Historical order data-based ship part purchase quotation optimization system and method

Abstract

The invention discloses a ship part purchase quotation optimization system based on historical order data, and the system comprises a data preprocessing module which is used for preprocessing the historical order data, and generating a complete order record; the standard selling price prediction module is used for predicting the optimal selling price of the current market by analyzing historical selling data based on the multi-dimensional market characteristics; the optimal quotation prediction module is used for predicting the order forming probability of the current quotation by analyzing the historical order forming rate of the quotation, and adjusting a quotation strategy based on a prediction result; the supplier selection module is used for calculating comprehensive scores of suppliers according to prices, quality and option weights on the basis of historical purchase data, and recommending optimal suppliers; and the interface integration module is used for being integrated with an enterprise management system through a back-end API. The invention also discloses a historical order data-based ship part purchase quotation optimization method. According to the invention, historical order data can be automatically analyzed, the order price and the order rate can be predicted, and the optimal supplier can be intelligently selected.

Description

Technical Field

[0001] The present invention relates to the technical fields of artificial intelligence and machine learning, and particularly to a ship component procurement quotation optimization system and method based on historical order data. Background Art

[0002] In the modern ship trading and component procurement industries, enterprises face fierce market competition. To gain an advantage in the competition, enterprises not only need to accurately predict the success probability of customer orders, but also need to make reasonable price adjustments during the quotation process to ensure that they can meet customer needs while maximizing profits. In addition, selecting the right suppliers is also an important link in reducing procurement costs and increasing profits. However, traditional quotation and supplier selection methods usually rely on manual experience, lack scientificity and systematicness, and are difficult to cope with the complex and changing market environment.

[0003] Therefore, a ship component procurement quotation optimization system and method based on historical order data are provided. Summary of the Invention

[0004] The purpose of the present invention is to provide a ship component procurement quotation optimization system and method based on historical order data to overcome the existing defects, which can automatically analyze historical order data, predict the success price and success rate, and intelligently select the optimal supplier.

[0005] The technical solution to achieve the above purpose is as follows:

[0006] The ship component procurement quotation optimization system based on historical order data according to one aspect of the present invention includes:

[0007] A data preprocessing module for preprocessing historical order data to generate complete order records;

[0008] A standard selling price prediction module for predicting the optimal selling price in the current market based on multi-dimensional market characteristics by analyzing historical sales data;

[0009] An optimal quotation prediction module for predicting the success probability of the current quotation by analyzing the historical success rate of quotation forms and adjusting the quotation strategy based on the prediction result;

[0010] A supplier selection module for calculating the comprehensive score of suppliers according to price, quality, and delivery option weights based on historical procurement data and recommending the optimal supplier;

[0011] An interface integration module for integrating with the enterprise management system through the backend API, supporting the simultaneous processing of multiple quotation information and returning real-time prediction results.

[0012] Preferably, in the data preprocessing module, the preprocessing operations include:

[0013] Clean the historical order data, removing unimportant fields and duplicate records;

[0014] Merge relevant data tables through inner join and left join methods to generate complete order records;

[0015] Fill in and process missing values.

[0016] Preferably, in the standard selling price prediction module, based on historical sales data, perform data screening and distribution analysis according to product quality, sales type, site, and currency characteristics, and combine multi-dimensional market characteristics to generate optimal selling price suggestions for the current market.

[0017] Preferably, in the optimal quotation prediction module, adopt a binary classification model to predict whether a quotation will be converted into an order, and optimize and adjust the quotation based on the order conversion probability.

[0018] Preferably, in the supplier selection module, based on historical purchase data, return the supplier that provided the lowest price for similar products in the past, which is the optimal supplier.

[0019] Preferably, in the interface integration module, achieve seamless integration with the enterprise management system through the backend API, support real-time processing of multiple quotation information, and provide functions such as quotation prediction, order conversion rate prediction, standard selling price prediction, and supplier recommendation.

