Generate personalized views according to user roles, including senior manager views, procurement specialist views and financial staff views;

[0028] Support data linkage between charts and drill-down analysis, including dynamic correlation of heat maps, Gantt charts and waterfall charts;

[0029] Limit data access range based on user roles, sensitive fields are hidden by default and need to apply for permission to view.

[0030] Preferably, the procurement optimization suggestions are generated based on the analysis results, and the impact of different procurement strategies is evaluated through simulation, and the corresponding decision support information is output, specifically:

[0031] Receive user input procurement strategy parameters, including supplier proportion adjustment, procurement batch size modification and market trend fluctuation simulation;

[0032] Simulate key nodes in the procurement process based on a discrete event simulation model, and quantify cost, risk and efficiency indicators;

[0033] Generate cost-risk-efficiency comparative analysis reports for multiple strategies, and output priority-ordered decision recommendations.

[0034] Preferably, it also includes a feedback mechanism, specifically:

[0035] Record user adoption or neglect behavior and feedback reasons for optimization suggestions;

[0036] Adjust the weight parameters of the machine learning model based on feedback data, including price prediction model, clustering model and supply chain simulation model;

[0037] Periodically retrain the model using new feedback data to improve the accuracy of decision recommendations.

[0038] Preferably, it also includes an alarm mechanism, specifically:

[0039] Set early warning thresholds for key indicators based on historical data statistical analysis, including procurement cost saving rate deviation, supplier delivery delay days, and inventory turnover abnormal fluctuations;

[0040] The key indicators are monitored periodically or by events to determine whether they exceed preset thresholds.

[0041] When an indicator is detected to exceed a preset threshold, an alert notification is sent to the designated user via email, instant messaging, or SMS, along with the current value, threshold, and trend description of the relevant indicator.

[0042] Secondly, the present invention provides a business analysis and decision support system, comprising:

[0043] The data acquisition module is used to collect procurement-related business data from multiple data sources and preprocess it to generate structured data in a unified format.

[0044] The storage module is used to store the preprocessed data in a centralized data warehouse and manage it based on a preset data governance strategy.

[0045] The analysis module is used to perform multi-dimensional analysis on the structured data based on statistical analysis models and machine learning algorithms, and generate key performance indicators (KPIs) and trend prediction results.

[0046] The display module is used to provide a customized visual display interface based on the user's role, and to present the KPIs and trend prediction results through an interactive dashboard;

[0047] The output module is used to generate procurement optimization suggestions based on the analysis results, and to evaluate the impact of different procurement strategies through simulation, and output corresponding decision support information.

[0048] Preferably, the acquisition module specifically comprises:

[0049] Data acquisition methods include one or more combinations of API interface calls, message queue transmission, distributed web crawling, and real-time uploading from IoT devices;

[0050] The data sources include internal enterprise resource planning systems, external supplier platforms, and market information databases.

[0051] The preprocessing includes cleaning, standardizing, and integrating the collected data. The cleaning process includes outlier detection, missing value imputation, duplicate record removal, and unstructured data parsing. The unstructured data parsing extracts key fields using natural language processing techniques.

[0052] This application discloses a business analysis and decision support method and system. The method collects procurement data from multiple data sources, cleans, standardizes, and integrates it, solving the problems of scattered data, inconsistent formats, and varying quality in existing technologies, thus improving the accuracy and efficiency of subsequent analysis. The pre-processed data is centrally stored in a data warehouse and managed using data governance strategies, achieving data traceability, controllable access, and lifecycle management, enhancing data security and management capabilities. By modeling structured data using statistical analysis models and machine learning algorithms, key performance indicators (such as procurement cost savings rate and supplier on-time delivery rate) can be dynamically generated and their future trends predicted, improving the depth and foresight of data analysis. Customized interactive dashboards are provided based on user roles (such as managers, procurement specialists, and financial personnel), enabling users to intuitively and efficiently obtain key business information, improving the operability and user experience of decision support. Procurement optimization suggestions are generated based on the analysis results, and the implementation effects of different procurement strategies are evaluated through simulation, providing quantitative evidence and risk prediction capabilities for procurement decisions, significantly improving the scientific nature and efficiency of decision-making. Attached Figure Description

[0053] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0054] Figure 1 A flowchart illustrating a business analysis and decision support method provided in an embodiment of the present invention;

[0055] Figure 2 This is a schematic diagram of the structure of a business analysis and decision support system provided in an embodiment of the present invention. Detailed Implementation

[0056] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0057] like Figure 1As shown in some embodiments of this application, this embodiment provides a business analysis and decision support method, specifically, the method includes the following steps:

[0058] Step S101: Collect procurement-related business data from multiple data sources and preprocess it to generate structured data in a unified format.

[0059] As mentioned above, this step involves the collection and preliminary processing of enterprise procurement-related business data, and is a fundamental link in the entire business analysis and decision support process. Its core lies in systematically integrating heterogeneous data from different sources, eliminating data format differences, information redundancy, and noise interference, and ultimately outputting standardized structured data that can be used for subsequent analysis.

[0060] Specifically, this step includes two sub-phases: the first phase, multi-source data acquisition; and the second phase, data preprocessing. Multi-source data acquisition refers to obtaining procurement-related business data from multiple systems through various data access methods (such as API calls, data scraping, message queues, etc.), covering multiple dimensions including orders, contracts, payments, logistics, and supplier evaluations. Data preprocessing refers to cleaning, standardizing, and integrating the collected raw data to ensure data quality meets analytical needs. Cleaning includes outlier identification, missing value handling, and duplicate record removal; standardization converts data from different sources into unified field naming, units, and format standards; and integration involves associating and merging scattered data according to business logic to form a structured form that can be used for modeling and analysis.