[0020] The method for optimizing the procurement quotation of ship components based on historical order data according to the second aspect of the present invention includes:

[0021] Step S1, preprocess the historical order data to generate complete order records;

[0022] Step S2, use the standard selling price algorithm to predict the optimal selling price in the current market based on the historical sales data of the product and multi-dimensional market characteristics;

[0023] Step S3, predict the order conversion probability for each quotation by analyzing the order conversion rate of historical quotations;

[0024] Step S4, calculate the comprehensive score of suppliers based on historical purchase data according to price, quality, and delivery option weights, and recommend the optimal supplier;

[0025] Step S5, send the prediction results to the existing management system of the enterprise through the backend API, enabling the enterprise to make real-time and intelligent decisions in a dynamic market environment.

[0026] Preferably, in step S1, the data preprocessing stage includes:

[0027] Clean the historical order data, removing unimportant fields and duplicate records;

[0028] Merge relevant data tables through inner join and left join methods to generate complete order records;

[0029] Fill in and process missing values;

[0030] In step S2, based on historical sales data, perform data screening and distribution analysis according to product quality, sales type, site, and currency characteristics, and combine multi-dimensional market characteristics to generate optimal selling price suggestions for the current market.

[0031] Preferably, in step S3, a binary classification model is used to predict whether a quotation will be converted into an order, and based on the order conversion probability, optimize and adjust the quotation;

[0032] In step S4, based on historical purchase data, return the supplier that provided the lowest price for similar products in the past, which is the optimal supplier;

[0033] In step S5, the prediction results include: quotation prediction, order conversion rate prediction, standard selling price prediction, and supplier recommendation.

[0034] The beneficial effects of the present invention are as follows: The present invention combines advanced data processing technologies and algorithm models, and can efficiently clean, integrate, and analyze historical order data; through standard selling price prediction, optimal quotation prediction, and order conversion rate prediction, the system helps enterprises generate accurate quotations that conform to market dynamics; at the same time, the system also analyzes historical purchase data and automatically recommends the optimal supplier to ensure the optimization of procurement costs on the premise of meeting quality and delivery requirements; in addition, the system is seamlessly integrated with the enterprise's existing management system through the backend API, can process and analyze multiple quotation information in real time, and quickly return optimized decision results; this not only greatly improves the decision-making efficiency of enterprises, but also reduces errors and risks caused by manual decision-making. Brief Description of the Drawings

[0035] Figure 1 is a module diagram of the ship component procurement quotation optimization system based on historical order data of the present invention;

[0036] Figure 2 is a flowchart of the ship component procurement quotation optimization method based on historical order data of the present invention. Detailed Embodiments

[0037] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, the terms "first", "second", "third" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0038] The present invention will be further described below with reference to the accompanying drawings.

[0039] As Figure 1 shown, the ship component procurement quotation optimization system based on historical order data includes: a data preprocessing module 1, a standard selling price prediction module 2, an optimal quotation prediction module 3, a supplier selection module 4, and an interface integration module 5.

[0040] The data preprocessing module 1 is used to preprocess the historical order data to generate complete order records.

[0041] In the embodiment, the preprocessing operations include:

[0042] Cleaning the historical order data to remove unimportant fields and duplicate records;

[0043] Merging relevant data tables through inner join and left join methods to generate complete order records, ensuring that the subsequent algorithm model can analyze and predict based on accurate and complete data;

[0044] Filling and processing missing values to ensure data consistency and integrity.

[0045] The standard selling price prediction module 2 is used to predict the optimal selling price in the current market by analyzing historical sales data based on multi-dimensional market characteristics.

[0046] In the embodiment, based on historical sales data, data screening and distribution analysis are performed according to product quality, sales type, site, and currency characteristics, and combined with multi-dimensional market characteristics, an optimal selling price recommendation for the current market is generated; this algorithm analyzes data on goods in different cost ranges to generate market selling price recommendations with high accuracy, thereby helping enterprises reduce pricing errors and maximize sales revenue.

[0047] The optimal quotation prediction module 3 is used to predict the order success probability of the current quotation by analyzing the historical order success rate of the quotation form, and adjust the quotation strategy based on the prediction result.

[0048] In the embodiment, a binary classification model is adopted to predict whether a quotation will be converted into an order. Based on the order conversion probability, the quotation is optimized and adjusted to improve the order conversion rate. For orders with a high order conversion probability, the system will further refine the quotation adjustment strategy to ensure that the quotation can not only meet the customer's needs but also maximize the enterprise's profit.

[0049] The supplier selection module 4 calculates the comprehensive score of suppliers based on historical purchase data according to the weights of price, quality, and delivery options, and recommends the optimal supplier.

[0050] In the embodiment, based on historical purchase data, the supplier that provided the lowest price for the same type of product in history is returned as the optimal supplier, which helps the enterprise optimize the procurement cost on the premise of ensuring product quality.

[0051] The interface integration module 5 is used to integrate with the enterprise management system through the backend API, support processing multiple quotation information simultaneously, and return real-time prediction results.

[0052] In the embodiment, seamless integration with the enterprise management system is achieved through the backend API, supporting real-time processing of multiple quotation information, and providing functions such as quotation prediction, order conversion rate prediction, standard selling price prediction, and supplier recommendation.

[0053] As Figure 2 shown, the method for optimizing the procurement quotation of ship parts based on historical order data includes:

[0054] Step S1, preprocess the historical order data to generate complete order records.

[0055] In the embodiment, the data preprocessing stage includes:

[0056] Clean the historical order data, excluding unimportant fields and duplicate records;

[0057] Merge relevant data tables through inner join and left join methods to generate complete order records;

[0058] Fill in and process missing values.

[0059] Step S2, use the standard selling price algorithm to predict the optimal selling price in the current market based on the historical sales data of the product and multi-dimensional market characteristics.

[0060] In the embodiment, based on historical sales data, data screening and distribution analysis are performed according to product quality, sales type, site, and currency characteristics, and combined with multi-dimensional market characteristics to generate suggestions for the optimal selling price in the current market.

[0061] Step S3, predict the order conversion probability for each quotation by analyzing the order conversion rate of historical quotations.

[0062] In the embodiment, a binary classification model is adopted to predict whether a quotation will be converted into an order, and the quotation is optimized and adjusted based on the order conversion probability.

[0063] Step S4: Based on historical purchase data, calculate the comprehensive score of suppliers according to price, quality, and delivery option weights, and recommend the optimal supplier.

[0064] In the embodiment, based on historical purchase data, the supplier that provided the lowest price for similar products in history is returned as the optimal supplier.

[0065] Step S5: Send the prediction results to the enterprise's existing management system through the backend API, enabling the enterprise to make real-time and intelligent decisions in a dynamic market environment.

[0066] In the embodiment, the prediction results include: quotation prediction, order conversion rate prediction, standard selling price prediction, and supplier recommendation.

[0067] The following embodiments are used for detailed description (1)

[0069] This system is based on the historical order data within the company, including 5 database tables: InventTable, QuotionDetail, SalesDetail, SalesRetur, and SalesReturnInter. The data is integrated through inner join and left join methods to generate complete order information; before data integration, some unimportant fields are removed to improve processing efficiency; at the same time, the associated fields are aggregated to ensure data consistency and accuracy. The collection of historical order data is specifically divided into the following five steps:

[0070] 1. Connect related tables

[0071] Connect QuotionDetail1000, SalesDetail1000, and SalesReturnInter1000 to synthesize the complete order information record of the customer:

[0072] (1) Use the inner join method for QuotionDetail1000 and SalesDetail1000, and the associated fields are as follows:

[0073] QuotionDetail1000.Related_SalesLineRecid – SalesDetail1000.RMS_QUOLINEREFRECID

[0074] For a large amount of data, use SQL query statements to initially filter fields and remove unimportant fields to save memory. The specific statements are as follows:

[0075] qs_sql = "SELECT t2.SalesId,t2.ItemId,t2.CUSTLINENO,t1.CustNumber,\

[0076] MAX(t1.SalesType) AS SalesType,\

[0077] MAX(t1.SiteId) AS SiteId,\

[0078] MAX(t1.CURRENCYCODE) AS CURRENCYCODE,\

[0079] MAX(t2.EXCHANGERATE) AS EXCHANGERATE,\

[0080] MAX(t1.UnitId) AS UnitId,\

[0081] MAX(t2.NetSalesPrice) AS NetSalesPrice,\

[0082] SUM(t1.Qty) AS Qty,\

[0083] SUM(t1.DiscountAmountStandard) AS DiscountAmountStandard,\

[0084] SUM(t1.DisAmountStandard) AS DisAmountStandard,\

[0085] SUM(t2.CusOrigQty) AS CusOrigQty,\

[0086] SUM(t2.CusOrigAmount) AS CusOrigAmount,\

[0087] SUM(t2.SalesQty) AS SalesQty,\

[0088] SUM(t2.DisAmount) AS DisAmount,\

[0089] SUM(t2.SalesAmount) AS SalesAmount,\

[0090] SUM(t2.CurrencySalesAmount) AS CurrencySalesAmount\

[0091] FROM QuotionDetail1000 AS t1\

[0092] INNER JOIN SalesDetail1000 AS t2 ON t1.Related_SalesLineRecid = t2.RMS_QUOLINEREFRECID\