[0061] For example, before conducting procurement data analysis, a company needs to collect data from the following systems: In the ERP system, obtain data such as internal purchase orders, inventory changes, and approval processes; In the supplier platform, obtain information such as supplier quotations, delivery records, and performance status through API interfaces; In the market information database, access publicly available data such as raw material price indices and industry supply and demand trends provided by third parties; In IoT devices, collect real-time data on material inbound / outbound status and inventory levels from smart warehousing equipment; In unstructured documents, such as PDF contracts and email communication records, extract key fields through natural language processing and convert them into structured entries.

[0062] After the data collection is completed, the system performs the following preprocessing operations on the data: remove invalid order numbers and duplicate purchase requests; convert currency units such as "yuan", "dollar", and "euro" into RMB; standardize the "supplier name" field, unifying "A Technology Co., Ltd." and "A Technology" into "A Technology Co., Ltd."; and integrate order data and logistics data to generate a complete procurement cycle record including order time, delivery time, and receipt time.

[0063] After the above processing, a unified format procurement transaction data table is finally output, which is used for subsequent KPI calculation, trend prediction and strategy simulation.

[0064] It should be noted that, in specific implementation scenarios, in addition to the above solutions, the following data collection expansion solutions can be adopted: Data collection is not limited to API interfaces or direct database connections, but can also include asynchronous transmission based on message middleware, distributed web crawling, and local caching and uploading from edge devices, applicable to data integration scenarios under different network environments and system architectures; Data type expansion solutions can be adopted, meaning that the collected "procurement-related business data" is not limited to structured tabular data, but also includes semi-structured data (such as XML and JSON), unstructured text (such as scanned contracts and email text), and multimedia data (such as product acceptance photos), all of which can be structured through appropriate parsing technologies; Preprocessing logic upgrade solutions can be adopted, meaning that, in addition to traditional data cleaning, machine learning models can be introduced to automatically identify abnormal data patterns, such as using clustering algorithms to detect abnormal order amounts, or using semantic analysis to identify risk points in contract terms, thereby improving the accuracy and automation of preprocessing; A data governance pre-processing mechanism solution can be adopted, meaning that the preprocessing stage can be combined with data governance strategies to predefine data standards, field mapping rules, and quality scoring systems, ensuring that the data meets the enterprise's data management requirements before entering the analysis process, avoiding repeated cleaning and adjustments later. All of the above optional solutions fall within the scope of protection of this application.

[0065] Step S102: Store the preprocessed data in a centralized data warehouse and manage it based on a preset data governance strategy.

[0066] As mentioned above, this step involves uniformly collecting the cleaned and standardized structured procurement business data into a centralized data warehouse, and implementing a series of data governance measures on this basis to ensure the integrity, consistency, security and availability of the data.

[0067] This process comprises two core parts: the first is the construction and storage of a centralized data warehouse; the second is data governance based on pre-defined strategies. The construction and storage of the centralized data warehouse involves loading pre-processed procurement-related data from different business systems into an enterprise-level data warehouse, breaking down existing data silos and achieving centralized data management. The data warehouse adopts a layered architecture design, supporting efficient querying and complex analysis needs. Data governance based on pre-defined strategies involves continuously managing and controlling key dimensions such as data quality, access permissions, and lifecycle after data is entered into the warehouse, in accordance with the enterprise's established data governance strategies, ensuring that the data remains controllable and reliable throughout its use.

[0068] For example, after a company completes the collection and preprocessing of procurement data, it categorizes the data according to business themes (such as suppliers, orders, contracts, payments, etc.) and loads them into its enterprise-level data warehouse.

[0069] The data warehouse adopts a layered structure: the Operational Data Layer (ODS) stores the raw structured data that is closest to the source system; the Detailed Data Layer (DWD) performs light aggregation and field standardization on the ODS data; the Summary Data Layer (DWS) builds thematic wide tables for procurement analysis to quickly respond to the query needs of upper-layer applications; and the data lake stores unstructured or semi-structured raw data (such as PDF contracts, log files, etc.) for subsequent mining.

[0070] Meanwhile, the system performs the following operations according to the pre-set data governance strategy: For the "Supplier Name" field, it checks whether it conforms to the unified naming convention, and if an anomaly is found, it marks it and notifies the administrator; it sets access permissions for sensitive fields (such as contract amount and discount terms), allowing only purchasing managers and above to view them; according to the data lifecycle strategy, it automatically cleans up historical order data older than 3 years, retaining metadata indexes for audit traceability; and it establishes a data lineage tracing mechanism to record the complete flow path of each piece of purchasing data from collection, cleaning to warehousing, facilitating problem tracing.

[0071] The above methods have enabled unified management and effective governance of procurement data, providing a high-quality data foundation for subsequent analysis and decision-making.

[0072] It should be noted that, in specific implementation scenarios, in addition to the above solutions, adaptation solutions for multiple data warehouse architectures can be adopted. This means that the data warehouse is not limited to traditional relational database architectures, but can also utilize distributed columnar databases (such as Redshift and BigQuery), real-time streaming data warehouses, or cloud-native data platforms to adapt to large-scale, high-concurrency, and low-latency business scenarios. Dynamic configuration capabilities can be adopted, meaning that the preset data governance strategies are not static but can be dynamically adjusted according to business changes or compliance requirements. For example, new field validation rules can be configured through a graphical interface, permission allocation strategies can be updated, or data retention periods can be modified. Automated data quality monitoring mechanisms can be adopted, introducing automated quality detection modules during the data ingestion process to identify issues such as missing data integrity and field logic conflicts in real time, and generating quality scoring reports to assist operations personnel in promptly fixing data defects. Cross-organizational collaborative data governance models can be adopted, meaning the system supports data sharing governance under a multi-tenant architecture. While sharing the same platform, each organization can independently define its own data standards, access control policies, and data lifecycle rules to achieve cross-enterprise joint procurement data analysis. All of the above optional solutions fall within the scope of protection of this application.