[0093] WHERE t1.Qty = t2.SalesQty\

[0094] GROUP BY t2.SalesId, t2.ItemId, t2.CUSTLINENO, t1.CustNumber"

[0095] By aggregating and processing the combined related fields to make them unique after joining, data errors are avoided. The specific statement is as follows:

[0096] SELECT SalesId, ItemId, CUSTLINENO, COUNT(*) AS COUNT FROM SalesDetail1000 GROUP BY SalesId, ItemId, CUSTLINENO ORDER BY COUNT DESC;

[0097] Convert the CUSTLINENO field from floating point to character type. Ensure data consistency and accuracy, especially to avoid type mismatches or unexpected calculation errors during data processing and table join operations.

[0098] (2) Use a left join between SalesDetail1000 and SalesReturnInter1000, and the related fields are as follows:

[0099] SalesDetail1000.SalesId – SalesReturnInter1000.ORIGSALESID

[0100] SalesDetail1000.ItemId – SalesReturnInter1000.ITEMID

[0101] SalesDetail1000.CUSTLINENO – SalesReturnInter1000.CUSTLINENO

[0102] Read SalesReturnInter1000 and perform aggregation processing by ORIGSALESID, ITEMID, and CUSTLINENO. The specific statement is as follows:

[0103] sri_sql = "SELECT ORIGSALESID,ITEMID,CUSTLINENo,CustNumber,SUM(Returngty)As Returngty

[0104] FROM SalesReturnInter1000

[0105] GROUP BY ORIGSALESID,ITEMID,CUSTLINENO,CustNumber"

[0106] Connect QuotionDetail1000, SalesDetail1000, and SalesReturnInter1000 through a left join to form a complete sales record. The specific statement is as follows:

[0107] merged_df = gs_df.merge(sri_df, left_on=['SalesId','ItemId','CUSTLINENO']

[0108] right On=['ORIGSALESID','ITEMID','CUSTLINENO'],

[0109] how='left')

[0110] II. Field Screening

[0111] To simplify the dataset, retain the keyword fields relevant to the analysis objective and eliminate redundant or irrelevant information to improve the efficiency of data processing and the accuracy of the results. The specific content is as follows:

[0112] (1) Eliminate nominal data: Since some nominal data (such as SalesId, ORIGSALESID, ITEMID, CustNumber) has no direct impact on the main objective of the analysis, these fields are eliminated to reduce data redundancy;

[0113] (2) Retain keyword fields: According to the purpose of the model, retain the keyword fields related to price such as quotation, unit cost price, and pre-sale price, and eliminate other redundant fields related to price but unimportant to ensure the simplicity and pertinence of the data;

[0114] (3) Exchange Rate Processing: The EXCHANGERATE field was removed to ensure that all price data is compared and analyzed on the same currency basis.

[0115] III. Handling Null Values

[0116] To ensure data consistency and integrity, missing values in the data were processed. The main content is as follows:

[0117] (1) Identifying the Source of Null Values: The missing values in the current data are mainly generated after a left join of the SalesReturn table with other tables. Since some orders have no return situations, null values appear in related fields;

[0118] merged_df['ReturnQty'].notnull().sum()

[0119] (2) Counting the Proportion of Missing Values: Through statistics, it is found that the proportion of missing values reaches 99.1%. Only 18,609 records show return situations, and most of the remaining records lack return data;

[0120] (3) Filling Null Values: Fill the empty ReturnQty field with 0 to handle the null value problem caused by missing return data. The specific statement is as follows:

[0121] merged_df['ReturnQty'] = merged_df['ReturnQty'].fillna(0)

[0122] merged_df.head(100)

[0123] IV. Feature Generation

[0124] To gain a deeper understanding of the price relationships and order conversion situations in the data and provide an important basis for further analysis and prediction of the model, it is mainly divided into the following three steps:

[0125] (1) Calculate the difference ratio of the quoted price relative to the net selling price through a formula. The specific formula is as follows:

[0126]

[0127] The specific statement is as follows:

[0128] merged_df['QNRate'] = merged_df['DiscountAmountStandard'] / (merged_df['NetSalesprice']*merged_df['Qty'])

[0129] (2) Calculate the difference ratio of the unit cost price relative to the net selling price through a formula. The specific formula is as follows:

[0130]

[0131] The specific statement is as follows:

[0132] merged_df['SNRate'] = merged_df['currencySalesAmount'] / (merged_df['NetSalesprice'] * merged_df['gty'])

[0133] (3) Calculate the order conversion rate through a formula. The specific formula is as follows:

[0134]

[0135] The specific statement is as follows:

[0136] merged_df['orderConversionRate'] = merged_df['OrderConversionRate'] = (merged_df['salesty'] - merged_df['Returngty']) / merged_df['SalesQty']

[0137] V. Perform One - Hot Encoding on nominal data

[0138] To improve the accuracy of the model when dealing with categorical variables, avoid biases caused by the direct use of nominal data, and improve the prediction effect, the One - Hot Encoding of nominal data mainly consists of the following two steps:

[0139] (1) Identify important nominal data fields, including SalesType, SiteId, and CURRENCYCODE;

[0140] (2) Perform One - Hot Encoding on these nominal data fields and convert them into a numerical form suitable for input to the machine learning model. The specific statement is as follows:

[0141] encoded_df = pd.get_dummies(merged_df, columns=['SalesType','siteId', 'CURRENCYCODE']) (II)

[0143] The core of this system is an algorithm model consisting of multiple modules, which are respectively used for standard selling price prediction, optimal quotation prediction, order success rate prediction, and supplier selection, etc.

[0144] By analyzing historical sales data, the system conducts data screening and distribution analysis based on multiple fields such as product quality, sales type, site, currency, etc., and finally generates a standard selling price suitable for the current market environment. This selling price algorithm can flexibly handle commodities in different cost ranges and minimize pricing errors. The specific selected fields are shown in Table 1 below:

[0145]

[0146]

[0147] Table 1

[0148] To improve the order success rate of quotations, this system first predicts the order success probability of each quotation, and then adjusts the quotation according to the customer's needs and market environment to increase the final order success rate. The model conducts binary classification prediction on the order transfer rate, effectively dealing with the impact of a large number of samples with an order transfer rate of 0 on the model accuracy. The specific field screening is shown in Table 2 below:

[0149]

[0150] Table 2

[0151] This system passes the ItemID and UnitId in the quotation and returns the procurement information of the k operators with the best prices that have been historically purchased for this ItemID for price inquiry. The specific returned fields are shown in Table 3 below:

[0152]

[0153] Table 3 (Three)

[0155] The system is integrated with the company's existing management system through the backend API. The API provides multiple functional interfaces, including quotation prediction, order success rate prediction, standard selling price prediction, and optimal supplier information query, etc. The system can process multiple quotation information at one time and return the prediction results, providing real-time support for the company's decision-making. Specifically, it includes the following five interfaces:

[0156] 1. This interface is mainly used to obtain the predicted quotation, order success probability, and order rate of the quotation.

[0157] The corresponding quotation information (one or more, forming a list) needs to be passed. Multiple quotation information can be processed simultaneously, so a single JSON data can pass multiple quotation information (such as the quotation information of multiple commodities in a whole ship's quotation), and the efficiency is higher than passing them individually.

[0158] II. Obtaining Quote Forecast and Order Success Rate: This interface is different from the one described in I. It only needs to obtain the order success probability and order rate of the quote. The order success probability and order rate can be predicted by continuously adjusting the quote manually.

[0159] Multiple quote information can be processed simultaneously. Therefore, multiple quote information can be passed in one JSON data (such as multiple product quote information in a quote for an entire ship), and the efficiency is higher than passing them individually.

[0160] III. Predicting Standard Selling Price: Predict the standard selling price of a certain product purchased by different users. If the predicted value is used as the value obtained by the original standard selling price algorithm of the company, then the latter is taken as the return result.

[0161] IV. Historical Optimal Supplier Information: Return the information of the optimal supplier for the historical purchase of a certain product.