[0073] Step S103: Perform multi-dimensional analysis on the structured data based on statistical analysis models and machine learning algorithms to generate key performance indicators (KPIs) and trend prediction results.

[0074] As mentioned above, this step is the core analysis component in the entire business analysis and decision support process. Its purpose is to extract key performance indicators (KPIs) with business guidance significance by modeling and analyzing the cleaned and integrated procurement-related structured data, and to predict future trends by combining historical data and external factors, so as to provide data support for subsequent visualization and strategy simulation.

[0075] This process mainly includes the following aspects: application of statistical analysis models; modeling and analysis using machine learning algorithms; calculation and dynamic updating of KPIs; and trend prediction and impact assessment. The application of statistical analysis models involves using traditional statistical methods (such as regression analysis, analysis of variance, and time series decomposition) to conduct basic analysis of procurement transaction data, identifying data distribution characteristics, abnormal fluctuations, and correlations between variables. Modeling and analysis using machine learning algorithms involves introducing supervised learning, unsupervised learning, or reinforcement learning algorithms to deeply mine and model complex issues such as procurement behavior, supplier performance, and price trends, improving the intelligence level of the analysis. The calculation and dynamic updating of KPIs involves calculating a series of key performance indicators reflecting dimensions such as enterprise procurement efficiency, cost control, and supply chain stability based on the analysis results, and dynamically updating them according to real-time data. Trend prediction and impact assessment involves combining historical KPI change patterns and external market factors (such as raw material prices and seasonal demand fluctuations) to build a predictive model, outputting trend prediction results for a future period, and assisting in the formulation of forward-looking procurement strategies.

[0076] For example, after a company completes the collection and preprocessing of procurement data, the system performs the following analyses based on this data: Using a regression analysis model, it identifies the main factors affecting procurement costs (such as transportation distance, order quantity, and supplier type) and establishes a procurement cost prediction formula; employing the K-means clustering algorithm, it classifies suppliers according to dimensions such as fulfillment rate, on-time delivery rate, and quality pass rate, dividing them into different levels such as high-value suppliers, ordinary suppliers, and risky suppliers; using the ARIMA time series model, combined with historical procurement prices and market data, it predicts the trend of commodity procurement prices for the next 6 months; using the isolated forest algorithm to detect outliers in procurement orders, such as orders whose unit price deviates from the average by more than two standard deviations, it marks them as potentially abnormal transactions for manual review; and dynamically calculates key performance indicators, including: procurement cost savings rate = (benchmark cost - actual cost) / benchmark cost × 100%, supplier on-time delivery rate = number of on-time delivered orders / total number of orders × 100%, and inventory turnover rate = cost of sales / average inventory value.

[0077] Ultimately, the system not only displayed the current KPI values, but also output KPI trend curves for future quarters based on the predictive model, providing management with scientific data support.

[0078] It should be noted that, in specific implementation scenarios, multiple machine learning model adaptation schemes can be adopted based on the above solutions. That is, the machine learning algorithms are not limited to a specific model; different algorithm combinations can be selected according to actual business needs, such as random forests for classification prediction, XGBoost for regression modeling, and LSTM for time series prediction. An automated multi-dimensional feature engineering processing scheme can be adopted, introducing an automatic feature engineering mechanism before modeling to automatically extract key business features (such as supplier historical performance scores, commodity price volatility coefficients, and procurement cycle standard deviations) from the raw data, improving the quality of model input and generalization ability. A trend prediction scheme integrating external data sources can be adopted, meaning that trend prediction not only relies on internal historical data but can also integrate external data sources (such as macroeconomic indices, industry supply and demand reports, and weather forecasts) to improve the accuracy and applicability of predictions. A dynamic configuration and expansion scheme for the KPI system can be adopted, meaning that the types and calculation methods of KPIs are not fixed but can be flexibly defined and adjusted according to the company's management objectives. For example, adding environmental indicators such as "green procurement ratio" and "carbon emission cost" can adapt to the sustainable development strategy. All of the above optional schemes are within the scope of protection of this application.

[0079] Step S104: Provide a customized visualization interface based on the user role, and present the KPIs and trend prediction results through an interactive dashboard.

[0080] As mentioned above, this step is the user-facing front-end interactive link in the entire business analysis and decision support process. Its core is to intuitively present the key performance indicators (KPIs) and trend prediction results generated in the previous steps to different types of users in a graphical way, thereby improving the efficiency of data understanding and the speed of decision response.

[0081] This step mainly includes two aspects: personalized display based on user roles and the construction and operation of interactive dashboards. Personalized display based on user roles means that the system automatically matches corresponding data permissions and display preferences according to the user's identity attributes (such as purchasing manager, finance personnel, senior managers, etc.), building customized visualization interfaces for different roles to ensure that each user can quickly obtain key information relevant to their responsibilities. The construction and operation of interactive dashboards involves integrating various chart types (such as bar charts, line charts, heatmaps, Gantt charts, etc.) into the visualization interface, and supporting users to perform interactive operations such as dynamic filtering, drill-down analysis, and linked refreshes, improving the flexibility and depth of data analysis.

[0082] For example, after a company completes a multi-dimensional analysis of its procurement data, the system automatically generates different visualization interfaces based on different user roles.

[0083] For senior managers: Display macro indicators such as overall procurement cost savings rate, supplier performance score distribution, and inventory turnover trend; provide a heat map with a global perspective to show the comparison of procurement expenditures in different regions; and support clicking on a specific region to view the corresponding supplier performance details.

[0084] For purchasing specialists: Displays the current status of pending orders, supplier on-time delivery rate rankings, and recent price fluctuation trends; integrates Gantt charts to show the difference between the purchasing plan and the actual execution progress; allows filtering by product category or supplier name via drop-down menus, and updates chart data in real time.

[0085] For finance personnel: It presents invoice matching information, payment cycle statistics, and analysis of discrepancies between contract amounts and actual settlements; it uses waterfall charts to display the composition of procurement costs, helping to identify key cost control aspects; and it supports exporting financial analysis reports for specified time periods for auditing.

[0086] In addition, all users can use interactive features such as dragging chart positions to customize layouts; setting time sliders to select specific analysis intervals; and clicking on a data point in a chart to update the data content of other related charts.

[0087] The above methods enable efficient visualization of KPI and trend prediction results, improving the usability of data analysis and user experience.

[0088] It should be noted that, in specific implementation scenarios, a multi-terminal adaptation solution can be adopted based on the above solutions. This means the visualization interface is not limited to PC browser access but also supports mobile apps, large-screen dashboards, and other terminal formats to adapt to different usage scenarios and device environments. An intelligent recommendation view solution can be adopted, where the system records user browsing history, frequently used filtering conditions, and attention metrics, and dynamically recommends the most likely chart combinations or analytical dimensions of interest using machine learning algorithms, improving ease of use. A modular configuration mechanism can be adopted, allowing users to freely add, delete, or adjust the position and style of visualization components through drag-and-drop, creating a personalized dashboard layout to meet the needs of different positions and management styles. All of the above optional solutions fall within the scope of protection of this application.

[0089] Step S105: Generate procurement optimization suggestions based on the analysis results, evaluate the impact of different procurement strategies through simulation, and output corresponding decision support information.

[0090] As mentioned above, this step is the core decision-making link in the entire business analysis and decision support process. Its purpose is to transform the key performance indicators (KPIs) and trend forecast results extracted in the previous steps into procurement optimization suggestions with practical guidance, and to evaluate the implementation effect of different procurement strategies through simulation, thereby providing scientific and quantitative decision support information for the final decision-maker.

[0091] This process mainly includes the following aspects: generating optimization suggestions based on analysis results, simulating procurement strategies, and outputting structured decision support information. Generating optimization suggestions based on analysis results means that the system automatically generates one or more sets of procurement optimization suggestions based on historical data, current KPI status, and market forecast trends, combined with the company's procurement objectives (such as cost minimization, risk control, and delivery assurance). These suggestions can cover supplier selection, order allocation ratios, and procurement cycle adjustments. Procurement strategy simulation involves building a virtual environment before actual implementation to simulate multiple rounds of strategies based on the proposed optimization suggestions, quantifying their expected performance in terms of cost, efficiency, and risk, and helping to identify the optimal or suboptimal strategy. Outputting structured decision support information involves organizing the simulation results into an easy-to-understand report format, including cost-risk-efficiency comparisons of each strategy, recommendation rankings, and explanations of key influencing factors, for user reference and final decision-making.

[0092] For example, after a company completes its procurement data analysis, the system generates the following procurement optimization suggestions based on current inventory levels, supplier fulfillment status, and future market demand forecasts: Suggestion 1: Change the main supplier of Category A materials from the current supplier X to supplier Y, because its on-time delivery rate has been 12% higher in the past 6 months, and its price fluctuations are smaller; Suggestion 2: Adjust the procurement frequency of Category B materials from once a month to once every two weeks to reduce the risk of inventory backlog; Suggestion 3: In anticipation of the upcoming peak demand season, make bulk purchases of Category C raw materials 3 months in advance to lock in the current lower price level.

[0093] Subsequently, the system simulated the three strategies using a discrete event simulation model. The simulation results showed that: Strategy 1 can reduce annual procurement costs by about 8%, but switching suppliers may bring short-term delivery delay risks; Strategy 2 slightly increases procurement management costs while reducing inventory holdings; Strategy 3 has higher returns in the context of price declines, but may lead to excessive capital occupation in the case of continuous price increases.

[0094] Finally, the system outputs a structured decision support report, including: cost change curves for each strategy; a risk level rating table (low, medium, high); a recommendation priority ranking; and an explanation of external variables that may affect the effectiveness of the strategies (such as raw material price fluctuations, logistics delays, etc.). Users can choose the most suitable procurement plan based on their own preferences and risk tolerance.

[0095] It should be noted that, in specific implementation scenarios, a multi-objective optimization suggestion generation scheme can be adopted based on the above solutions. This means the optimization suggestions are not limited to a single objective (such as cost minimization) but can also support multi-objective collaborative optimization, such as controlling costs while considering sustainable development goals like carbon emission reduction and improved supply chain resilience. A suggestion generation scheme based on the fusion of rule engines and AI can be adopted, meaning the generation of optimization suggestions is not limited to machine learning models but can also combine expert knowledge bases, industry best practice libraries, and rule engines to form a more logical suggestion system. A multi-dimensional procurement strategy modeling scheme can be adopted, meaning the simulation is not limited to changes in price or order quantity but can also simulate various combinations of procurement strategies, such as: order allocation ratios among different suppliers; comparison of multiple procurement models (centralized procurement vs. decentralized procurement); and the impact of different payment terms (prepayment, payment terms, letters of credit) on cash flow. A dynamic parameter input and assumption configuration scheme can be adopted, meaning users can manually set different assumptions during the simulation process (such as market volatility, transportation delay probability, exchange rate fluctuation range, etc.) to test the stability and adaptability of the strategy under different scenarios. All of the above optional schemes are within the scope of protection of this application.

[0096] In some embodiments of this application, to effectively address the problems of scattered data sources, diverse formats, and inconsistent quality in procurement operations, and to build a unified, high-quality data foundation for enterprises to support subsequent business analysis and decision support processes, the collection and preprocessing of procurement-related business data from multiple data sources specifically includes:

[0097] Data acquisition methods include one or more combinations of API interface calls, message queue transmission, distributed web crawling, and real-time uploading from IoT devices;

[0098] The data sources include internal enterprise resource planning systems, external supplier platforms, and market information databases.