[0162] Multiple quote information can be processed simultaneously. Therefore, multiple quote information can be passed in one JSON data (such as multiple product quote information in a quote for an entire ship), and the efficiency is higher than passing them individually.

[0163] Note: One more parameter is passed compared to the previous interface: CurrencySalesAmount (i.e., the total pre-discount quote given by the company to the customer).

[0164] V. Obtaining Quote Forecast and Order Success Rate: Return the historical selling price, amount order rate, and gross profit margin according to the customer ID and product ID.

[0165] 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 foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A ship parts procurement quotation optimization system based on historical order data, characterized by: include: Data preprocessing module, used to preprocess historical order data and generate complete order records; Standard selling price prediction module, which is used to predict the optimal selling price in the current market by analyzing historical sales data based on multi-dimensional market characteristics; The optimal quotation prediction module is used to predict the probability of success of the current quotation by analyzing the historical success rate of quotations, and adjust the quotation strategy based on the prediction results; The supplier selection module calculates the comprehensive score of suppliers based on price, quality, and delivery weights based on historical procurement data, and recommends the best supplier; The interface integration module is used to integrate with the enterprise management system through the back-end API, support simultaneous processing of multiple quotation information and return real-time forecast results.

2. The ship parts procurement quotation optimization system based on historical order data according to claim 1 is characterized in that: In the data preprocessing module, the preprocessing operation includes: Clean historical order data and remove unimportant fields and duplicate records; Merge related data tables through inner join and left join to generate complete order records; Fill and process missing values.

3. The ship parts procurement quotation optimization system based on historical order data according to claim 1 is characterized in that: In the standard selling price prediction module, based on historical sales data, data screening and distribution analysis are performed according to product quality, sales type, site and currency characteristics, and combined with multi-dimensional market characteristics to generate the optimal selling price recommendation for the current market.

4. The ship parts procurement quotation optimization system based on historical order data according to claim 1 is characterized in that: In the optimal quotation prediction module, a binary classification model is used to predict whether the quotation will be converted into an order, and the quotation is optimized and adjusted based on the probability of order completion.

5. The ship parts procurement quotation optimization system based on historical order data according to claim 1 is characterized in that: In the supplier selection module, based on historical procurement data, the supplier that historically provides the lowest price for similar products is returned, which is the optimal supplier.

6. The ship parts procurement quotation optimization system based on historical order data according to claim 1 is characterized in that: In the interface integration module, seamless integration with the enterprise management system is achieved through the back-end API, supporting real-time processing of multiple quotation information, and providing quotation forecasting, order success rate forecasting, standard selling price forecasting and supplier recommendation functions.

7. A ship parts procurement quotation optimization method based on historical order data, characterized in that: include: Step S1, pre-processing historical order data to generate complete order records; Step S2, using a standard selling price algorithm, based on the product's historical sales data and multi-dimensional market characteristics, predicting the optimal selling price in the current market; Step S3, by analyzing the order success rate of historical quotations, predicting the order success probability of each quotation; Step S4, based on historical procurement data, calculate the comprehensive score of suppliers according to price, quality, and delivery weight, and recommend the best supplier; Step S5, sending the prediction results to the enterprise's existing management system through the backend API, so that the enterprise can make real-time and intelligent decisions in a dynamic market environment.

8. The ship parts procurement quotation optimization method based on historical order data according to claim 7 is characterized in that: In step S1, the data preprocessing stage includes: Clean historical order data and remove unimportant fields and duplicate records; Merge related data tables through inner join and left join to generate complete order records; Fill and process missing values; In step S2, based on historical sales data, data screening and distribution analysis are performed according to product quality, sales type, site and currency characteristics, and the optimal selling price recommendation for the current market is generated in combination with multi-dimensional market characteristics.

9. The ship parts procurement quotation optimization method based on historical order data according to claim 7 is characterized in that: In step S3, a binary classification model is used to predict whether the quotation will be converted into an order, and the quotation is optimized and adjusted based on the probability of order conversion; In step S4, based on the historical purchasing data, the supplier that has historically provided the lowest price for similar products is returned, which is the optimal supplier; In step S5, the prediction results include: quotation prediction, order success rate prediction, standard selling price prediction and supplier recommendation.

Images