[0099] The preprocessing includes cleaning, standardizing, and integrating the collected data. The cleaning process includes outlier detection, missing value imputation, duplicate record removal, and unstructured data parsing. The unstructured data parsing extracts key fields using natural language processing techniques.

[0100] As described above, in this embodiment, the data acquisition methods include, but are not limited to, one or more combinations of API interface calls, message queue transmission, distributed web crawling, and real-time uploads from IoT devices. Specifically, API interface calls are used for data interaction with internal enterprise systems or other external platforms; message queue transmission is used to achieve asynchronous data synchronization and decoupled communication; distributed web crawling is used to extract unstructured data from web pages or platform pages; and real-time uploads from IoT devices are used to acquire dynamic data from physical devices such as smart warehousing equipment and logistics tracking terminals.

[0101] Data sources include the company's internal systems, external collaboration platforms, and third-party databases. Specifically, internal systems mainly include Enterprise Resource Planning (ERP) systems, used to obtain core business data such as purchase orders, contract information, and inventory changes; external platforms include supplier management systems and e-commerce platforms, used to obtain information such as supplier quotations, performance records, and transaction flows; and market information databases provide external reference data such as raw material price indices and industry supply and demand trends.

[0102] After data acquisition, the system performs preprocessing operations to ensure data consistency, integrity, and usability. Preprocessing includes three stages: cleaning, standardization, and integration. Cleaning identifies and corrects errors and anomalies in the data, including outlier detection, missing value imputation, and duplicate record removal. For unstructured data such as PDF contracts, emails, and scanned documents, the system uses natural language processing to parse and extract key fields, such as contract amount, delivery date, and supplier name, transforming the unstructured data into structured data suitable for analysis. Standardization unifies data from different sources and formats into consistent field naming rules, unit systems, and time expressions to facilitate subsequent modeling and analysis. Integration merges the cleaned and standardized data according to business logic, generating a complete, subject-oriented data table as the foundation for subsequent analysis.

[0103] In some embodiments of this application, to achieve centralized management, efficient utilization, and security of procurement business data, and to build a stable, reliable, and scalable data infrastructure platform for enterprises to support subsequent multi-dimensional analysis, visualization, and decision optimization processes, the preprocessed data is stored in a centralized data warehouse and managed based on a preset data governance strategy, specifically as follows:

[0104] The data warehouse adopts a layered architecture, including an operational data layer (ODS), a detailed data layer (DWD), and a summary data layer (DWS), and combines a data lake to store raw unstructured data.

[0105] The data governance strategy includes metadata management strategy, access control strategy, and data lifecycle management;

[0106] The metadata management strategy defines data lineage, data dictionary, and field descriptions through the metadata management system;

[0107] The access control policy is based on a role-based access control mechanism to ensure that sensitive data is only accessible to authorized users.

[0108] The data lifecycle management is based on a preset data archiving strategy to automatically clean up invalid data.

[0109] As described above, in this embodiment, the data warehouse adopts a layered architecture design, including an operational data layer (ODS), a detailed data layer (DWD), and a summary data layer (DWS). The operational data layer stores cleaned data closest to the original source, maintaining its original structure. The detailed data layer performs light processing on the ODS layer data, such as standardization, deduplication, and field normalization, to form detailed tables suitable for analysis. The summary data layer aggregates and calculates the detailed data according to business themes, generating wide tables for analytical queries, improving subsequent analysis efficiency. Furthermore, the system incorporates data lake technology to store unstructured raw data, such as log files, scanned contracts, and supplier email attachments, for later on-demand extraction and use.

[0110] Building upon data storage, the system further implements data governance strategies to ensure the controllability and compliance of data throughout its entire lifecycle. These strategies include metadata management, access control, and data lifecycle management. The metadata management strategy, through the deployment of a metadata management system, defines and manages metadata for all data entering the data warehouse. This strategy covers recording data lineage, tracing the complete flow path of each piece of data from collection and processing to final use; it also includes maintaining a data dictionary to describe the meaning, value range, and update rules of each field; and unified management of field description information to ensure consistent data understanding among different users, improving data readability and sharing capabilities.

[0111] Access control policies utilize a role-based access control (RBAC) mechanism, assigning appropriate data access permissions based on different user roles. For example, purchasing specialists can only view purchase orders and supplier performance related to their assigned product categories, while finance personnel can access payment records and invoice matching data. Fields involving trade secrets or sensitive information, such as contract prices and discount terms, are hidden by default and accessible only to specifically authorized users, thus ensuring data security.

[0112] Data lifecycle management strategies are used to regulate the storage period and disposal methods of data. Based on preset data archiving strategies, the system automatically identifies and cleans up historical data that has exceeded its retention period, such as expired order records from three years ago, while retaining necessary metadata indexes to support audit traceability. For data that still needs to be retained but is accessed infrequently, the system can migrate it to a low-cost storage area, balancing data availability and storage costs.

[0113] In some embodiments of this application, in order to provide enterprises with comprehensive intelligent analysis capabilities covering supplier evaluation, price forecasting, anomaly monitoring, and performance evaluation, and effectively improve the scientific nature and responsiveness of procurement decisions, the multi-dimensional analysis of the structured data based on statistical analysis models and machine learning algorithms specifically includes:

[0114] Multidimensional business features are extracted from the structured data, including supplier fulfillment rate, price volatility coefficient, delivery cycle standard deviation, and inventory turnover rate.

[0115] Price trends are predicted using time series forecasting models, supplier classification is identified based on cluster analysis models, and transaction anomalies are identified using anomaly detection models.

[0116] Based on historical data, we dynamically calculate procurement cost savings rate, supplier on-time delivery rate, and inventory turnover rate, and combine external market factors to predict KPI trends.

[0117] As described above, in this embodiment, the system first extracts multi-dimensional business features from the cleaned and standardized structured data. These features include, but are not limited to, supplier fulfillment rate, price volatility coefficient, delivery cycle standard deviation, and inventory turnover rate, which can comprehensively reflect core indicators such as efficiency, cost, risk, and supply chain stability in the procurement process.

[0118] Subsequently, the system further analyzes the aforementioned features using different modeling methods. Specifically, the time series forecasting model is used to model and predict future price trends, helping companies prepare for market price fluctuations in advance when formulating procurement plans; the clustering analysis model is used to classify suppliers into different categories based on factors such as historical transaction behavior and performance, such as high-value suppliers, ordinary suppliers, and risky suppliers, providing data support for supplier management; and the anomaly detection model is used to identify abnormal behaviors in procurement transactions, such as abnormally high-priced orders, duplicate orders, and abnormal payments, assisting companies in promptly identifying potential risks.

[0119] Building upon this foundation, the system also dynamically calculates several Key Performance Indicators (KPIs) based on historical data. For example, the procurement cost savings rate is calculated by comparing the difference between the benchmark purchase price and the actual transaction price; the supplier on-time delivery rate is calculated based on the proportion of on-time deliveries according to historical delivery records; and the inventory turnover rate measures inventory management efficiency as the ratio of cost of sales to average inventory value. Simultaneously, the system incorporates external market factors, such as raw material price trends and changes in industry supply and demand, to predict future trends in these KPIs, generating forward-looking analytical conclusions.

[0120] In some embodiments of this application, in order to provide personalized data presentation methods according to the needs of different users, improve the intuitiveness and interactive experience of data analysis, and at the same time take into account data security and access control requirements, so as to build an efficient, secure and intelligent decision support environment for enterprises, the provision of customized visualization interfaces based on user roles, and the presentation of KPIs and trend prediction results through interactive dashboards, includes:

[0121] Personalized views are generated based on user roles, including views for senior managers, purchasing specialists, and finance personnel.

[0122] It supports data linkage and drill-down analysis between charts, including dynamic association of heatmaps, Gantt charts and waterfall charts;

[0123] Data access is restricted based on user roles. Sensitive fields are hidden by default and can only be viewed after requesting permission.

[0124] As described above, in this embodiment, the system automatically matches the corresponding display content based on the user's identity. For example, the view for senior managers mainly displays macro indicators such as overall procurement cost savings rate, supplier performance score distribution, and inventory turnover trend; the view for procurement specialists focuses on operational information such as order execution status, supplier on-time delivery rate ranking, and price fluctuation trend; while the view for finance personnel focuses on the presentation of financial data such as invoice matching status, payment cycle statistics, and differences between contract amount and actual settlement.

[0125] To enhance the flexibility of data presentation and the depth of analysis, the system supports data linkage and drill-down analysis between charts. For example, after selecting a region in a heatmap, a Gantt chart will update to display the procurement plan progress for that region, while a waterfall chart will simultaneously show the corresponding cost structure changes. In addition, users can click or filter on specific parts of the data in the charts to explore the data in greater depth and obtain more granular information.

[0126] While ensuring data security, the system controls data access scope based on user roles. Fields involving sensitive corporate information, such as contract amounts, discount terms, and supplier cost prices, are hidden by default and can only be viewed after a user applies for and obtains the appropriate permissions. This access control mechanism, combined with role-based access policies, ensures that users in different positions can only access data content relevant to their responsibilities and within their authorized scope, preventing information leakage and misuse.

[0127] In some embodiments of this application, to provide scientific and quantitative simulation analysis results before the actual implementation of procurement strategies, helping users identify potential problems and optimize procurement plans, thereby improving the foresight and decision-making quality of procurement management, the process of generating procurement optimization suggestions based on the analysis results, evaluating the impact of different procurement strategies through simulation, and outputting corresponding decision support information specifically includes:

[0128] The system receives procurement strategy parameters input by the user, including supplier ratio adjustment, procurement batch modification, and market fluctuation simulation.

[0129] The key nodes in the procurement process are simulated based on a discrete event simulation model, and cost, risk and efficiency indicators are quantified.

[0130] Generate a cost-risk-efficiency comparative analysis report for multiple strategies and output decision recommendations with priority ranking.

[0131] As described above, in this embodiment, the system first receives procurement strategy parameters input by the user. These parameters include, but are not limited to, supplier ratio adjustment, purchase batch modification, and market fluctuation simulation. Specifically, supplier ratio adjustment is used to set the order allocation ratio among different suppliers; purchase batch modification is used to adjust the quantity of a single purchase to cope with inventory changes or price fluctuations; and market fluctuation simulation allows the user to set the range of changes in external factors such as raw material prices and transportation costs to test the adaptability of the strategy under different market environments.

[0132] After acquiring the strategy parameters, the system simulates key nodes in the procurement process using a discrete event simulation model. These key nodes include purchase order generation, supplier response, logistics and transportation, inventory updates, and payment settlement. By simulating the process execution under different strategies, the system can quantify the overall performance of each strategy in terms of cost, risk, and efficiency. For example, one strategy may have an advantage in procurement costs but may lead to increased delivery delay risks; another strategy may improve supply chain stability but may cause inventory backlog problems.

[0133] After completing the strategy simulation, the system summarizes and analyzes the execution results of each strategy, generating a cost-risk-efficiency comparison report for multiple strategies. This report presents the differences between different strategies in the form of charts and text, and, combined with user-defined priority preferences (such as cost priority, risk control priority, etc.), outputs decision recommendations with a suggested order. For example, the system can sort the strategies according to their comprehensive scores and mark the main influencing factors for each strategy, such as "Strategy A is optimal in terms of cost control, but has a higher risk in terms of supplier concentration."

[0134] In some embodiments of this application, in order to achieve a closed-loop optimization process from "generating suggestions—user feedback—model optimization—regenerating suggestions," and effectively improve the intelligence level and practical application value of the procurement decision support system, a feedback mechanism is also included, specifically:

[0135] Record user behavior regarding whether they adopt or ignore optimization suggestions, along with the reasons for their feedback;

[0136] Adjusting the weight parameters of machine learning models based on feedback data, including price prediction models, clustering models, and supply chain simulation models;

[0137] Regularly retrain the model using new feedback data to improve the accuracy of decision recommendations.

[0138] As described above, in this embodiment, the system automatically records the user's adoption or disregard of optimization suggestions during use, and allows the user to input feedback reasons on the interface. For example, the user can explain why a suggestion was not adopted: "the current supplier is irreplaceable," "the strategy risk is too high," or "market conditions have changed." This feedback information is captured and stored by the system, serving as an important basis for model optimization.

[0139] Based on the collected feedback data, the system adjusts and optimizes the parameters of several key models. Specifically, the system can adjust the feature weights in the price prediction model to more accurately reflect market trends; optimize the classification boundaries of the clustering model to improve the rationality of supplier classification; and improve the behavioral parameters in the supply chain simulation model to better reflect the decision-making logic in actual business scenarios.

[0140] In addition, the system has a regular model update mechanism. Within a preset time period (such as weekly or monthly), the system uses newly accumulated feedback data along with historical data to retrain the relevant machine learning models, thereby continuously improving model performance. This mechanism ensures that the system can adapt to changes in the procurement business environment and maintain the timeliness and accuracy of its recommendations.

[0141] In some embodiments of this application, in order to notify relevant personnel immediately when key business indicators show abnormalities, assisting enterprises in taking timely countermeasures, thereby effectively reducing procurement risks and improving supply chain response efficiency and management accuracy, an alarm mechanism is also included, specifically:

[0142] Early warning thresholds for key indicators are set based on historical data statistical analysis. These key indicators include deviations in procurement cost savings rate, supplier delivery delay days, and abnormal fluctuations in inventory turnover.

[0143] The key indicators are monitored periodically or by events to determine whether they exceed preset thresholds.

[0144] When an indicator is detected to exceed a preset threshold, an alert notification is sent to the designated user via email, instant messaging, or SMS, along with the current value, threshold, and trend description of the relevant indicator.

[0145] As described above, in this embodiment, the system sets early warning thresholds for several key business indicators based on historical data analysis results. These key indicators include, but are not limited to, deviations in procurement cost savings rates, supplier delivery delay days, and abnormal fluctuations in inventory turnover rates, reflecting significant trends in procurement cost control, supply chain stability, and inventory management efficiency. The system models historical data using statistical methods and automatically or manually sets reasonable early warning threshold ranges to ensure the early warning mechanism has high accuracy and practicality.

[0146] After setting the early warning rules, the system continuously monitors the aforementioned key indicators. The monitoring method can be timed polling, such as evaluating the relevant indicators once per hour or per day; or it can be an event-driven trigger mechanism, such as immediately evaluating when the order status changes, inventory levels are updated, or price fluctuations reach a set range.

[0147] When the system detects that a key indicator exceeds a preset threshold, it will automatically trigger an alarm process. The system will send an alert notification to the designated user via email, instant messaging tools (such as the company's internal communication platform), or SMS. The alert notification includes the current value of the indicator, the set threshold, the extent of the exceedance, and a description of the indicator's historical trend, helping users quickly understand the severity and potential impact of the anomaly. Furthermore, the system allows users to configure alert rules, including threshold adjustments, notification method selection, and alarm level settings, to adapt to the needs of users in different roles and the company's internal risk management strategies.

[0148] Compared with existing technologies, this application discloses a business analysis and decision support method. This method collects procurement data from multiple data sources, cleans, standardizes, and integrates it, solving the problems of scattered data, inconsistent formats, and varying quality in existing technologies, thus improving the accuracy and efficiency of subsequent analysis. The pre-processed data is centrally stored in a data warehouse and managed using data governance strategies, achieving data traceability, controllable access, and lifecycle management, enhancing data security and management capabilities. By modeling structured data using statistical analysis models and machine learning algorithms, key performance indicators (such as procurement cost savings rate and supplier on-time delivery rate) can be dynamically generated and their future trends predicted, improving the depth and foresight of data analysis. Customized interactive dashboards are provided based on user roles (such as managers, procurement specialists, and financial personnel), enabling users to intuitively and efficiently obtain key business information, improving the operability and user experience of decision support. Procurement optimization suggestions are generated based on the analysis results, and the implementation effects of different procurement strategies are evaluated through simulation, providing quantitative evidence and risk prediction capabilities for procurement decisions, significantly improving the scientific nature and efficiency of decision-making.

[0149] Based on the same inventive concept as the methods described above, this application also proposes a business analysis and decision support system, such as... Figure 2 The diagram shown is a structural schematic of a business analysis and decision support system, which includes:

[0150] The data acquisition module is used to collect procurement-related business data from multiple data sources and preprocess it to generate structured data in a unified format.

[0151] The storage module is used to store the preprocessed data in a centralized data warehouse and manage it based on a preset data governance strategy.

[0152] The analysis module is used to perform multi-dimensional analysis on the structured data based on statistical analysis models and machine learning algorithms, and generate key performance indicators (KPIs) and trend prediction results.

[0153] The display module is used to provide a customized visual display interface based on the user's role, and to present the KPIs and trend prediction results through an interactive dashboard;

[0154] The output module is used to generate procurement optimization suggestions based on the analysis results, and to evaluate the impact of different procurement strategies through simulation, and output corresponding decision support information.

[0155] Preferably, the acquisition module specifically comprises:

[0156] Data acquisition methods include one or more combinations of API interface calls, message queue transmission, distributed web crawling, and real-time uploading from IoT devices;

[0157] The data sources include internal enterprise resource planning systems, external supplier platforms, and market information databases.

[0158] The preprocessing includes cleaning, standardizing, and integrating the collected data. The cleaning process includes outlier detection, missing value imputation, duplicate record removal, and unstructured data parsing. The unstructured data parsing extracts key fields using natural language processing techniques.

[0159] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0160] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0161] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0162] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process.Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A business analysis and decision support method, characterized in that, include: Collect procurement-related business data from multiple data sources and preprocess it to generate structured data in a unified format; The preprocessed data is stored in a centralized data warehouse and managed based on a preset data governance strategy; The structured data is analyzed from multiple dimensions using statistical analysis models and machine learning algorithms to generate key performance indicators (KPIs) and trend prediction results. Provides a customized visualization interface based on user roles, and presents the KPIs and trend prediction results through an interactive dashboard; Based on the analysis results, procurement optimization suggestions are generated, and the impact of different procurement strategies is evaluated through simulation, outputting corresponding decision support information.

2. The method as described in claim 1, characterized in that, The process of collecting and preprocessing procurement-related business data from multiple data sources specifically includes: Data acquisition methods include one or more combinations of API interface calls, message queue transmission, distributed web crawling, and real-time uploading from IoT devices; The data sources include internal enterprise resource planning systems, external supplier platforms, and market information databases. The preprocessing includes cleaning, standardizing, and integrating the collected data. The cleaning process includes outlier detection, missing value imputation, duplicate record removal, and unstructured data parsing. The unstructured data parsing extracts key fields using natural language processing techniques.

3. The method as described in claim 1, characterized in that, The step of storing the preprocessed data in a centralized data warehouse and managing it based on a preset data governance strategy specifically involves: The data warehouse adopts a layered architecture, including an operational data layer (ODS), a detailed data layer (DWD), and a summary data layer (DWS), and combines a data lake to store raw unstructured data. The data governance strategy includes metadata management strategy, access control strategy, and data lifecycle management; The metadata management strategy defines data lineage, data dictionary, and field descriptions through the metadata management system; The access control policy is based on a role-based access control mechanism to ensure that sensitive data is only accessible to authorized users. The data lifecycle management is based on a preset data archiving strategy to automatically clean up invalid data.

4. The method as described in claim 2, characterized in that, The multi-dimensional analysis of the structured data based on statistical analysis models and machine learning algorithms specifically includes: Multidimensional business features are extracted from the structured data, including supplier fulfillment rate, price volatility coefficient, delivery cycle standard deviation, and inventory turnover rate. Price trends are predicted using time series forecasting models, supplier classification is identified based on cluster analysis models, and transaction anomalies are identified using anomaly detection models. Based on historical data, we dynamically calculate procurement cost savings rate, supplier on-time delivery rate, and inventory turnover rate, and combine external market factors to predict KPI trends.

5. The method as described in claim 1, characterized in that, The provision of a customized visualization interface based on user roles, presenting the KPIs and trend prediction results through an interactive dashboard, includes: Personalized views are generated based on user roles, including views for senior managers, purchasing specialists, and finance personnel. It supports data linkage and drill-down analysis between charts, including dynamic association of heatmaps, Gantt charts and waterfall charts; Data access is restricted based on user roles. Sensitive fields are hidden by default and can only be viewed after requesting permission.

6. The method as described in claim 1, characterized in that, The process involves generating procurement optimization suggestions based on the analysis results, evaluating the impact of different procurement strategies through simulation, and outputting corresponding decision support information. Specifically: The system receives procurement strategy parameters input by the user, including supplier ratio adjustment, procurement batch modification, and market fluctuation simulation. The key nodes in the procurement process are simulated based on a discrete event simulation model, and cost, risk and efficiency indicators are quantified. Generate a cost-risk-efficiency comparative analysis report for multiple strategies and output decision recommendations with priority ranking.

7. The method as described in claim 1, characterized in that, It also includes a feedback mechanism, specifically: Record user behavior regarding whether they adopt or ignore optimization suggestions, along with the reasons for their feedback; Adjusting the weight parameters of machine learning models based on feedback data, including price prediction models, clustering models, and supply chain simulation models; Regularly retrain the model using new feedback data to improve the accuracy of decision recommendations.

8. The method as described in claim 1, characterized in that, It also includes an alarm mechanism, specifically: Early warning thresholds for key indicators are set based on historical data statistical analysis. These key indicators include deviations in procurement cost savings rate, supplier delivery delay days, and abnormal fluctuations in inventory turnover. The key indicators are monitored periodically or by events to determine whether they exceed preset thresholds. When an indicator is detected to exceed a preset threshold, an alert notification is sent to the designated user via email, instant messaging, or SMS, along with the current value, threshold, and trend description of the relevant indicator.

9. A business analysis and decision support system, characterized in that, include: The data acquisition module is used to collect procurement-related business data from multiple data sources and preprocess it to generate structured data in a unified format. The storage module is used to store the preprocessed data in a centralized data warehouse and manage it based on a preset data governance strategy. The analysis module is used to perform multi-dimensional analysis on the structured data based on statistical analysis models and machine learning algorithms, and generate key performance indicators (KPIs) and trend prediction results. The display module is used to provide a customized visual display interface based on the user's role, and to present the KPIs and trend prediction results through an interactive dashboard; The output module is used to generate procurement optimization suggestions based on the analysis results, and to evaluate the impact of different procurement strategies through simulation, and output corresponding decision support information.

10. The system as described in claim 9, characterized in that, The acquisition module is specifically: Data acquisition methods include one or more combinations of API interface calls, message queue transmission, distributed web crawling, and real-time uploading from IoT devices; The data sources include internal enterprise resource planning systems, external supplier platforms, and market information databases. The preprocessing includes cleaning, standardizing, and integrating the collected data. The cleaning process includes outlier detection, missing value imputation, duplicate record removal, and unstructured data parsing. The unstructured data parsing extracts key fields using natural language processing techniques.

Images